THE UNIVERSITY OF MICHIGAN SCHOOL OF SOCIAL WORK Progress Report CLIENT CAREERS AND PUBLIC WELFARE STRUCTURES Rosemary C. Sarri John E. Tropman Matthew Silberman Edward J. Pawlak Kenneth Badal ORA Project 03089 supported by: U. S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE SOCIAL AND REHABILITATION SERVICE GRANT NO. CRD-425-C1-9 WASHINGTON, D.C. administered through: OFFICE OF RESEARCH ADMINISTRATION ANN ARBOR March 1970

TABLE OF CONTENTS Page LIST OF TABLES vi LIST OF FIGURES x LIST OF CHARTS xi INTRODUCTION 1 Research Personnel Part I. Client Career Patterns and Organizational Characteristics: Child Welfare 7 1.1. INTRODUCTION 9 1.2. AGENCY MANDATES FOR CHILD WELFARE AND JUVENILE CORRECTIONAL PROGRAMS 10 Agency Manuals 11 1.3. ANALYSIS OF INFORMATIONAL FORMATS IN FIVE STATES 13 Comparative Analysis of Information Formats 13 Chart Glossary 15 Child Welfare Agency Formats and Variables 21 Correctional Agency Formats and Variables 25 Juvenile court data 40 Probation formats 40 Institutional formats 42 Discussion 46 l.4. CONCEPTUAL FRAMEWORK 47 Variables 48 Independent variables 48 Dependent variables 49 Intervening variables 49 Adequacy of the State Corrections Data 50 1.5. SOME BASIC DETERMINANTS OF CLIENT CAREER TENURE 51 Data 553 Methodology 54 Findings and Discussion 57 Part II. Survey of Local Welfare Administrators 67 2.1. INTRODUCTION 69 2.2. SURVEY FINDINGS 70 Agency Size and Staff Workload 71 iii

TABLE OF CONTENTS (Continued) Page Turnover Rates 78 Staff Growth 80 Bureaucratic vs. Professional Promotion Criteria 83 Program Emphases 87 Child welfare services vs. family-centered problems in child welfare agencies 88 Problems in Analysis 89 Missing Data 89 Caseload size and structure 91 Supplementary Data 94 2.3. METHODOLOGY 94 Survey Foci 94 Selecting the States for Intensive Study 100 Selecting the National Sample 100 Collecting the Data 104 Structural complexity and supplementary information 105 Operational Problems in Processing Data 106 Structural complexity 106 Functional differentiation 107 Analysis of sentence completion items 108 Reliability of open-ended codes 110 Findings from Analysis of Sentence Completion Items 111 2.4. STRUCTURAL COMPLEXITY IN STATE PUBLIC WELFARE SYSTEMS 112 Integrated Structures 117 States in which all local public welfare programs are administered through a single county office 117 States in which there is a single-county office for all PA programs, usually excluding general assistance 117 States in which the director is responsible for a district or two or more counties 117 Bifurcated Structures 120 States in which child welfare is administered separately 120 Program Abbreviations 124 Part III. Social Indicators and Welfare Structures 125 3.1. INTRODUCTION 127 3.2. COMMUNITY COMPOSITION-THE EMPIRICAL DEVELOPMENT OF COMMUNITY INDICATORS 129 3.3. THE SOCIAL MEANING OF SOCIAL INDICATORS 131 Social Stratification 131 Race or Mobility Coagulation 133 iv

TABLE OF CONTENTS (Concluded) Page Community Complexity 137 Change and Stabilization 141 Summary and Conclusions 142 3.4. THEORETICAL APPROACHES TO COMMUNITY STRUCTURE 144 3.5. THE COMMUNITY AND THE ORGANIZATION 148 3.6. A DEMOGRAPHY OF THE SELECTED STATES 151 3.7. WELFARE STRUCTURE OF THE STATES 154 Welfare Data 154 3.8. WELFARE DEMOGRAPHY 157 3.9. RATES AND UTILIZATION BY COUNTY 166 Part IV. Research Data Systems: An Exploration 173 4. 1. INTRODUCT ION 175 4.2. INFORMATION PROBLEMATICS 176 4.3. SECURING THE DATA 176 4.4. UNDERSTANDING THE DATA 179 4.5. DATA PROCESSING AND MANAGEMENT 180 Appendix 1, A COMPUTER APPROACH TO EVALUATING PROGRAM ALTERNATIVES FOR CHILD WELFARE CLIENTS 185 2.A. SURVEY QUESTIONNAIRE ENTITLED "TO PROVIDE HOPE" —FOR ADMINISTRATORS 201 2.B. HOW RESPONSES WERE SOUGHT FOR COUNTIES WITH IRREGULAR ORGANIZATIONAL DESIGNS 228 2.C. RESPONDENT AND REQUEST RATES 235 2.D. PROCEDURES FOR HANDLING QUESTIONNAIRE DATA IN RELATION TO COUNTY ECOLOGICAL DATA 240 2.E. GENERAL CODING INSTRUCTIONS 243 3.A. CRITICAL DIMENSIONS OF COMMUNITY STRUCTURE 271 5.B. THE AMERICAN STRATIFICATION SYSTEM 289 3.C. SOCIAL STRATIFICATION, SOCIAL INTERVENTION, AND COMMUNITY COMPETENCE 297 v

LIST OF TABLES Table Page 1.1. Variables for Which Information is Collected Re Child Welfare, in Four States 22 1.2. Variables for Which Information is Collected Re Corrections Institutions in Four States 43 1.53 Percent Distribution and Rank of Predictive Efficiency of Selected Background for Length of Commitment 56 1.4. Percent Distribution and Rank of Predictive Efficiency of Selected Background Variables for Length of Commitment 59 1.5. Percent Distribution of Relational Fertility of Selected Variables for Predicting Length of Commitment, By Reason for Service 60 1.6. Distribution of Wards in West Central State, By Race and "Problem Label" 65 2.1. Number of Client Applications by State 72 2.2. Size of Staff By State 72 2.3. Size of Staff By Number of Client Applications in West Central State 73 2.4. Client/Staff Ratio and Client/Professional Staff Ratio, By State 74 2.5. Client/Staff Ratio and Client/Professional Staff Ratio By Type of Agency in Eastern State 76 2.6. AFDC Caseload and Child Welfare Caseload, By Size of Agency 77 2.7. Turnover Rates By State 78 2.8. Turnover Rates By Type of Agency in Eastern State 79 2.9. Growth Rates By State 80 2.10. Growth Rates By Type of Agency in Eastern State 81 vi

LIST OF TABLES (Continued) Table Page 2.11. Growth Rates By Size of Agency in West Central State 82 2.12. Importance of Promotion Criteria By State 84 2.13. Importance of Promotion Criteria By Size of Agency 85 2.14. Importance of Competitive Examinations By Size of Agency with Middle Group Considered Separately 85 2.15. Importance of Promotion Criteria By Type of Agency in Eastern State 86 2.16. Actual vso Ideal Agency Effort Allocated to Mental Health and Addiction By State 87 2.17. Actual vs. Ideal Agency Effort Allocated to Family-Centered Problems and Child Welfare Services in Eastern State's Child Welfare Agencies 88 2.18. Percentage Missing Data for Client Applications and First Time Applications By State (and Program) 90 2.19. Percentage Missing Data for Total Client Applications and First Time Applications By Professionalism Among Child Welfare Administrators in Eastern State 91 2.20. Mixed Caseload By State 92 2.21. AFDC Caseload Size By State 92 2.22. Child Welfare Caseload By State 93 2.23. Agency Objectives and Administrators' Preferences for AFDC Program 95 2.24. Administrators' Report of Proportion of Clients "Completely Helped" By State 96 2.25. Problem Producing Situation for AFDC Clients by State 96 2.26. Importance of Promotion Criteria 97 2.27. People's Beliefs About AFDC Mothers By State 97 vii

LIST OF TABLES (Continued) Table Page 2.28. Percent of Administrators With M.S.W. Degree By State 98 2.29. Social Class Distribution of Administrators' Fathers 98 2.30. Opinions of Administrators and Their Friends on Welfare Issues By State 99 2.31. Sampling Rates 103 2.32. Reliability of Codes 110 2.33. States With a Single County Agency Administering All Major Programs 118 2.34. States With Single County Agencies For All Programs Except General Assistance 119 2.35. States With District and Multiple-County Administered Programs 119 2.36. States With Bifurcated Agency Structures 120 5.1. Percent Distribution By Educational Level of Men 20 to 64 Years Old and Their Fathers, By Color: March 1962 135 3.2. The Intercorrelation Matrix of Community Composition Variables, Using Selected Other Variables, 154 Communities 143 353. Values and Ranks for Selected Demographic Variables, By State 152 3.4. Number of Cases and Number of Recipients for General Assistance Programs, Five States, 1968 156 3.5. Recipients of Public Assistance, and Total Assistance Payments By Program, State, June 1964, June 1966, and February 1968, For Five States 158 3.6. The Rate, Per Thousand Population, of Use of Public Assistance Programs and the Payments Under Those Programs, Five States, For June 1964, June 1966, and February 1968 159 viii

LIST OF APPENDICES Appendix Page I CALORIMETER RESPONSE........................ 55 II RATE EFFECT................................. 59 III BEAM RATE MEASUREMENT....................... 6 IV CALORIMETER BACKSCATTER..................... 63 V GEOMETRIC RESPONSE FUNCTION FOR THE TRANSMISSION COUNTERS...................... 65 ix

LIST OF FIGURES Figure Page 1.1. Total sample. 61 1.2. Neglected children. 62 1.3. Children whose parents have a problem. 63 1.4. Children with their own problem. 64 3.1. Salient community characteristics. 130 3.2. Complexity as a step function of system size-some possible relationships. 140 35.5 General systems paradigm of formal organizations. 149 3.4. Reporting formats for 1960, 1964, 1966, and 1968 recipients of public assistance. 155 x

LIST OF CHARTS Chart Page I. Information Formats about Child Welfare Clients in Four State Departments of Social Welfare and One County Juvenile Court 16 II. Information Formats about the Families of Child Welfare Clients in Four State Departments of Social Welfare and One County Juvenile Court 17 III. Information Formats about the Careers of Child Welfare Clients in Four State Departments of Social Welfare and One County Juvenile Court 18 IVo Information Formats about Juvenile Corrections Clients in Four State Departments of Social Welfare 26 V. Information Formats about the Parents of Juvenile Corrections Clients in Four State Departments of Social Welfare 32 VI. Information Formats about the Careers of Juvenile Corrections Clients in Four State Departments of Social Welfare 34 VII. State Structural Patterns for Local Welfare Program Administration by Regions 115 VIII. Distribution of Public Assistance and Child Welfare Agencies, by Regions in Texas 122 xi

INTRODUCTION This report covers activities during the second year of the research project entitled, "Client Careers and Welfare Structures," and presents plans for completion of the study and preparation of final research reports.* This project has four main objectives which are identified below along with the progress to date toward their achievement. The first objective is to study the organizational careers of cohorts of clients served by agencies within state departments of social services (and corrections where applicable) in five states, over a two- to three-year period of time. Relationships between personal and social attributes of clients and characteristics of agencies are being examined to determine the factors which influence and/or are associated with differential career patterns. Children who are wards of the state because of dependency, neglect, or delinquency were chosen for this phase of the study because of the availability of the information necessary to test some of our propositions in the sample states. The development of instruments for assessment of careers was a part of this objective, because we did not wish to use tenure as the sole criteria of differential careerso Analysis of goals, organizational structures, and processes of decisionmaking in local public welfare agencies was an additional objective. This agency was selected because of the crucial role which it plays in the initial nomination of clients to particular programs, and it was assumed that differences among these organizations would have consequences for career patterns. The third. objective was to relate demographic and social indicator data about cities and counties to characteristics of local public welfare agencies and to client career patterns. We hope to identify some of the factors which influence career nominations and processing, and also to factor out interactions between ecological and organizational phenomena as they affect our understanding. Formulation of tentative guidelines for the development of informationprocessing systems in state departments of social services was the last objective. These guidelines relate particularly to the decisional structures and *The first progress report covering the period August, 1968, to March, 1969, was submitted earlier. It included information about the research design, instrument preparation, and preliminary findings. This report builds from the earlier statement and from the original research proposal, so the reader who is interested in the overall design is referred to those statements. 1

processes which affect client problem labelling, and the organization and delivery of services. Far too often in the past elaborate classificatory systems were developed but they bore little or no relationship to operational patterns and decision contingencies. It is hoped that some preliminary test of these guidelines will be possible to assess the feasibility of engineering more efficient and effective procedures. Millions of children are processed as child welfare wards each year in the United States. Some remain as wards for very brief periods of time; for example, infants to be placed for adoption are often processed within a few days after birth. In contrast, other children may remain as wards for many years, being transferred among a variety of agencies and programs. The label assigned in the client's early processing as a ward has qualitative implications which may have positive or negative consequences for the client's future life chances. Unfortunately, far too little is known about the actual careers of dependent, neglected, or delinquent children despite the obvious policy implications of having such information available on a systematic basis. In this study we have found that it was possible and viable for research purposes to obtain longitudinal data about these children from state agencies and analyze it extensively to identify career modalities and the factors with which they are associated. In the analysis of informational formats of five states we observed that these formats are partial indicators of agency goal priorities and operational patterns. Marked differences among the states were noted as well as between different programs within a given state. Analysis of the tenure of client careers in two states where length of commitment was examined as a dependent variable revealed that the concept of categoric risk explains the agency operations which produce longer tenure and more negative sanctions for lower class, non-White children. It was also observed that organizational variables could be examined separately from personal attributes despite the possible interaction between these sets of variables. The organizational variables were further observed to be very important determinants of career patterns. Further study is now underway to refine the analysis procedures, to develop typologies of careers, and to formulate additional criteria for assessment of the impact of different types of careers. No human service organization in this country is beset with more crises and trouble than are local and state welfare agencies in nearly every state of this country. One needs only to follow the daily newspaper to be aware of the limelight in which these agencies exist today. They are subject to sharp criticism from all sides; from special interest groups, from professional organizations, from client associations, from taxpayers, and from legislators. Certainly it is a time when policy makers and administrators as well as direct service practitioners could benefit from greater knowledge about the operation of local welfare agencies-so as to be able to place priorities rationally, to know which modifications will achieve certain outcomes, and to know how to promote and/or adapt to major change which is probable in the welfare system within the 2

next five years. Although we have not attempted a comprehensive study of public welfare, we have tried to place our survey in a context which will permit more general assessment of the meaning of the findings beyond the nomination and processing of wards. In this context the decision was made to pursue the national sample survey. Results from the intensive survey of all counties in four states and from the national sample survey have been gratifying, particularly from the interest shown about the study among local administrators across the country. That this is a time of crisis and concern for them appeared to us to be evidenced by their interest in this research and its findingso From our analysis it became quickly clear that local welfare departments do not exist uniformly on a county basis, nor is there uniformity in the range and types of program offeredo There are important differences among local welfare agencies in structural complexity and functional differentiation and these influence perspectives and operational patterns. Other differences were noted in agency size, in client/staff ratios, in staff turnover and growth rates, in administratorsv priorities for services, and in indicators of bureaucratization. The relative adequacy of information-decision systems at the local level was evidenced in the ability of respondents to provide us with certain kinds of data about numbers and types of clientele. Plans are underway for further analysis of the four states and of the national sample with attention being given to both intra- and interorganizational variables. Ecological factors, such as rural-urban differences, level of poverty, education, stratification migration and mobility will be studied to assess their import for welfare agency operationso In addition there will be study of parent organization policies as these are reflected in agency manuals of procedure, We hope to link variables relevant to understanding local welfare agency operations with variables relating to career patterns of child welfare wards and with ecological variables, The third phase of this project involves a modest attempt to conceptualize and measure some social indicators which could be used to understand further operational patterns of local welfare agencies. A wealth of ecological data about counties and states has been obtained from the City and County Data Book, from a tape prepared by the Office of Economic Opportunity of health, education, poverty, welfare, and other information about counties, and from federal reports of welfare recipients and payments. Some dimensions of community structure have been conceptualized including stratification, mobility coagulation, complexity, and change. Measures of these variables will be linked with data about welfare structures. The last objective of this research is to utilize knowledge gained from the above research to develop guidelines for information-decision systems in local and state welfare agencies. Our experience with data from these agencies suggests that such systems are sorely needed if efficiency and effectiveness 5

are to be increased. Many of the problems which we have encountered in collecting, managing, processing, and analyzing data have meaning for the design of agency information systems. There is every reason to believe that agency staff will respond favorably to recommendations for the development of such systems. Research Personnel This has been a very busy year for the staff and other researchers who have been associated with this project. Dr. Sarri, Principal Investigator, was on sabbatical leave for part of the year, and was able to devote about 30 percent of her time to the project. Dr. Tropman spent 50 percent of the Spring-Summer term, 1969, with the project and 20 percent or more of his time during.the remainder of the year. Dr. Wolfgang Grichting served full-time as a Research Associate until May, 1969, when he left for Taiwan where he is now conducting a research project in affiliation with the National University. Miss Suann Hecht served as a Research Assistant until November, 1969, when she resigned to become a program director in the international headquarters of Care, Inc., in New York. Mrs. Nancy Berla served as Administrative Assistant until November, 1969, for this project and for the other research projects affiliated with it under the direction of Professors Thomas and Litwak. She has been succeeded by Mrs. Ann Smith who will continue to function in this role and in other areas in the year ahead. No project of this scope could ever succeed without creative and responsible work by many staff members. In addition to those who have been employed on a full-time basis or have primary responsibility for the overall project effort, a number of persons have assisted as part-time staff or in a voluntary capacity as consultants. Mr. Matthew Silberman is a part-time Research Associate. Along with Miss Hecht and Mr. Kenneth Badal he has carried major responsibility for the survey of local welfare administrators. Mr. Silberman is completing his doctorate in sociology. Mr. Badal was a member of the public welfare agency staff in Fulton County, Georgia, until January, 1969, when he enrolled in the graduate program in social welfare administration. Among the faculty colleagues of The University of Michigan School of Social Work who have provided consultation in the survey phase of the project are Professors Wayne Vasey, Philip Booth, Paul Glasser, Robert Vinter, and Yeshekel Hasenfeld. Their advice has been invaluable and they have provoked us with penetrating questions. Primary responsibility for analysis of client-career patterns was assumed by Dr. Grichting and subsequently by Dr. Sarri and Edward J. Pawlak, a doctoral student in sociology and social work. Mr. Pawlak is completing his doctoral dissertation with the project. He is the recipient of a Public Health Service Special Fellowship, HS0034-03, and therefore, is not paid from project funds although he has devoted nearly full-time effort to the project since September, 1969. Additional part-time assistance with this phase has been provided by 4

two graduate students in social welfare administration who were formerly public welfare workers in Montana, James McCabe and Thomas Egano Dr. Tropman assumed responsibility for the third part of the project, that involving the study of social indicators and welfare structures. He has been assisted by several graduate students from the social welfare policy specialization and the joint doctoral program in sociology and social work, including Howard Hammerman, Ted Reid, Diana Wright, and others. Of great importance to the total project has been the hard work of those who have assisted with the coding, computer programming, and analysis. These staff include Suzy Southon, Marlene Cohen, William Murphy, and Len Harding. Last, but far from least has been the contribution of the three secretaries who have very ably assisted with all phases of project activity: Mrs. Sheila Horsley, Mrs. Margaret Lemley, and Mrs. Barbara Hiebbner. Many of the above staff will continue to be associated with the project during 1970-71, and in addition, it is expected that several other masters' and doctoral students in social work and sociology will also be involved. 5

PART I CLIENT CAREER PATTERNS AND ORGANIZATIONAL CHARACTERISTICS: CHILD WELFARE 7

1.1. INTRODUCTION A major objective of this total research effort has been to obtain knowledge of client career patterns of child welfare wards and of the major variables which produce, influence, or are associated with differential careers. Longitudinal study of careers has long been recognized as important by students of human service organizations, but until recently comprehensive studies were almost impossible because of the methodological limitations. As a result, one had to be content with the results from elaborate case studies of individual careers. It was impossible to generalize to larger populations of clients because of lack of knowledge about the relevant dimensions for those populations. Modern computers and new information processing procedures now make elaborate analyses possible and relatively inexpensive. Knowledge about client career patterns is of as great importance to the social policy maker and agency administrator as it is to the social scientist. Millions of children in the United States today are child welfare wards and as such are the objects of powerful intervention by the state, Formal labels and statuses are applied by the state in the processing of child welfare wards, and these social classifications may have great positive or negative significance for the life chances of the person who is the object of the action. From the perspective of the organization and society, tremendous resources are expended in the operations of these programs, and yet little is known about differential outcomes for clients who experience differential types of programs and services. For example, few social agencies could report regularly and accurately the length of service for different categories of clients despite the obvious importance of such information for planning and evaluation. Increasingly administrators, legislators, and others in crucial decision-making roles are requesting such information on a routine audit basis so that decision-making can be a more rational and objective process. Recently Secretary Robert Finch spoke about the great need for more objective information about clients and services for decision-making at several governmental levels. Our research to date clearly indicates that in the five states where we have obtained information about the career patterns of child welfare wards and juvenile offenders all of them collect far more information than they ever analyze or use for policy making and other types of organizational decisions. Most of the information is used only for fiscal reporting and auditing purposes rather than for organizational problem-solving, for evaluation of performance, or for staff development, to mention only a few of the areas for which objective information is critically needed. Efficiency and effectiveness of organizational performance cannot be improved unless there is greater objective knowledge of actual operational patterns and their outcomes. The availability of electronic data processing procedures today makes this task much easier and far less time-consuming that it was a short time ago. But, 9

these procedures are of little value unless the appropriate data are carefully collected and prepared for such processings. Lastly, organizational mechanisms must be created to facilitate utilization of this information for problem-solving throughout the organization. In our research thus far we have developed an even greater appreciation about the nature of the above problems and their consequences for the operation of social agencies. At the same time we believe that some of the findings from this research have many significant implications for examining current practices for processing child welfare wards, for modification of some policies and agency manual provisions which reflect situations that are no longer prevalent, and for the development of a system which will permit more rapid feedback of information about client career patterns for staff decision-making. In this section of the progress report we wish to highlight the following research efforts and findings which have a bearing on the problems and issues cited above. (1) Selection of child welfare wards for study of client careers and illustrations of the mandates for services to operating agencies. (2) Presentation and analysis of informational formats and categories used in the sample states. (3) Methodologies for analysis of client career patterns, including technical problems in data collection, management, reduction, and synthesis. (4) Progress in development of a theoretical framework, including identification of independent, intervening, and dependent variables, and statement of hypotheses. (5) Presentation of findings from analysis of the tenure of client careers. (6) Preparation of an instrument for obtaining professionals' ratings of the quality of service in different programs and agencies for clients with specific types of problems. 1.2. AGENCY MANDATES FOR CHILD WELFARE AND JUVENILE CORRECTIONAL PROGRAMS If one is to study client career patterns, it is necessary to obtain longitudinal data about client experiences into, through, and out of social agencies, and also to be knowledgeable about the major contingencies influencing careers. Earlier studies by the project director of juvenile court case processing and of public school student cohorts indicated that it was possible to obtain routine 10

data from these organizations which could be used to delineate client career patterns. Some preliminary contacts were made with several public social agencies to determine the feasibility of this approach and to learn about the types of data which were recorded routinely about individual clients so that it could be made available for research purposes. Our earlier research progress report, submitted in March, 1969, delineated the rationale for the selection of child welfare wards, for the choice of the five-state sample, and for determination of the time periods to be included. By way of a brief summary, we decided to study the careers of all children who were wards of the state in five different states during a two-year period of time between late 1964 and 1968.* Child welfare wards were defined to include children officially labelled as dependent and neglected and juvenile offenders so adjudicated by the juvenile court. Both population groups were selected because in some states they are handled as a single population group, whereas in others they are handled separately. Data have been obtained from the five states. In this report, these states are identified as follows: Eastern, Western, Central, East Central, and West Central. Analyses are nearly complete for East Central and West Central states; it is well underway for Central and Eastern states; and it is in the beginning stages for Western state. As we shall note subsequently, preparation of the data for research-type analysis required more time than we had originally anticipated, and a major problem in slowing our processing was the condition of the data. It became quite obvious from the amount of "missing data," "improper codes," and "wild punches" that these data were not extensively analyzed by staff within the several states. If they had been, many of the problems which we uncovered would have been corrected. Agency Manuals Sections of state agency manuals as they pertain to child welfare and juvenile correctional populations are being studied to determine the mandates and formal goals under which these agencies operate. These manuals also detail career contingencies and formal operating procedures which are to be carried out by workers providing services within each of the program agencies. In most cases these manuals are extremely detailed about these procedures, and many also contain obsolete procedures no longer in use. Nonetheless, they must be examined to determine formal procedures and contingencies. Examination of agency manuals reveals marked interstate differences in the definition of agency mandates and in their conception of the problems with which *Because of data gathering and processing procedures in the several states, we had to provide a range of time during which we could use two-year longitudinal data. In all cases but one the data are 1966-68 data. 11

they are confronted. At the same time there are many areas of agreement and commonality among the states. In the five manuals of the sample states, it was made explicit that every child has a right to care and protection as a part of the basic American culture, and that this right is reinforced by the legal system. Parents are viewed as primarily responsible for a child's welfare, but if they are absent, or if they fail to do so, organized society must assume this responsibility. The various state departments of social services are typically assigned this formal responsibility. Some states limit their child welfare services to explicit problems and to those who meet financial eligibility criteria, or are without parents. In other states the services and eligible clientele include a broader range of situations and problems. In Western Central state, for example, children to be considered as eligible for child welfare services include those with the following problems: neglect, a child separated from his family, illegitimate parenthood mental illness, delinquency, truancy, school drop-out, educational handicap, and indigence. In addition families with a broad range of problems may be offered services if it would help children to remain in their own home. These include financial assistance, marital and other counseling medical treatment, and day care. The range of services in Eastern state is broadly defined to include state payment for local police services for juveniles. The manual states the objective of the program is: "...prevention and control of juvenile delinquency and the protection of children and their families through special services within police departments for the handling of juveniles who come to the attention of police." Various agencies are mandated with responsibility for child welfare services in the different states. In nearly all cases, however the juvenile courts, county (or district) welfare agencies, voluntary agencies, and public and private residential institutions are included in the network of agencies providing services to child welfare wards. In some states these agencies are, or view themselves as highly interdependent, and even the law may require coordinated actions. In others they are quite independent and act almost without reference to the other agencies in their "organization set." State laws generally place overall responsibility with the state director of social services, but structural differences within and between states produce markedly different patterns. For example, in Eastern state, there is a proliferation of voluntary agencies with the result that state and county agencies contract these voluntary agencies to provide services to child welfare wards. In contrast, Central state provides the bulk of service through public county and state agencies. West Central provides a third pattern with complex interorganizational relationships between voluntary and public agencies. Study of state agency manuals will be completed and we plan to integrate 12

this finding with those from the survey of local welfare administrators. Following that we plan to examine these findings with reference to the actual career patterns of clients within and among the five states. 1.35 ANALYSIS OF INFORMATIONAL FORMATS IN FIVE STATES One objective of the client careers project is the development of a methodology for the processing and analysis of client career data which can be used by state departments of social welfare and other public and private agencies to rationalize decision-making about clients. A requisite for the development of such a methodology is the concurrent development of informational desiderata for the study of client careers. We are utilizing two approaches to achieve the latter objective. First, we are doing a comparative analysis of the informational formats provided to us by five states' departments of social welfare about their child welfare and corrections clients. The specific purposes of the analysis are: (1) to determine the similarities and differences in the formats; (2) to determine the kind of analysis that is possible with each state's data; and (3) to identify and to merge the best features of each of the formats. Secondly, we are formulating a formal or theoretical model of client careers to determine the kind of data that are needed to measure certain aspects of client careers. In this section of the report we will first discuss the findings from the comparative analysis of the formats. Subsequently, we will describe our theoretical model and we will use it to determine what is problematic about client careers and to evaluate the adequacy of the state data for the study of client careers. Comparative Analysis of Information Formats Client career is a concept that refers to the experience of individuals in their roles as recipients of services and/or objects of organizational processing and decision-making. The term experiences refers to: (1) the events that determine or that are associated with the occupancy and termination of an organizational position; (2) the duration of occupancy in a position; and (3) the pattern of movement through organizational positions and organizations. In order to study these and related phenomena five state departments of social welfare were asked to provide us with the information which they routinely collect on their child welfare and juvenile corrections clients. Only one state was not able to provide any information about its child welfare clients because their data are not processed by means of an electronic data processing system. The remaining state departments provided us with the information we requested in the form of magnetic computer tapes or IBM cards which contain client data taken from admissions, release, and other official reporting forms. The forms, 13

the instructional booklets for their completion, the code books, and the computer tape layouts were also sent to us. The forms and their respective instructional booklets identify the kinds of information which the state department requires of its county subsidiaries. The code books and the computer tape record layouts indicate how the information is organized and stored. A review of the forms and code booksreveals e that they appear to be organized in a manner that facilitates the recording and coding of information. This kind of organization of the code books makesit difficult to comprehend the type and range of variables within a particular state and it makes it virtually impossible to acquire an appreciation for the similarities and differences between the states. In order to overcome some of these shortcomings and to facilitate the use of the data for our analytic purposes we decided to compile all of the variables into several master charts (see Charts I to VI below). The charts are divided into two main sets: the Child Welfare Set-Charts I to III; and the Corrections Set-Charts IV to VI. Within each set the data are divided into three classifications: (1) information about the clients, i. e., their personal and social characteristics; (2) information about the i parents or family of the client; and (3) client career information. The client career variables are organized into several sub-classifications: admissions or intake information; criminal, clinical, and correctional history (Chart VI); information processing and decision-making pertinent to placement and disposition; experience or adjustment in program or institution (Chart VI); tenure of client career; and termination of or change in client career. The charts are organized such that the information collected by the state department is listed in the row margins and the name of the state department is listed in the column margin. The cell entries in the chart specify the categories within each variable. Most of the variable names in the charts are listed exactly as they are found in each state's format. However, in the interest of parsimony and comprehensibility certain conceptually related variables were clustered under a generic variable name, e.g., see variables 10, 18, 20, 29, 30, and 31. Most of the categories in the cells of the charts are also listed exactly as they are found in the state formats. However, in some cases a cell entry merely gives a general description, because an exact listing would have revealed the identity of a particular organization or state, e.g., see East Central State's variable 29. In some instances an exact listing of categories was prohibitive because of the number of categories, e.g., for variable 29 Western State listed 87 agencies; certain law violation codes enumerated over fifty specific offenses (see variable 62). Thus far, we have discussed what is germane to all the charts. In the following sections we will present and discuss the charts separately. 14

Chart Glossary 1. A zero (0) in a cell means that the variable in that particular row does not appear in the state's information format. 2. A check mark () in a cell means that a state department does collect information about a particular variable, and that given the variable the cell entry should be self-evident, e.g., there is a check mark in the cells referring to the variables Case Number and Age which means that the state department assigns a number to every client and that it records his age. 39 Mo - Mother 4. Fa - Father 5. NR - Not Reported 6. Inap - Inapplicable 7. Unk - Unknown 8. Mos - Months 9. Yrs - Years 15

Chart I. Information Formats about Child Welfare Clients in Four State Departments of Social Welfare and One County Juvenile Court, T____ __Eastern State. n East Central State West Central State Western State Urban County Variables Adoptions Nondeliquency Court Corrections & Child Welfare Dependent & Neglected Child Welfare Service ~_____..Records_ _ 1. Case Name / / / / 2. Case Number / / / / / 3. Birthdate/Age Age in yrs & mos Age in yrs & mos Month, Day, Year Month, Day, Year Month, Day, Year 4. Birthplace Within county, In 0 East Central State, U.S., 0 0 Eastern State, Out Canada, Europe, Central or state, Out of U.S., South America, Other, Unk. Unk. 5. Birth Status Born out of wedlock 0 0 Legitimate, Illegitimate- 0 Born in wedlock, Other, Paternity established, Unk. Illegitimate-Paternity not established, Foundling, Unk. 6. Sex Male, Female Male, Female Male, Female Male, Female Male/female are combined with race & legal status 7. Race White, Negro, Other, White, Nonwhite White, Negro, Indian, White, Negro, Indian, Unk., N.R. Other, Unk. White-Indian, White- White, Negro, Indian are Negro, Negro-Indian, combined with race & legal Other, Unk., N.R. status 8. Religion 0 Roman Catholic, Roman Catholic, Protestant, Roman Catholic, Lutheran, O Protestant, Jewish Jewish, Other, None Other Protestant, Jewish, Other, Unk. 9. Age at time of: Admission/Referral / Petition / Commitment Placement An approximation can An approximation can Can be obtained by cross Can be obtained by cross Closing obtained by complex obtained by complex tabulating birthdate & tabulating birthdate & date manipulation & manipulation & date of action of action codes controlling interpretation of interpretation of controlling for type for type of action the data the data of action 10. Living Arrangements-Eaot Public Agency/Institution Centuat State Voluntary "/ " Mo & Stepfa Mo & Stepfa Type of Pfacement PLio0 Independently by Parent Fa & Stepmo Fa & Stepmo O to DecAee-EaoteAn State " by attorney, physician, Other family, home, Relatives House of relatives State Adoptions or minister relatives County foster home Chkid'& Abode-EateAtn " by other individual Foster parents Foster parents Voluntary agency foster home State-CouAt Recotds No placement-child in home Institution, Agency Institution Institution Wheheabout -WestetAn " "-child not home Adoptive parents Hospital House of parents Staute Unk, N.R. Both parents Both parents Day care center Mo only Mo only Homemaker Fa only Fa only Voluntary agency maternity Parent & paramour Elsewhere, Unk. home, elsewhere 11. Child Characteristics 0 0 0 Physical handicap 0 Dull intelligence Mentally defective or epileptic Personality or behavior problem, older child Minor physical handicap Severe physical handicap 12. Previous Contact 0 New charge New New case Rehearing Readmission 0 Previous termination Never in court before First time this yearknown previous years 2nd time etc. 16

U CQ 4 0 0 IZI < 4-) U X Q 4 4 0 04 - n?J bo 0o ) H o I0 O 4 4-4 4- r U 4-c H > 0 04 t 0. ^ -~ 4J 4 HJ -H - 0 4 i-.3 4. -.,-4 4) ) a) 0 0 0 0. 4- -H C 40 C U - 0-U4-4 U C (4-A Q ( U & U U U 000 C (0 0 * 0 *i *Hr 4-) 4-1 0 P 4-J U 0 0 0~ - U, U f) * )g O -UOH — H - Ua 0 aO U 1 o4-4J Z -t - z -H co 4- CO 3 U t u U (La c o 4J -i. cc o U -HU c H C ) * HU.I W UOU oQU Q -4 D p 0 a) 0 co * 00 m. H c (L) co X r=E +-> a!3Q QW O > v X S u >S c ^ >,'o 4 = Z:2: n C V)') p U0 X L 4 Z M *4 Z C Q o U 03' U U r), 4-) cU.3 a- - U *r- -3 (0.S 0 4 —.4,) (.) -''0 ) 4 C 3U 44 U U U> toc-0 H a) U J 040 =3 1 0 UU 0UH 2E, - - o~ 0~ 0, 0 ->t t)... - ~ -. 3 r O o 4 H'-4-) Q 0 UUUnoUHUO0)3. 4 c ) 4 Z-I t- a)?^ H * * * 0 1LI UO -_ ( ol - C 4 —3 ~-< O U E co; Cl 0 * U C aJ > *'IQ E.-' 45) 00 U W W U W —.,C - U w' u E EU. ) U04U -H' (0 U 0 U, O H U U E q n.- C o a o - 44 4> 0 0 r4 Ua c ca o U 0 0 4 -3 U U U I H r-.I O 2 0 4UOUOU 4-Jl *H H -P O ).., O t - 0 U-OO- *H4 4- C( c 0 < U 4- ) 4 4- > C SU, 0v0 Q. * 0 0 0 c 4 4. 0 3 UU 4)

Chart III. Information Formats about The Careers of Child Welfare Clients in Four State Departments of Social Welfare and One County Juvenile Court. _____E___astern State East Central State West Central State Western State Urban County Variables Adoptions Nondeliquency Court Corrections & Child Welfare Dependent & Neglected Child Welfare Service Records Admissions Information 16. Date of Admission-Ea5t Month, Day, Year Month, Day, Year Month, Day, Year Month and Year Month, Day, Year CentrAat State. (Date of Petition) (Date of REferral) (Date of Petition) We-teAn State 17. Source of Referral 0 Department of Public 0 0 0 Welfare Individual Parent or relative School authorities Other social agencies Court authorities Social agency Eastern State Child Protective Assoc. 18. Reason for Service-West 0 Dependent "Type of Commitment" 0 CentAat State does Dependent with Dependent not have ouch an delinquent behavior Dependent & Neglected indoamationat Neglect Neglect or improper care categorly. Abuse " (Associate with day Neglected Infomunation simitar Abandoned care placement) If recommitted mentally to EasteAn State & Immoral environment " (Associate with full Deficient (MD): East Centraa State'd Custody time placement) Dependent & MD "Reason dot SeAvice" Physical handicap Employment of mo Dependent & Neglected & MD is categorized undeA Feeble minded Illness of Mo Neglected & MD "Type of Commitment" Fa unable to care for "Parents Characteristics" S "Patents' Chatac- children Abandonment tecistics." Marital discord affecting Physical neglect child Medical neglect Mo not married to Fa of Moral neglect children Rejecting child Child in conflict with law Physical incapacity Other behavior problems Mental Relationship problems In mental institution Illness of child In penal institution Physical handicap of child Proceszing and Dec-ison Making PeAtinent to Piacement/and DVipoaition 19. Diagnostic Services Psychological-Yes/No Provided 0 Refer to Diagnostic 0 0 Psychiatric-Yes/No Center Psychiatric Studies Medical Examination Social Studies. 20. Type of Placement Public Agency or No Detention 00 0 Prior to Decree- Institution Relative home EasteAn State- Voluntary agency or Hospital Adoptions Institution Department of Public Independently by parent Welfare (social CaAe Pending or relative agency-public) DVisposition-Ea6teAn Independently by attorney, Eastern State Child State-Urban County doctor or minister Protective Assoc. Independently by other (social agencyindividual private) No placement-Child in Detention center home No placement-Child not in home Unk., N.R. 21. Investigation of Court staff 0 O 0 0 Petition Made for Public child caring Court By: agency Voluntary child caring agency State Department of Public Welfare Other None or not required Unk., or N.R. 18

Chart III (cont.) _________ astern state East Central state west Central state Western State Urban County Variables Adoptions Nondeliquency Court orrections & Child Welfare Dependent & Neglected Child Welfare Service Records 22. Relationship of Not related 0 G 0 0 Petitioner to Child Own parent Step parent Grand parent Aunt-Uncle Other relative Unk. 23. Consent to Adoption Mo. 0 0 0 0 By: Fa. Mo. & Fa. Guardian Agency Other individual Unk. 24. Child's Consent to Yes/No 0 0 0 Adoption Given Not applicable-children under 12 Unk. 25. Type of Commitment- Consent only 0 0 Termination of parental 0 West Centftae State Abandonment found by rights Type o0 TeAmination o6 the court Dependent and Neglect Pitentae Rights Combination of above Neglected Decree of voluntary If recommitted mentally relinquishment deficient (MD): Decree of abandonment Dependent and MD Combination of above Neglected and MD Unk. Dependent and Neglected and MD 0 26. Supervising agency 0 0 0 Thirty public, private and sectarian agencies are listed, e.g., Family Service Agencies, Children's Homes, etc. 27. Eligibility for Medical 0 0 0 0 Child welfare service program Care only, Public assistance program, not eligible 28. Payment Plan 0 O 0 0l |O| Requirements-given in dollars'for board & room, clothing, personal maintenance, school supplies. Source of payment-eight sources listed e.g., earnings and contributions, Social Security, various public fund sources. 29. Type of Placement or Adoptive home Dismissed or Discharged Free foster home Free home Adopted by foster parents Disposition Court Supervision Paid foster home Work wage Adopted by relatives Commitment: Foster placement planning Adoptive home Agency X facilitated adoption Custody of individual Foster care supervision Relative's home Home of parents Custody of relative for non state ward Boarding home Home of relatives To Department of Pre-adoptive study of I Own home County foster home Public Welfare ch Schild Scool Voluntary agency foster home To delinquent Adoptive placement Self support (extensive list of names institution planning Hospital, Sanitorium of voluntary agencies) Custody of parent Supervision of child in Boarding home Day care foster home To private agency adoptive home i Boarding school Day care center Adoptive home Institutions: Public, Homemaker Other: Work or wage home i Private, for Blind, for Voluntary agency maternity Continued for Child Relatives home I Deaf, for Crippled home (extensive list of Protective Independent Living Children. names of voluntary Association Report Arrangement Inactive: agencies) Child Protective Services to juvenile Armed Forces; Institutions (extensive list Association to unmarried Mo. Correctional institution; of names of private and Supervise Day care center placement Mental hospital; public institutions) Referred to public Day care home placement Self support; mental health center Child care screening No supervision necessary; To live with Mo. Child case intake Moved out of state; To live with relative Institutions: The names Whereabouts unknown; or friends East Central State's Foster care-no supervision Visitation privilege reception centers, necessary; granted institutions, camps, State residential Other cottages and halfway treatment center. houses are recorded. Runaway After care-group homes After care services to The names of the state children in own home institutions are recorded. No private institutions (16 state institutions are listed are named.) Client status-Inactive, Placement in private On leave of absence, institution can be Truant from training determined by scanning schools. "Supervising Agency' (Category above) 19

0 co 4) 0 )d Z 0 0 -4 0 a o ro - 0 HOl T ) -4) 41 -4 I -l -4>.-l 00 0) 0) 4-1 4-1 4-44.: 0C 0 0t 4-1 4-H1 C 0 0. 0 n 0 *iO00U 1d 0 H.4H U -4 -4 Hl 0 0 00) 0 *H0 0 U i" 0H 44JH0 -4Ela-4 4 10 0) 00 T0'0H 4- 0) 1c rt -4 a0000>) 001 " -4- 1-40 0.> >.C 00 0 H-i 0 ro tf) C0004- H H C 0 Ul(At440 0)m00 0.0-4H -4 4- 10.0 -40 00 4-1 -H40o001 4-1 4iH-l4 4- 4i1-4 -1 0. ri-' 4 4-1r >0 ( > 4. 0 0 /) 0 >0 o co 0 4 0 > 00 0c 0 o.0) r n4.a 00 4J = 0. 0- u 4 4 4 a 0 4.-40 0 -444- Q) 4?.000 001-4'''00)4O'O - 0)4-) 40.00000o 0 1- 4. 0.0 0 0 4- 44 00 0 H( 0) (0 i00 0 4- 40 - 00 0 0 0) - -40)40)0. 440 0 H P44 00H) 0 B - 00)0 -o w'.4~I 0) 0.00.4 -4 0C4 44Ji - HI0)c-0000r 0 03C3 C 044 1 0 04 (1) o 04-14- c 0 0 OH 0 1-4.- I n 1 1 > i- 1t1 040 0.- H>I H 0 <4u4) 0 M 0 0 4J W r Mi (u3 < 3 C d Q1 0 0 0 w 4 0H 000$ 00.0000)UC> T34-~!4-U 040000TcI 4C 3 0m0 O.. 04- 0 -4 (Q 0 Hi 044.00- )Hi 00m)'0000000i 0 > >0)4-40 0) t 4) 0i-4 440) 000i 04- C (344 00-I H 00 0 0000i000 003 0000 ) 00 A1< 044 0004 Cl - 0. 00H 0rd4) 0) )H0H- HH4J (U-4 r )HHH-4 OS 0.H4-4'0000.0)44 a) 0)4-4H o > 4-~i H0.0e i.44 >4 C0 HO0)U 0044 0H0N -. 0 4- 404'-4' 30'04'0*'000 >0000 0 4J OH 0'H 4 0 4 H n) O0)OHH0)HI >4H>4HO mt~~-r(r >ritr-c 1- T o 3 ( r44 -i 4i 44 - H H H 0 00 O cc O a ) 4 Cr-4d oJ Q) ~ v v aa000.0.04-4I.0 0.0 ( 2 ) 4-1 P 0 a- 1 ( c dr )r 04) 0 0O0 0 0: 0 4o Z 4w 4U-4 0 0 <~ 4-) 3>'00uo n m0 a)41 C H - 44 0) 0 ) C4-4'0 0) I 0) 44'0 Q> (I'Oco 01 P ( o = ta l W 10 0) 44 4-0) 0) 44 M. 0 0 0 00 0 C4 04 ) C0.4 4 44 4 H ) 04'0 0d0 44 4r-4 0 0) T4-0 0001 4) 0 10 0)OCS0 0 0 ) 00-~ 0 00 000 - 4. H 4g~40).0 H0.4lU'00 >00. 00'0-'00 0 4 nl Q)0H 0H 0 0PP W 4 la Cd i — 44 0. 0) 440440)0)0.00..044> I 0 4 -404 0.0 ~000 0H' 4-40 00 4-40 (060'0 >44 t = 00 04 O' 0 > 044H'0'i 0'0 OH H 0i) to m "> 4Ju 3 C 00 4d 4'0'0 0)0 4 H'004- c 0) 4 4.00)0)o 0 444 44000'.a0440)4 44d 04-4 0 OH 4- O0 0 0 iHa-) 0 0 0 d 0l-O O C 4-4d 4O 4 4 4 l-, 0 <U 4- Ai la Fe Q c.4 0)c'4-O 0 0.,I 0 00 10 +0 0'0 to0))<U4n (B6 0l U 44 -44 4H ~ 0 a0) 3 0C0 W 40) 44 1[ I 00 H 0 AI w,P0) 44 H 0 0.0 I l p ~ HW 0. 0044 H H 4 00 44 44 0)0 V)I 0 c 00'0 44 0) 0 0 0 0 0 34-1~i 4l 0 4000 4 4 4 e $4ico c Z ~4 d..o'He tU H 0'0 0oa> i 0 0H40) 4 00 0 41) U 00).0.0.0 H.0 0 44 0.0 440 00)_ 4 044 * 0'0 0)44 44 44 4444 44 0. OH 00) 000 0'0 0d 9) Q) k sc ~ ~OC LI 0 0 0 0- d ~0 a 4-4 d 4 0 0 0 0 H OH 0) O t ~t-^'~l~-OOurl0) 0)0'4 0 H40) 4);4.00 4 44ri0) 44'0H HH4-4>4 0) 0 0).<4 4i40)00g400 44' f440)0 H 00 00 0)0 di Mu 4__________ 0040) 0000 044 0u 0 >H;0I 0o 004 H 4 4 Co P 4-14 c; f 0004 H O_.O H 004H 44 0 0)0 44 0 *44 d 44H 044' 440 00 00'44> 440.)04 ) H444 0r. 000) 00000. 44 0-l4 0 0^ CO i l 0 0 0) Q)" 00 00) 43400 0 +'0aS0 00) OH 0 1-Li 0)0 0.0)000 o H 44 HJ 0 3 0>4J.-0 0.4J (U( o H 0.00 0.0'0 0O- $ 5 Ai 20

Child Welfare Agency Formats and Variables Once the charts were prepared we considered the ways in which we could attempt a comparative analysis across states. An examination of the charts indicated that by merely counting the number of variables and the categories within a variable we could ascertain the variation in the formats among the states. The results of this tally are presented in Table 1.1 on page 22. The table shows that all of the states with the exception of West Central State collect just a bit more information about clients than about any other dimension. West Central State collects the same amount of information about clients as it does about organizational variables pertaining to client careers. Very little information is collected about the parents or families of the client (see Chart II). This is quite surprising considering that a child's parents are usually instrumental in his becoming a client in a child welfare program. There is marked variation among the states in the amount of information that is collected pertinent to placements and disposition. The categories of ihis variable are inflated by the number of agencies and institutions where a child can be placed or supervised. For example, in East Central State 13 of the 35 categories refer to public institutions; in West Central State 46 of the 78 categories refer to public and private agencies and institutions; in Western State 87 of the 115 categories refer to institutions, foster homes, and maternity homes. If we take these figures into account and consider that East Central State combines its child welfare and corrections data, then West Central State turns out to have the most comprehensive and differentiated child welfare information collection system. On the other hand the differences in the number of placements in each state are probably a reflection of the interorganizational relations between the state department of social welfare and other social service agencies and institutions. Although the above table does help us to comprehend some of the differences in the formats among the states it doesn't tell the whole storyo The formats can also be described by ascertaining the number and percent of the variables that appear in each state formato Of the 32 variables in Charts I, II, and III, only 8 (25%) variables appear in the formats of all four stateso The variables and their code numbers in the chart are: (1) Case Name; (2) Case Number; (3) Birthdate or Age; (6) Sex; (7) Race; (16) Date of Admission; (29) Type of Disposition or Placement; and (32) Location of Residence (County). Three (9%) additional variables appear in the formats of three stateso The variables and their code numbers in the chart are: (9) Age at Admission; (10) Living Arrangements; and (31) Termination of or Change in Client Career. Eight (25%) other variables appear in the formats of only two states and 11 (35%) variables appear in the formats of one stateo All but two of these variables are from Chart III which pertains to client careerso An examination of Eastern State's Urban County format reveals that two variables do not appear in any state formatSource of Referral and Siblings. These figures indicate that there is considerable diversity among the states in their child welfare formats. Thus far, 21

CH a)s1H H I?4 r4C r4A 0) r. 0 r a) *0 - O r-I v I I I o 1 rt E-i Co 5~D ~~ LC\ CO H0 Hr * H - H *-p0U) 0 C r -T) H H @.H H U) d z a) Ut) -rd c d 0i r N *A rr l0 r ^H EI I HO O - r-l Ss U3 C) k z c n ^n o (X -CU OH co C O CC) Q\ H H | O0 _z r -- H H ct O.~ O O o qQ c o C o H i.. r-, Lf) 0: cii P p C Cd Co 0 0 H rrd H 0 a) H I C 4-) rd ro O V.H O O -rl d.0 -~ r O Q/I +) E3 QP O -r1 *r1 r-'ii ~^ _d H CO 4-) Cd cn3 w *rl C1 *1 (Q 4 a, o o 0 c f 4O ) a, o r-1 Q. *H *H OH M 3 CO O> 0)-^) (l 4 OJ ) -q cd> O r- ) o o k4 Pi 0, O a _: 0 + o rW -1 - d Cd t a co* CH * 0 (U w V- *S 0 0 q 1: V O! k ~ O P PH ^^ E- -- r+rO H ~ rO P -H H!h U F^ V *r1H 2 gd CQ. O o V H H V V E~ c6 22 P^ CH r-1 ^~~~~2

we have relied on quantitative methods of comparisono We will now make a few qualitative comparisons. Earlier we indicated that the state departments of social welfare were asked to provide us with the information they routinely collect about their child welfare clients. The column labels underneath the pseudonyms for the states identify the titles on the information formats which the states were able to provide to uso In other words, with respect to child welfare programs Eastern State has an electronic data processing system only for its adoption clients. (Eastern State's Urban Ccunty is not an object of analysis. It was included in the charts merely to provide us with a contrasting format.) The data collection system of the other states particularly West Central State is more comprehensive in that it includes information about adoptions as well as other child welfare programs. On the other hand Eastern State's variables 20 to 25 show the extent of differentiation that is possible in the information processing system of one child welfare program, io e, Eastern State collects more and different kinds of information about adoptions than any other state. East Central State's format must be examined carefully because its information processing system combines child welfare and corrections data. Although some of the categories within certain variables obviously apply only to one program, at times it is difficult to determine whether an item pertains to the dependent population, to the delinquent population, or to both populations, e. go, see variables 29 and 30. Another problem with East Central State's format is that certain categories of a variable are not mutually exclusive (e.g., see variable 31) and there are no instructional booklets for the resolution of coding dilemmaso In contrast the other states have conceptually clear categories and instructional booklets to resolve coding dilemmaso Although the charts are useful for certain comparative analyses, they do not reveal certain features of the information formats, The official report form and the code books must also be described in order to convey the character of the informational formats, This is especially true of Western State's format. One half of that state's report form pertains to the determination of the client's financial needs and to the authorization of payments in behalf of the client. Eastern State's and West Central State's formats are well organized and the codes correspond to the computer tape layouto On the other hand, East Central State's codes are unclear, poorly organized, redundant, and they do not correspond to the computer tape layout. Subsequently, in our report of the analysis of the career patterns for this state we shall present some of the findings which indicate that these code inadequacies result in methodological and processing errors. Up to this point we have been addressing ourselves to some of the holistic features of the formats, io e, we have attempted to characterize and compare the formats per seo In the section below we will examine and compare some of the variables and the categories within variables, 23

In Chart I it is interesting to note that there are variations among the states not only in terms of whether information is collected about a particular variable, but also in the selection of categories within a variable. East Central State specifies the foreign birthplaces (4) of its clients, while Eastern State specifies the county of a client born in its state. Two states are concerned about the child's birth status (5) but only one of the states is concerned with information about paternity. All of the states collect information about race (7), but they vary in the specificity of their discriminations between racial groups. West Central State specifies the racial mix of clients who are the children of parents with different racial origins, whereas Eastern State classifies such clients as "Other. " In view of the fact that religion (8) is still a salient factor in the adoptive process and that sectarian agencies support many child welfare programs, it is surprising that only two states collect information about religion. The salience of religion in West Central State is reflected in the distinction it makes between Lutherans and Other Protestants. Only West Central State does not collect information about the child's living arrangements (10) prior to the first contact with the agency or during organizational processing. On the other hand it is the only state that collects information about the child's physical, psychological, and social problems which are ofter decisive determinants in making child welfare placements. Earlier we indicated that a child's parents are usually instrumental in his becoming a client in a child welfare program. Yet, Chart II shows that only West Central State collects information about the factors which led to the child's acquisition of client status in a child welfare program (see variable 14). In Chart III the first item of significance is the fact that none of the states collect information about the source of referral (17). Consequently, none of the states can identify who is instrumental in initiating client careers in child welfare programs, i. e., the deviance nominators or the audience that is privy to child welfare problems cannot be studied systematically. Furthermore, the differences in the deviance nominators cannot be studied to determine if they have differential affects on other phases of client careers. In the two states that collect information about the Reason for Service (18) there are striking differences in the categories within the variables. In West Central State the reason for service is a function of parental deviance or misfortune; in East Central State it is also a function of the child's deviance or misfortune. West Central State identifies the type of neglect experienced by the client, i.e., physical, medical, moral, whereas East Central State s categories seem to convey the degree of neglect, io e., neglect associated with day care placement or full time placement. East Central State's categories are more specific whereas West Central State's categories are more generic, eo g., some of East Central's categories are examples of or can be subsumed by the categories in West Central State's format. The juxtaposition of the categories in these two states helps us to appreciate the role that judgment plays in the definition and labelling of deviance, e.g., what West Central category would the reader use to classify East Central's "employment of mother" or 24

"Father Unable to Care for Children?" We will skip variables 19 to 28 because most of these items were discussed in previous sections of the reporto The categories within the variables Type of Disposition or Placement (29) and Termination of or Change in Client Career (31) are quite similar across all formats. Only two states record information about the tenure of the client's career (30). However, the information can be derived from the other two states by manipulating the dates of referral and the dates of disposition or closing. This completes our analysis of the Child Welfare Charts, We will defer our discussion of the meaning and significance of the findings until we complete the presentation and discussion of the Corrections Charts IV to VI. Correctional Agency Formats and Variables The juvenile correctional programs and formats were selected for separate study and analysis because, with the exception of East Central State, child welfare programs are separately administered-one for dependent and neglected children and one for juvenile delinquents and/or offenders. Marked differences were quickly observed between these two programs when the information formats and code books were examined. Therefore it was decided to analyze them separately. In preparing the Corrections Charts a certain set of variables were excluded, namely, the variables from Central State's probation and institutional release formats that pertain to the organization's assessment of the client's personality and social adjustment. The specific variables that were excluded are: (1) measures of the clientVs feelings toward himself, toward adults and toward others, ego, accepting, indifferent, rejecting; (2) measures of whether the client has accepted responsibility for his delinquent activity, e.g., accepts entirely, mostly, only slightly, doesn't accept, etc.; (3) measures of the client's feelings toward his parents, e.go warm, indifferent, hostile; and (4) measures of the client's hostility, emotional stability, judgment, and ability to plan for the futureo This set of variables does not appear in the format of any other stateo Consequently, there was little to be gained from its inclusion in the charts which were prepared to facilitate comparisons among the states. Thus, with the above exception, Charts IV to VI contain all of the variables that appear in the state formats. 25

' H E 440 H)E PL 3 uc.~rc 0 Ha a) 4'*^~~~~~~~~~~'H*i-l~~~~~~~~~ 0HU( o H OH HO H - a) O'r ) OH. 0 HHEc' H 0Q) r I a H4 3 O-4 H H C HUH O H H O4'4 - 44 H > 4 MU ( -i 44 ")~^2 o o H H 4 oH 4405'HO OH h pr u ^ o H o HoH -! O4 Ha 000'H 04 H(- HBHOH 00 C ) U -l *' H H t H Ht HHH14L'HHa OH) 0HH -0 HH O S Or, U'Hti U'w'o C tc 0o u C' O, 0W H co - e 1^ ^ ^ 0 ^ 1 az ~a \ Q) o ~ ~ co a) Hz HHHHHH'HHH c o HHO HH oIC 00041* w ^i Q) a, 41. ) 4 -' HHH'~oc-o-j04 H - I 41 CO UP) o En cHH 00H' u H' y og H O >-1*'-'O H H U'44 H HCCU C!'-!*H H Hie iO C -l-'S 4- i4 x H t'C- ). 0 -) o^'H'H ca'4 4)4 Hr Q) 4 Q C C ~ coc-,H'HO)- H)co) 0 4) *H W4 3 4 4)4Il 04 f H ^ ^ HO 4 ^ ^ ^^s0 H'HO H)n I44 UHO+^ 0"> H+c -K^C M S r y 0 HOD HH (HU4-4 4Jc4+4)Q oS>Q 44-l^ OH~ < N S ke p-d Nl 45HU HH3 OO OIi+ 00 O -+-) -^SH S ^..^ S H JO 040 HO H 0 HU co40 44O 0 H 0 ~ w OH HHHO'O' 0 Ha Q L x (0 Z a)Fga) 0 HH 4O H H H H H O H O H4-4 H H, 04 04 Hco04 H 44H 0 Ha ~ ~ 6 a E Z 40 a _, Q) F5 -H 1-Z 0 0 04' H H HO H-H H, Od HOD, Ou HH HOl HIa C, O d CZ OH- H 0co HO 000 H O 0 Ho HHU HP H HHH o CU cco HHHH L) -' 0) H 0, ) (,, 00,0 coH 4 H - Ca HHH 0'a)HO 0 4- 41 -I0 H 0o H1 H - o' O H H 0 U H HO H H o- H 0a) 1 IU a H -C _ 0H! c =Q4 ic42 ru:3 1:11 M41 4J >,;j LQ ul co 4 -U t. J_ I I I ~ I!~w n D ~~~~-LL U ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~ E Od ri C EO h~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~d. I $,.,,, U ~~~~~~~~~~~~~~~o (U~~~~~lcl~~ O akoo ~c d uI ah OE~ caXC Hs~FiG~ p~F HHO~ 33~n~r G C00m0 6~0 e 26

Chart IV (cont.) __________________^West Central State _ Admissions & Terminations Variables Corrections Admissions Corrections Releases Field Supervision 33. Case Number / / / 34. Case Name / / / 35. Birthplace-CitizenJh6p Citizen by birth, 0 0 WeIst CentAat State Naturalized, Alien, Unk. 36. Birthdate/Age Month, Day, Year Month, Day, Year Month, Day, Year 37. Sex Male/female Male/female Male/female 38. Race White, Negro, American White, Negro, American White, Negro, American Indian, Mexican and Indian, Mexican and Indian, Mexican and Other Other Other 39. Religion Roman Catholic, Jewish, 0 0 Lutheran, Methodist, Episcopalian, Baptist, Presbyterian, Other Protestant, None, Other 40. Age at Time of: / Inap Not given but can be Admission, Referral, easily determined or Commitment Closing, Release Inap / Not given but can be or Disposition easily determined 41. Living Arrangements- Natural parents 0 Natural parents Mo & stepfa Mo & stepfa Fa & stepmo Fa & stepmo Mo only Mo only Fa only Fa only Adoptive parents Adoptive parents Relatives, Friends Relatives, Friends Foster home, Group Foster home, Group home, Work home home, Work home Maternity home Maternity home Independent living Independent living arrangements arrangements Residential institution, Residential institution, Halfway house, Treatment Halfway house, Treatment institution, Correctional institution, Correctional institution, Jail, institution, Jail, Workhouse, Other, Unk. Workhouse, Other, Unk. 42. Location and Length Metro 0 Metro of Residence-Aft Urban Urban Stoate provide Rural non-farm Rural non-farm the name of the Rural farm Rural farm county. GkosL Transient Transient ditstinctions between urban and ruratl can be determined by county of Leesidence EasteAn ls the only state that has infomaotion on length of rLeidence 27

o 00N 0 m o o o.0 4 El. 1 a) -O~~~~~~~0 ~ |0S8 O,o o 00 0 I0I ~ o L o o o S " I 0 4 9 c O I CO co co i. 0, 0 o'' 0'................................. T4 ~, I.,_, -4 ~1 t o 4 o o OH *00 o 000O0 u 0 000- 0 o C OV)s 000o WO O0 0 0 04 00 Op 00 O -O00 -0 0 0 4 ~ 0 E0c ~ <'0 ~ ~i - o o o ~o~~0o 00, a) ( 0 00 E=. 0 0 0 _0 t t E o 0 OH OH', o I Ho H 0 0.o 0 00 2 o.g.E~ oJ1 s3 (1) r-l U ( t ~ U L O 00P oO C 0O 0.: 1 28 co H 0000 00~~~~~~~~~~~~~: 00> S- Z) I 0 ~~4 > Q0 2 8 -0a — 0-

Chart IV (cont.) West Central State Admissions & Terminations' Variables Corrections Admissions Corrections Releases Field Supervision 43. Intelligence Quotient Superior 0 0 Bright normal Average Dull normal Borderline Defective Untestable 44. Achievement Test 0 0 0 45. Educational Attainment Highest academic grade 0 Highest academic grade and School Adjustment completed is recorded completed is recorded A. Years of School Completed B. Grade Placement 0 0 0 in Relation to Age C. School Adjustment 0 0 0 46. School and/or Not employed-not in school Inap School Status Employment Status Academic or technical Not employed-in school Full time Employed-in school Part time (At time of apprehension) Home bound Work study Evening only Correspondence only Not in school Reason For Dropout: Expelled, Not admitted, Mutual agreement, Refused to start Economic Physical incapacity Employment Status Umemploy Full time Part time Seasonal, Irregular, Unk. 47. Occupational Professional 0 Professional Classification Managerial Clerical & Sales Semi-Professional Agricultural Technical Building Trades Clerical Industrial Agricultural Fishing-Hunting-Guiding Military Service-Domestic Retired Other Housewife Other 29

o40 0o 0.. ^0 rX o.n a0 a) o 0,a U o W4 ( o O0 0 0 r-i o o + o o o 1. Co ~ oj._,. _ -0. 0) 000 0 0 t 41 p.v e0 o 0.) 0 0 o tfl ^ 60 4 0 0 0 00 0, 4 o o 3 o o i o o 4o 0 04 I - 0^ 0.-'t0)-00O00 000'.0 00.0 0 [OC.>'0T.0-f&0 0.0. En00 0 ED ~ 0 o o 32 oP r * > 0)'0 0 04 co 0.H 0' V ~ D 0 O H 0 00 0 0,.) cu> 3 =~E u0 0 00 Z X E E E.11 0 0 0 0 0.0 0.0 0 0. 0. H 4 0 0> 0 0 0)0.0 c o4C 14 -l O 0.-0'0(f0 00 C0. 00'00 ).) 0.'0 0 c a) 0 0 0). 0. n JU s i-1 3 fi'^ >^ ^ T)(4

Chart IV (cont.) L __________________________ ___ ______West Central State Admissions & Terminations Variables Corrections Admissions Corrections Releases Field Supervision 48. Occupational Skilled 0 Skilled Skill Level Semi skilled Semi skilled Unskilled Unskilled 49. Marital Status Single 0 Single & Adjustment Married Married Annulled Annulled Legal Separation Legal Separation Non-legal Separation Non-legal Separation Divorced Divorced Widowed Widowed Non-legal Association Non-legal Association 50. Dependency Number of dependents Type and Number of: 0 Obligations Dependents Wife Children Parents Siblings Other 51. Military Service 0 Less than 1 year 0 1-2 years 3-4 years 5-10 years Honorable General Other than honorable Dishonorable 52. Physical Problems Remedial 0 0 Chronic Both None 53. Drug and Alcohol Drugs 0 0 Problem Alcohol Both None, Unk. 51

"^ a 41 aI 14444-4~~~~~~~~~ ~-4 i o o-,I o 4 44 4 34 j4 3 3 4-4 44 444 4444 0U () u 4 E 00 44-4 44 W W S ^ 444 |.o = 3 u o v ~ W a ~ 3 2 (u 44 ^ 44-4tl044'5.~- 0'a 00000 4 444444a o 3 34 4 c I1 X U) 0 1. ) 44 44 -i-.) o.; U ) oo _ _. 44 40 0-o 0 C (.40 4 > 44-4 44 0 + 4-4 00 f tQ ~ *~ 41 LLJ W r 4J > _U._. 4. 4 4.O 04 44 0) 0 O- 1 3- 44 <4 *S' C 0 0 0 0 S- 4-4 44 -4 > 4- 444 o Id- 0,4 I44 44 444.- 40 4 E 0 4 4 44 4 * o 44 044 -4400 0 (-3 4-~ 4? H b-4 44P 44 3 3 44 - 4 44 d- 4i 444 044 cn333CT4 o0 44 VO.c-c4o 4 4 3 44 O rd 04 - 4 4 * 4 44 * A Uafnt i = o o 44 Ao0 S; 4 0) 4 0 0 4-t _ d -4- Oq C=W 4 44 004 4 ) 44 044 04 - 4 (U " E4 04 O C 4O, A0 44 I a Q a o z 44 4 1 s o Ca C o U En ^ e u I I 0 W W l t 44 14-4 P4 44)4'444C0> 44 oP 44 3- ~o 4 ) 52

Chart V (cont.) 11____________________________ __ft West Central State Admissions & Terminations Variables I Corrections Admissions Corrections Releases Field Supervision 54. Marital Status of Parents 55. Sources of Support 0 0 Self and Family Income Public Assistance Parents Relatives Friends Spouse &/or children Unemployment Compensation Other 56. Nativity of Parents 0 0 O 57. Supervision of Youth 0 0 0 by Father and Mother 58. Affection by Mother 0 0 0 and Father 59. Cohesiveness of 0 0 0 Family 55

Chart VI. Information Formats about The Careers of Juvenile Corrections Clients tn FoUr State Department* of Social Welfare -Et*n Sta East Cntral Stte Central Stated, Variables banle coari Re-oods C fa d abiooZ A eaaitss Coaotoa U vl C el d t jlf 2abo *IerVuios Isfkitffanl lllasa ftieaba-.aoon Intobk Information _ ____ —__ 60. Date of Ade1"aion, Mobth, Day, Yo.r Oouatha, "ay. Yeer Mo0, Dy, m ath, Year -h, Day, Ye-r Month, Year Inap f orr l 61. Purpose and Type of 0 PurpoX: TV IMip is Admission Training Under supervision:, coitntmei Study Oriiamal comaltamnl am com tmant-prvlop i ly Detention Juvenile probation committed & discharged Type!50~ laanm-aaaa~Ttra lte RIeturn fro Aftercaren Court Other Raturn for replacemnt Voluntary or radical care Other Infornm lon re: ak nreturns & trans fers betwann speclfic I institutions ia! also recorded 62. Reason for Service The speclfic. offenase Offenses Leading to Neglect or taiproper care All cllents are coded Major current delinquent Inap Inap committed by the Admission: assoclated with day care "Juvenile Doalinquency acts prior to adamissiaon client isa recorded. Offenaes aalnst: placement Sane (The codes contain a Property, persons, Neglect associated with Aggravated assault detailed & an authority. full time placement Sa slsconduct extensive list of Sex offence Eaployent of no Auto theft of fenssa.) Illness of so Burglary Fa unable to care for Vandalism children Other theft Marital discord affecting Uncontrolled child Run away Mo not marriled to Fa of Drinking children Carelesaness or Children in conflict with mischief law Other behavior problems Illness of child Physical handicap Other 63. Referred by Law enforcent t Court Court Courc Court Inap loo agency School Social agency Probation Parent/relative Other court 64. Copanions Involved in 0 0 0 0 None 0 0 Deliaque.Jcy before One Admisaion Two or oare 65. eadission 0 N, Readmission New, Readmission O 0 I taor reasan for Inap Inap Informatian Date of ssadalsion Xeturn from Aftercare as o f re admission toI b n pble Nota readmit, In panstso hm Parole violation in foater hon e isbehavior Poor school adjustmaent Rearrest, unk. Run away Elapsed time oIw violation Between readmit & previous admission Also see Type and Purpose of admission above for readnit informtion Criina%, Corn tional jPrior Police Contacts Clinioal Hitory O. 1-3, 4, 7 and over, 66. Criminal-History 0 0 Correctional history not Number of Prior Felaoy 0 0 0 sncordmd as suchk, t Coenictions eays La other Lstitutoi a Type of Prior Felony - dotemid by Convitions 3w latin ra n Hdoneaio Sate Qodss. The poaltl i that Criminal homicide oam -0 to het for the ape deta. Other sex offense Robbery Aggravated assault Burglary Larceny Auto theft 67. Correctional History Number of Prior Delinquenc Nomber of Court Appearances 0 0 Prior Official Probation 0 0 Referrals-.rience Current and previous years Nuber of-times on probation Number of prior placements Department of Public Welfare in institutions for: Co aty welfar e Dependents Private child welfare Truants fParent/relativeo Delinquents JWenaile court worker I Othr __ __

Chart VI (cont.) West Central State Admissions & Terminations Variables Corrections Admissions Corrections Releases Field Supervision Intake Information 60. Date of Admission, Month, Day, Year Month, Day, Year Month, Day, Year Referral 61. Purpose and Type of Type: Inap Juvenile Admission New court commitment Youth Transfer from X institution or X hospital, Return from Parole: Return from medical parole For violation of technical rules With new court sentence or commitment New offence For replacement Admitted from probation for the above five reasons 62. Reason for Service The offense committed The offense committed 0 by the client is by the client is recorded. The codes recorded. The codes contain a detailed contain a detailed & an extensive list & an extensive list of offenses. of offenses. 63. Referred by Court Inap Inap 64. Companions Involved in 0 0 0 Delinquency before Admission 65. Readmission See Type and Inap 0 Information purpose of Admission above for readmission information Criminal, Correctional & Clinical History 66. Criminal History Number of prior convictions 0 0 charged for each offense. Number of prior felony convictions. 67. Correctional History Juvenile Number of paroles current 0 Youth, Adult sentence or commitment. City, County State record in other state Court Status at Time of Arrest None Juvenile court probation District Court probation Municipal court probation Reception center probation Parole Escape Fugitive _____________ 55

Chart VI (cont.) ________________________ Eastern State East Central State I Central State Acsdmeions to Juvenile'Probation a Parole Variables Juvenile Court Reoords Training School Admissions Corrections & Child Welfare Admissions to Probation Institutionsa InaittilonaI Rleae Termination 68. Clinical History 0000 Agency Experience in Re: 0 0 To Delinquency State Department of Public Welfare Private child welfare Child guidance lcobinations of above Information Processing and Decision Making Pertinent to Type of Placement or Disposition 69. Diagnostic Services Psychiatric 0 0 0 0 0 Psychological Social, Medical Indicated & provided Indicated but not available Not indicated 70. Care Pending No detention 0 0 0 0 0 0 Disposition Jail-Police station Detention home Foster family 71. Type of Disposition Date of disposition The na of the institution If applicable the nae of Type of Disposition The nae of the institution The nae of the institution The nae of the institutior or Placement Waived to criminal in which the client is the institution in which Withheld in which the client is from which the client was from which the client was court placed is recorded, or the client is placed is Imposed & stayed placed is recorded. released is recorded. released is recorded. Complaint not substantiated- foster home recorded Juvenile supervision dismissed Or foster care, Sex deviate Complaint substantiated Aftercare, Halfway house Intra state No Transfer of legal Status: Truant from Juvenile custody custody: institution, Inaqtive, Dismissed-warned, Leave of absence, Child Adjusted, Counseled case screening, Child Held open without further case intake, Preaction adoptive study, Probation officer to Adoptive placement planning, supervise Supervision of child in Referred to another hoe, Day care center agency for service placement, Day care home or disposition placement Runaway returned to Transfer of legal custody: Public institution for delinquents Public agency or department Private agency or institution Individual, Other. The name of the institution in which the client is placed is recorded. Experience and Adjustment in Program or Institution 72. Educational and Inap Inap 0 Inap nap Inapplicable School adjustment during Vocational Progress Excellent supervision Satisfactory Inap No progress Satisfactory Regression Unsatisfactory No evalution made 73. Work Record Inap Inap 0 Inap Inap No participation 0 Satisfactory Unsatisfactory No evaluation 74. Change in Institutional Inap Inap 0 Inap Inap No change 0 Adjustmet cDuring Better djustent Stay Worse adjustment Unk. 75. Institutional Inap Inap 0 Inapp xcellent 0 Adjustment at Time Better than average Youth Was Considered Average for Release Poorer than average Extremely poor Unk. 76. Social Workers and Inap Inap 0 Inap Inap Will return to Not likely to comit a new Cottage Counselors institution offense Prognosis for Youth Will not return to Might commit a new offense Not recorded new offense 36

Chart VI (cont.) West Central State Admissions & Terminations Variables Corrections Admissions Corrections Releases Field Supervision 68. Clinical History Previous Psychiatric Time Spent in Mental Health 0 Treatment Hospital This Sentence Hospitalization 1 month Outpatient 2-6 months Both 7-14 months Unk. 1-2 years 3-5 years 6-10 years 11-20 years > 20 years Information Processing and Decision Making Pertinent to Type of Placement or Disposition 69. Diagnostic Services Pre-Sentence psychiatric 0 0 evaluationYes, No, Unk. 70. Care Pending Disposition 0 0 0 71. Type of Disposition The institution in which the The name of the institution The name of the institution or Placement client is placed is from which the client was from which the client wap recorded. released is recorded. paroled is recorded Intensity of Supervision. Intensive Standard Minimum Fugitive I______________________________ _ __ _________ _________________ Other jurisdiction Experience and Adjustment in Program or Institution 72. Educational and Inap No participation 0 Vocational Progress Satisfactory Unsatisfactory No evaluation 73. Work Record Inap 0 0 74. Change in Institutional Inap 0 0 Adjustment during Stay 75. Institutional Adjustment Inap 0 0 at Time Youth was Considered for Release 76. Social Workers and Inap 0 0 Cottage Counselors Prognosis for Youth 37

44 - 5-4 4 4.0 o *i44'a o;; C 44 040 0 04 i3'~ ^ I *s a rQ E~~~~~~~~~~~~~~~~~l U ~~~~~~~~~~~~ > SU> ll ) - 0 *T-I ~i-l tfl *44 r4.4 0 44-4 )4-.14co4. 4444 SL, r-l ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~4,ti-1-( 00 0 0 0. 40.44044 v o 444-4 44. 4 4404444-4 0 ) 4a44.0 0 ID 0 44 444 0 4d a) U 41 0 - 444.4~~~~~~~~~~~~~~~~~~~~~~~~~~P. 44U0 44 44 &.4444. 4444 0w > 0, o i' ^.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. 1'4444 44.0 44 0 -'1~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~('a i;i3SL 44 44 a) (U4o o a0 a) 4444004444.)44..44<&~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Z ~~~~~~~~44 4-4 - 0 4 44~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~a 0 Q 0*~-3 Cl^ 0)0~~~~~~~~~~~~.,10, 00i 00~~~~ 0 ~~~~~~~~~~~~~~~~~~~~~~- 4 >, co Q~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0 4 -4..i. 40 S T3 U 4)I +-> >-) ni~~~~~~ui- u~<i r u a 44. 44 44 44-4 44~~~~~~~~) J 4 -4 -4 - to a) 4 wc 440.H 3~~~~~~~~~~~~~~~~~~~~~~~~~~~'5 -4 co44 4444 444-4 0 4-0 $L 4~~~~~~~~~~~~~4 4 44 0 44 4 44 to"I8 r- - 44.4-.444 4 T.0,-. oo ooo v e~~~~~~~~~~~~~~ 4.4-104 44 to ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~.044.04 44 44 4044 44 C)~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0 rQI 4 0 U 0! C 0) C ^ *i-i,0 T-I >~~~~~~11) -4 <0 TO*H C~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~o0~ + 4.0..4.044 4 404 4a)4C) 4 Ca )i044 4 44.0 4 4 (4 4 3 0 co4 0 44!.- 4 C04-4LI oc 4 1U 0' 44 44. ) C).44 -4 444~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~4 44U 4 O- > M- CD 44340.I., 0. 0.a)0.H0.0 0O 0 444 4. 4 Ca Ca co 44 -4 44 0> 44 -. > X " > 4-4 44-4 -4 -4 -4 -4 -4 4444 -4~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. E,:3 -4 E 444~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~4.044 44.44~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~n 4 0 044444444 04444~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~10; 0 0.44 4444~~~~~~~~~~~~~~~~~~~~~~~~~~~~) U)c o )6 44, co'e~-i~~~~~~~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ":n4I~4>ci~o~ coc>u1 )PHC j4H' 4-~ Si -'-<0.(-inlr-ltOT-l fXO~4444.0.D..444.i (0444) 4.004444i0Z 0 0 - i i 44440 44 04 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~a )co0 H 11 > ): 40 C 41 H 0 cu — I co.,4 P. 0 -14.4444444.44444.4 04co44 0 4.444444 4- w44 n "1 L4. C) 0 Q) ~: Q)044444444.4P44404 0.0.4444 04 44 4444~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~U 440440. 0m n-, o j 4i r). 4, C 4.44 0 0. 0. 0. 0. 44~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~C. 0~4444444 404 0H- ) ~ OC )a ~ H H -:0 r.0 -4 -4 -4 -4 4444~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ COc 4444P Ur I 4~ W, 4COI 0 u 44 4444 44 44 044~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. )C -I- - 0 oC o H a o o o o04444 0o 44 444444r;,~;,~:c!? (Ut~fri~u ~ citfl>< u 0ZC Ci ~~~~~~~~~~~~~~~~~~Cd CO~~~~~~~~~~~~~~~~0 4 44 44 44 44 44.0.044 -4~~~~~~~~~~~~~~~~~~.L 004rS;(~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Uu 0 44).4444 44.0 440.-4 44.44 -4440. 0 4 (0WC~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -i 0 44 0 cfo) iu f o n) < u 6.0 44.44.4444 4444 444444 44 4444 04 04~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 0. 44 C a44^ 441 44044.0440 44044 0 4400 4444 440 4.40 44 440 44440 0044 4-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. 4.44 C44 4 e g ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ iu~~~~~~~~~~~~~~~~ocoo~~~~~~~~~~~~~~~~~~~o eU- C c 4444 4404444 ~ -- 444'440 4444- i 44 444 0.~~~~~~~~~~~~~~~~~~~~4 -~~~~~~ co.4 0 4.4 0 C:.1-140.44444 to0. 44444 044444 0 0 4->*4 44 44 0 44 0 4 O-4 0 > ^T<(B~~~~~~~~~~~~~~~~~~~~~~~~l~~~~~~~ft. ^ S >^~~~~~~~~~~~~~~~~~~ -Iw0 1." I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~c ID M M) M M 0.;>, 0. 0. 4444 4444 44 44.44 44 44-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0 04 4 44 )4 O 4444.4 r. 4 4444 444 4 4 04 4 tiT iti i-a cf38-4 ( L

Chart VI (cont.) W___est Central State Admissions & Termination Variables Corrections Admissions Corrections Releases Field Supervision Privilege loss Inap 77. Penalties During Inap None, 1-25, 25 or over Institutionalization Lock ups 0-25, 25 or ovei 78. Time Lost As a Inap,1 month Inap Fugitive Or On 2-6 months Escape On This Etc. Sentence 79. Relationship With Inap 0 Inap Delinquent Peers Prior to Release 80. Use of Institution Educational or Inap 0 0 Vocational Training Cooperation with Agent 81. Was it Possible to Carry Out Probation Inap 0 n and Parole Plan 82. Tenure of Client Maximum sentence Years & months served in Length of stay in field Career Total number of years to reporting institution supervision can be be served for all this sentence or commit- determined by complex sentences ment manipulation of data. How Sentence is to be Served Concurrent Consecutive Date sentence began; Release date Release date Expiration date ___________________________ Termination or Change in Client Career 83. Type of Release of Inap Probation Discharges Termination Parole Satisfactory adjustment Continue probation & Expiration of sentence parole 21st birthday Pardon or commutation 25th birthday Nonsatisfactory adjustment Discharge Juvenile Death From juvenile commitment To Department of Public By court order Welfare District court Youth & Adult Municipal court Expiration of sentence State or federal commitment To complete sentence or to serve new sentence Return to Institution Transfer to another Violation of rules only institution New delinquent act- no ajudication- or prosecution New delinquent actadjudicated or convicted Replacement 84. County Data County of residence 0 County of residence while and commitment under supervision 39

Earlier we indicated that five state departments of social welfare were asked to provide us with the data which they routinely collect about their corrections clients. All of the states were able to provide us with such information. However, only four states are the objects of our analysis in Charts IV to VI. Western State Department of Social Welfare did not submit its materials in time for inclusion in this report, but its formats are being analyzed at the present time. The column labels underneath the pseudonyms for the states identify the titles on the information formats which the states were able to provide to use. All of the states provided us with data about institutional admission. East Central, Central, and West Central States also provided us with institutional release data. In addition, Central and West Central States provided us with probation and parole data. JUVENILE COURT DATA Only Eastern State provided us with juvenile court data. It collects data which is useful in studying early phases of client careers, i.e., probation and institutional phases of a client career are usually preceded by a juvenile court career. Eastern State also has information about the pre-court phase of the client's career in that it collects information about the Source of Referral (63). The categories within this variable refer to situations in which a child could have been launched on his client career, e.g., a school, or a social agency. Variable 71 Type of Disposition is another significant item in Eastern State's juvenile court format, because the categories within the variable identify thirteen different career paths through which a client might be processed. We can only examine three career paths in West Central and Central States: probation, institutionalization, and parole. Variable 69 Diagnostic Services is also a significant variable because it allows us to examine the affect of mental health services on client career paths while holding constant the nature of the offense. The other states have detailed and more information than Eastern State about the client's institutional and post-institutional experiences, i.e., about the later stages of a client's career. Central and West Central States also collect detailed information about the probation phase of client careers, whereas Eastern State can only indicate that a court placed a child on probation. Having examined some of the general features of the formats we will now analyze the probation, and the institutional formats separately and in more detail. PROBATION FORMATS Central State provided us with probation data for the years 1964, 1965, 1966, and 1967. In 1966 Central State changed its probation admission format to include the items identified in the opening paragraph of this section, i.e., variables that pertain to the organization's assessment of the client's personality and social adjustment. The charts only reveal some of the change in the formats. In any case it is significant to note that these new items dominate the 1966 and 1967 information format, i.e., more information is collected 40

about the new set of variables than any other set. In view of the subjective character of these measures and their questionable evidential adequacy one can only wonder about the validity and reliability of the measures. At the same time, however, it is important to note the character of these measures, and their appearance in Central State's format. These measures are evaluations of the client's behavior, and an examination of both the Child Welfare and Corrections Charts reveals that no other state collects this kind of information. A comparison of the probation admissions and probation termination formats reveals that Central State cannot systematically examine the social and sex variations in probation careers. Information about racial origins (38) and about the sex (37) of the client is collected at the time of admission, but not at the time of termination. For the sake of argument, someone could counter that it is not essential to have the information in both formats, because the client's worker or agent knows the client, or in the case of the organization it can obtain the information merely by looking through the other forms in the clientvs folder. If we were concerned about information collection and processing systems in terms of the organization's ability to conduct transactions with its clients then the above criticism would have merito However, we are not working at that level of analysis. We are concerned about the aggregate analysis of large amounts of data and the information collection and processing system at the state level. The department of social welfare at the state level does not have ready access to client case folders nor does it know the clients. Consequently, what may appear to be a duplication of information processing at the local transactional level may in fact be essential redundance for the processing of information on the aggregate levelo Thus Central State is limited in the kind of analysis it can perform with its probation data simply because it failed to duplicate information from its admissions format on its termination format. In contrast West Central's Field Supervision formats do not have such limitations because the same form is used for admissions and terminations. We will explore this problem further in a subsequent section. West Central Statets Field Supervision Formats pertain to Probation and Parole Admissions and Terminations. There are two major differences between Central State's and West Central State's formats. First, Central State collects information about the client's experience and adjustment while on probation whereas Central State does not (see variables 72, 76, 80, and 81)o Secondly, West Central State collects more and detailed information about the reasons for the termination of probation (see variables 83). These format differences raise questions about the rationale for the collection of certain information and about the meaning of these differences for client careers. We plan to explore these questions and problems more systematically in the coming year. 41

INSTITUTIONAL FORMATS Two modes of analysis will be employed in our examination of the institutional formats. We will do a separate comparison of the admissions and release formats among the states and then compare the admissions and release formats within the same state. Before proceeding with the admissions comparisons, it is essential to point out that East Central State's singular format not only contains Child Welfare and Corrections variables, but unlike the other states it also combines admissions and release dates in the same formato Thus, some care must be excercised in the analysis of that state's format. In our discussion of the Child Welfare Charts we indicated that once the charts were prepared we considered the ways in which we could attempt a comparative analysis across states. We discovered that by merely counting the number of variables and the categories within a variable we could ascertain the variation in the formats among the states. This analytic strategy was also used with the Corrections Charts and the result of this tally of the institutional formats are presented in Table 1.2 on page 43. Admissions Formats The column totals in the admissions formats in Table o1.2 show that there is marked variation among the states in the amount of information collected at the time of a client's admission to a correctional institution. West Central State collects almost twice as much information as Central and East Central States and 25 percent more information than Eastern State. An examination of the cells in the table shows that there are marked differences in the amount of information collected about the clients and about their careers. Again, West Central State collects the most information about these dimensions. In order to determine the factors that account for West Central State's dominance, its formats and the charts were reexamined. We found that four variables pertaining to the client's occupation, marital status, and dependency were only measured by West Central State (see variables 47, 48, 49, and 50 in Chart IV). Our search also revealed that the state formats were designed for adult and youthful offenders as well as juvenile offenders. Consequently, it appears that these four items which are more relevant to youthful and adult offenders, may have inflated West Central State's tally. However, even if we do not consider these four items West Central State still maintains its dominance. We finally decided to keep the items in the charts because it points to the fact that West Central State's institutional corrections program is more specialized and differentiated than the programs of the other states. Also, a preliminary analysis of the data from Central State revealed that the age range of its juvenile offenders corresponds in part to the age range of West Central State's juvenile and youthful offenders. Thus, the items were retained for a second reason. Namely, to retain format items that pertain to comparable client groups. The other item that contributes to West Central's dominance is Tenure (see 82 in Chart VI)o The items pertaining to this variable may also be more relevant 42

O - t- H -- Co^ K\ 00~~~~~~~~W r-c r\ K 0 aD C\J CjM H1\ Cr r-H co o 0 H1 *H H H - $O O_ H- CC 0 r- 0 0\ |C r-4 OH (1 H C) CC 00 CU C I 0 ~ ~ ~ ~ ~ ~ ~ ~ C C H LCN r- - H~ P H r.H - 0 ~ b H H0 0 H - J t rc oN 1\ CN r-H 0\ 0 \O |\ \D I- ()r-I r-I K\ H c6 oOr H OL( C M H V EI ~ 0 v -R NN CM C H 0 a n;*H r;'^ 0'0 ^ o.- r,c' P U -P rY r.rbl0 r-I D r- - ~~CC/3~~~~ +203 $=!~~- C\" Z P *Ha) C CM ON CM rc Z +2H - U O H 0 CO C\J H 0 cxi d~ -PH a Hd? H-j 4 O O C ci H H H V R Ca) 4). I C C C, CMq E-10) LCN CM HVa4 H H H EMoO o ~ O.-p 0 t O H w~ ~. rC o P- CC *. o o o - - o O ~ ~.H H C o - ~ C-i Hd -t 4 H ~ O H c 4 OH O H H E-^ 0 ^ 0a) n^, CM cc -It 0 4)+ 0 Cr C\J CC- C H H H Q) \p- -'0 CC a) H u-i-i'C" CMG.0 CC-i H1 Hr-1 CM C ad rH C III PI CCa C *Ha)C 0 C 4+) 4- 4+) 0 *HO Da) *H P- *H* o Ca) r-+ *Hr-+ a PC)* CC 4 -a cxi x i 4 CCUa) CP O Hd C cx P-C C 0 Pi a)+4 0 C C=1 0 ^^ >a IOCa CC! * 4-) C *HCa *HCa C cxiC0 cxH~iC 0 co )S- o 0 45

to nonjuvenile offenders, but the code books do not make any distinction about to whom these items apply. If all the "questionable" items (nine in all) are subtracted from West Central State's tally, it still turns out to have a more differentiated and comprehensive information collection system than any other state. Table 1.2 also shows that three states collect a negligible amount of information about the client's parents and that West Central State does not collect such information (see variables 54-59 in Chart V). Although there is no variation among the states in the amount of information that is collected pertinent to placements and dispositions, there is marked variation in the categories within this variable. The categories refer to the different institutions to which a client can be sent. Thus the categories reveal the number of placement options available to the juvenile courts and the number of interorganizational relations between the state department of social welfare and other social service agencies. Although Table 1.2 does help us comprehend some of the differences in the admissions formats it doesn't tell the whole story. The formats can also be described in terms of the number and proportion of the variables that appear in each state format. Of the 34 variables in Charts IV to VI pertinent to admissions only 10 (29%) appear in the formats of all the states (see variables, 553, 34, 36-38, 60, 62, 64, 71, and 84). Seven (20%o) additional variables appear in the formats of three states (see variables 35, 39, 41, 435, 45, 61, and 67). Four (12%) other variables appear in the formats of two states and 13 (539%) variables appear in the formats of only one state. In other words approximately 50 percent of the variables appear in the admissions formats of at least three states, (In contrast only 34 percent of the variables appear in the Child Welfare formats of three states. We will have more to say about the differences between Child Welfare and Corrections in a subsequent section.) Up to this point we have been addressing ourselves to some of the holistic features of the admissions formats, i.e., we have attempted to characterize and compare formats per se. We will now briefly examine some of the variables and the categories within variables. In Chart IV it is interesting to note that the sex of the clients in Central State can only be determined by obtaining the name of the institution to which the client was admitted (see variable 37). Three out of four states collect information about the client's religion, but only West Central State is concerned about the client's specific Protestant affiliation (see variable 39). Only West Central collects information about the location of the client's residence (see variable 42). Consequently it is the only state that can explore whether a client's residence has any relationship to his client career. When one considers the importance of physical problems, and educational attainment in the placement and admissions process it is surprising that only two out of four states collect such information (see variables 45, 46, and 52). In Chart VI only East Central State does not identify the offense patterns which led to 44

the client's institutionalization (see variable 62). Release Formats East Central, Central, and West Central State provided us with information about clients released from correctional institutions. However, East Central State will not be included in this analysis because it combines its admissions and release data. Its inclusion would have required considerable recalculation and a departure from the mode of analysis employed in this section. Table 1. 2 shows that there are neglible differences between the two states in the amount of release information that is collectedo However, Table 1.2 does not include the ten variables pertaining to the organization's assessment of the client's personality and social adjustment which we discussed in the introduction to the corrections sectiono Thus when all variables are considered Central State collects more release information than West Central State. When type of information is considered Central State collects information about the client's parents, whereas West Central does not. It may be that Central State weighs such factors as parental affection for the child, family cohesiveness, and related variables (see 54-59) to determine the type of release, eg.o, direct discharge vso release to aftercare. West Central State's slight dominance on the career dimension is due to the repetition of certain admissions information on the release format, namely, the date of admission and the offense that led to institutionalization (see variables 60 and 62 Chart VI)o The table also shows that Central State collects more information about the client's experience or adjustment in the institution. An examination of these items in Chart VI (variables 72-79) reveals that one of West Central State's variables identify what the institution did to the client, io eo, privileges lost, number of lock ups, and the other two variables are gross evaluations of vocational and educational performance. All of West Central's variables are organizational evaluations of the client's behavior in the institution. Admissions Release Formats Within States The rationale for comparing different formats within the same state is quite simpleo It shows us that organizations collect different kinds of information at different phases of the client's career. This suggests that the variation in the information collected is a function of the processing or decision making tasks which confront the organization and the client. For example in Table 1.2 note the drop in the amount of information collected about clients from the admissions format to the release format-from 9 to 7 variables in Central and from 19 to 8 in West Central. Also note the increase in the information collected about the parents in Central State from one variable to five variables. Or, take either one of the two states and simply go down the 45

admissions and release columns in the charts to see the changes in the formatso At this point in our research we are not prepared to say anything conclusive about these observations. We merely want to indicate that this mode of analysis may prove to be fruitful in raising questions about organizational rationales for differential use and nonuse of informationo For example, why is information about the client's family important at the time of release but not at the time of admission? Why is it important in Central State, but not in West Central State? The kind of comparative analysis suggested here will help organizations to rationalize their information collection and processing as well as their decision-makingo Discussion In contrast to child welfare it is quite obvious that much more extensive and intensive information is collected about juvenile offenders who are being served under probation, institutional or parole programso Furthermore, examination of the substance of the categories and variables within the corrections formats indicates that more specific data related to the processing and rehabilitating of offender population is collected and used, eogo, details about offense patterns, types of supervision, psychological and psychiatric evaluation, social adjustment, etco One can hypothesize that the level of knowledge re delinquents is greater than that about children who are dependent, neglected and in need of protective serviceso Also the technology for intervention and treatment is better developed and rationalized more completely throughout the agencyo Last but not least this difference may be a reflection of the goal priorities between the two programs as they operate within a single agency or in two separate agencieso The goal for child welfare service may largely involve general protection and maintenance of the child whose family is unwilling or unable to care for him. To achieve this goal it is nct perceived that one must obtain highly detailed information which is routinely recorded-only information to insure maintenance and care is required. In contrast juvenile offenders present serious problems of social controlo Agencies are likely to perceive the need for much more detailed and differentiated information if control and rehabilitation goals are to be achieved. Society will not tolerate a mass undifferentiated custody of juveniles. Instead they are to be controlled and rehabilitated in a differentiated mannero That this hypothesis is likely to be supported in subsequent data analysis is borne out by findings from an earlier study by the project director of juvenile correction institutionsr In that study institutions seeking only custodial goals obtained very minimal information about the clients, io eo, only as much as was necessary to maintain physical control. In contrast those correctional institutions which placed a high priority on treatment and rehabilitation of 46

offenders obtained comprehensive and detailed information about each offender.* 1. 4. CONCEPTUAL FRAMEWORK Throughout our report of the findings from the comparative analysis we indicate that the states have different variables in formats pertaining to similar programs, i.e., at the time of' admission to an institution one state collects information about the client's living arrangement and the other does not. We also reported that when similar variables are found in formats pertaining to similar aspects of a state program, often the categories within the variables differ, e.g., two states collect information about birth status but one state also collects information as to whether paternity was established. These findings show that clients are differentially labeled or categorized in organizational settings and it raises questions about the consequences of such labeling on organizational processing. We must keep in mind that the formats within which these variables and categories are found specify the kinds of information which the state expects its county institutional, or organizational subsidiaries to collect. It is our assumption that the format also informs and cues organizational functionaries to the factors that should be considered in processing and making decisions about clients, i. e., the format or form provides a decisional set for the agent from which he selects criteria to process the client in one way or another. Thus the format is a social control mechanism for the organization in that it constrains the agent to consider and to weight a certain set of variables. With this perspective in mind, the type of status an individual has as a deviant or as a client is not entirely a function of parental deviance nor of some intrinsic feature of the client. Instead deviant and client statuses are seen as outcomes of organizational structures and operating modes, i.e., the organization is structured to give selected responses to certain individuals. Labeling or categorizing serves the organizational function of defining, simplifying, and standardizing the way in which organizational functionaries are to operate toward certain clients. Given this perspective, our attention is drawn to such dimensions as: (1) the acquisition of deviant identities, and client statuses, i.e., who applies the deviant label to whom; what characteristics of an individual or of his performance are salient to organizational screening processes, e.g., in Central State 90 percent of the probationers are White, 7 percent are American Indian and 3 percent are Negro; in the institutions 63 percent are White, 27 percent are Negro, and 10 percent are other**; and (2) the consequences of labeling, i.e., given that paternity has not been established does this make a child less adoptable? *See D. Street, R. D. Vinter, and C. Perrow, Organization from Treatment, N.Y.: Free Press, Division of Macmillan, 1966. **In the following section we present some finding on analysis of client characteristics and career tenure for East Central. 47

The notion of career suggests that the organization's decision to categorize or label an individual is not a simple act but a transitory process with identifiable stages. According to Erikson there are three related stages: (1) confrontation between the client and the organization; (2) the announcement of a judgment about the nature of the client's deviance; and (3) an act of social placement. * We are presently engaged in an elaboration of this theoretical framework as a basis for developing a more complete set of propositions to be tested within and between the several states. We have formulated a partial list of the variables to be used in this process. Variables INDEPENDENT VARIABLES 1. Client qua Client a. Personal and social characteristics of clients prior to entry into the organization: age, race, social class of family, residence, physical handicaps. 2. Client Cohort a. Summary of individual variables using proportions and rates for groups of clients. Personal characteristics, external diagnostic labels might be used to develop cohort groups for analysis. 5. Organizational Level a. Source of referral b. Definition of client's problem, e.g., type of problem-dependent, neglected delinquent, mentally ill. c. Reason for service d. Types of information used to assess, process and treat clients, e.g.- influence of clinical diagnosis on disposition and career. *Kar Erikson, "Notes on the Sociology of Deviance," in H. Becker (ed.), The Other Side: Perspectives on Deviance. Glencoe, Illinois: The Free Press 1964, p. 16. 48

eo Organizational assessment of clientvs adjustment, e. g., how does it influence type of discharge and intensity of supervision. f. Range of programs and integration of agency programs, e.g., affect of institutional vso public school on institutional careers of clients. DEPENDENT VARIABLES lo Tenure of Client Careers ao Length of stay in institutions; length of time taken to process a client through the organization. b. Number of years committed; maximum sentenceo co Elapsed time between release and readmissiono 2. Modal career patterns for various cohorts in different types of institutionso 3o Reason for Termination and Type of Discharge, e.g., direct discharge to parents, release to aftercare-group home vso parole. 4o Type of Admission or Placement, io eo, type of institution, institutionalization vso probation. 5, Evaluation of career alternatives by professional judges-rating quality of service relative to need, 6. Discrepancy scores measuring differences between formal requirements as stated in manuals and empirical career patterns. INTERVENING VARIABLES 1o Characteristics of staff who serve clientso 2. Use of data from survey of welfare administrators to assess some of the conditions which influence labeling and processing, e.go, caseload size, organizational structure, staff tenure, staff trainingo 35 Organizational factors (Institutions providing services to clients ) Data about residenti al institutions for children have been obtained from Donnell Pappenfort from his study sponsored by the Children's Bureau of the Department of Health, Education and Welfare: Grant No, He E.Wo-WA-PR-700o Variables include organizational size, residential 49

capacity of institution, staff training and qualifications, services provided, type of treatmento In our report about the findings from the comparative analyses of the formats we made indirect references to our plans for the analysis of the relationship between some of the variables. We are presently engaged in a more systematic development of hypotheses and strategies of analysiso Adequacy of the State Corrections Data There are many criteria for determining the adequacy of data, eo g., validity, reliability etc. Although we plan to evaluate the validity and reliability of the data, for the task at hand we will consider two other criteria: (1) do the state formats have the variables which we consider important in the study of client careers; and (2) are the data collected, processed and stored in a way that the data can be retrieved and manipulated for analytic purposeso Given the data in hand it is not possible to do a longitudinal study of corrections' careers that spans the clients juvenile court, probation, institutional, and parole experiences. An examination of the formats reveals that different states can optimize the study of selected phases of a client's career and that certain states have more adequate data about personal characteristics whereas another state has more adequate client career informationo For example, Eastern State with its juvenile court data can study the early phases of a client's career but it has no data about the tenure of client careerso Central and West Central States with their parole data can study the later phases of a client careero West Central State has the most complete data about the client's personal and social characteristic, with the exception of one important variable-incomeo Thus the adequacy of the state corrections data varies depending upon what kind of analysis one wants to do. For example, if one were interested in an extensive analysis of client social characteristics and their relationship to type of placement then West Central State has the most adequate datao On the other hand if one were interested in examining the influence of the institution on client careers then Eastern State is most adequate because it has many institutions and they vary in terms of auspices and function~ Our experience with the preliminary analyses of the data indicates that for certain states the data are not organized in a way that allows for their manipulation for analytic purposeso For example, Central State's formats represent different files on a computer tape. Information from one file cannot be cross tabulated with information from another file, eo go we cannot examine the relationship between the offense committed and the length of stay. Thus, the computer tapes had to be reorganizedo This meant that the files had to be scanned for similar identification numbers and then placed in a new and more comprehensive fileo In addition we have discovered double punches in columns, wild codes and missing datao All of these problems are not insurmountable but 50

they do delay the analysis of data and they place temporary restrictions on the kind of analysis that is possible. 1L 5. SOME BASIC DETERMINANTS OF CLIENT CAREER TENURE GoffmanVs cogent description of careers of mentally ill persons and his conceptualization of the "total institution" has provided both a stimulus and a challenge to many students of human service organizations.* Subsequent analysis by Vinter, Brim, and Wheeler, and others indicates that there is a broad range of organizational behavior manifested by various people-processing and people-changing agencies, with subsequent differential outcomes for clients who are primary members of these organizationso** From the original socialization or "prisonization" to the final discharge there are a variety of interesting problems about client careers which have preoccupied the minds of researchers and administrators alikeo In much of this literature, the process of identification with the institution, its goals and its staff has come to be viewed as a sine qua non for effective change.*** To treat personal and institutional identification of clients as an independent variable for predicting and explaining socially desirable outcomes, however, is only meaningful if one is able to account for the variation in the explanatory factor: identificationo One's ability and willingness to identify as well as his actual identification with an organization or institution is thus viewed as a dependent rather than an independent variable in this papero Ultimately of course, one must come to grips with the problem of the relationship between identification and subsequent behavior. If one views identification as a dependent variable, the *See Erving Goffman, Asylumso Garden City, NoYo: Anchor Books, Doubleday and Co., 1961. **Vinter poses a range of human service organizations, using goal priorities as a basis for developing his typologies. See Robert Do Vinter, "The Analysis of Treatment Organizations," Social Work 8 (July, 1963) 3115; and R. D. Vinter, "The Juvenile Court as an Institution," in President's Commission on Law Enforcement and Administration of Justice, Task Force Report: Juvenile Delinquency and Youth Crimeo Washington; Uo So Govt. Printing Office, 1967, pp. 84-90; and Orville Brim and Stanton Wheeler, Socialization After Childhood. New York: Wiley, 1966. ***Adamek and Dager report findings which show that the degree to which indiduals are changed is related to the extent to which they identify with the staff and the institution~ See Raymond J. Adamek and Edward Zo Dager, "Social Structure, Identification, and Change in a Treatment-Oriented Institution," ASR 5533 (December, 1968), 931-944o 51

quest for independent variables of identifiability becomes mandatory. Needless to say, a host of variables can be used which at least theoretically affect and determine potential and actual identification of clients. Perhaps some of the more important ones are a client's prior experience and background, the organizational context and institutional environment, and his prospects of being released from the institution within a certain period of time. It is precisely this last factor that we are interested in here. We have observed as have many others that clients are often aware that career contingencies operate in very different ways to affect the length of stay. For example, certain types of offenses such as narcotics using and draft evasion have often merited far longer commitments than have more serious offenses such as embezzlement or robbery for the same general classification of offender. Unfortunately, researchers and administrators often have only noted this difference in a very general way, but have not tried to assess its consequence. We wish to identify the different types of career patterns and then study these in relation to personal characteristics and organizational and societal behavior. We assume that the tenure of client careers is a consequence of the interaction of personal and social system variables. More than a few studies in the recent past have addressed themselves to the implications and consequences of length of stay of clients in different closed institutions. Little doubt remains that tenure is an important factor in explaining subsequent behavior of the committed person after institutionalization. As Street, et al., have reasoned, average length of stay may be viewed as a measure of effectiveness insofar as it required a certain time before institution staff are ready to say that the committed person is ready for release.* This amounts to saying that length of stay can be seen as an indicator of the social distance between socially desired and actually demonstrated behavior of the client. The entire argument thus is predicted on the assumption that peoplechanging processes not only take place over time, but that they are significantly affected by the very boundaries of this time parameter. Whereas there is considerable consensus about the fact that time affects the final outcome of clients, opinions widely diverge about the way in which length of stay interacts with different types of institutional contexts and their graduates. Thus Clemmer in his classical work as well as McCorkle and Korn posit a steady increase in negativism in the custodial prison with prolonged stay because clients become more fully socialized to a negative inmate culture.** On the other hand, Wheeler observed a U-shaped pattern of resocialization and he questions the validity of the traditional solidary opposition *David Street, Robert Vinter, and Charles Perrow, Organization of Treatment. New York: Free Press Division of Macmillan, 1966, pp. 195-197. **Donald Clemmer, The Prison Community, New York: Rinehart, 1958; and Lloyd McCorkle and Richard Korn, "Resocialization Within the Walls," The Annals, 293 (May), 88-89. 52

model of inmate cultureo Street, et al., points out that although length of stay may be important, it does not account for the differences between type of institutions that subscribe to different treatment philosophies.* In the treatment-oriented institutions greater proportions of the clients expressed positive perspectives than in custodially oriented institutions at every point in time. But, there was a drop-off in the middle months and then an increase toward the end of the client tenure. In contrast, in the custodial institutions a continual increase in negativism was observed over time. Adamek and Dager also show the influence of length of stay on identification and change in a treatment-oriented institution.** This divergence of research findings calls for a search of the determinants of length of commitment in various peoplechanging organizations. If we succeed in shedding some light on the basic factors that explain length of commitment, we hopefully will be able to say something more definitive about the consequences resulting from this rather ambiguous variable. Data Because tenure of client career in an agency depends on many factors-one of which is the very nature of the organization-it is necessary to have data for a large population distributed over several types of agencies in order to allow for the required controls in singling out any one factor. To collect the data required to do an extensive analysis would be a formidable job with respect both to funds and research efforts. Fortunately, however, an abundance of data is available on official records. Such data have been collected in the routine processing of children as they become wards of the state through the juvenile court or other agencies. These data are seldom analyzed by the agencies which collect and retain them except for financial reporting and accounting purposes. They were made available to us for more comprehensive analysisboth for methodological and substantive inquiries, which in turn should provide information useful to legislative, administrative and professional groups in the field of social welfare. Needless to say the advantage of having ready access to these data is somewhat paid for by the fact that one must engage in considerable data preparation and secondary analysis. Thus, we had to cope with predetermined code categories which were often poorly defined, with lack of information, and with improper coding. Nevertheless, we decided that much could be learned from more intensive analysis. In this section we are reporting about the total population of wards of the state in one Midwestern state, identified as East Central in our larger *Street, po cit., p. 209. **Adamek and Dager, op. cit., po 938 53

study of client careers. The total population of wards for the period, 19651967 was 5,041. East Central has a total population of about eight million people, and the state department of social services administers a broad range of programs for children who are dependent and neglected, in need of protective services, and delinquents. East Central is a midwestern state with considerable industrialization, but it is comparable to many northern states with regard to race, residency pattern, education, income, and welfare programs. Methodology The overall length of stay for the 4,857 cases analyzed (184 cases were dropped from the analysis as there was no information available for them on one or several variables under scrutiny) was 25.15 months. This means value by definition conceals all underlying differences among the various groups committed to wardship of the state. Previous analysis of several other states sensitized us as to the most crucial variables accounting for intergroup differences of wards of the state. It was found that age at commitment, race, type of institution, client's residence, reason for service, marital status of client's parents, client's living arrangement, and his religion best predict just how long he would be committed as ward of the state.* Because we were not interested simply in predicting patterns of association but rather in identifying causal modes of interaction between these independent variables and length of stay simple cross tabulations and zero correlations would not do. Moreover, the latter impose linearity and additivity assumptions on our data which are largely of a nonparametric nature. Further the intercorrelations between explanatory variables and the interaction between them and the dependable variable make difficult the construction of precise theoretical models reflecting chains of causation, particularly where the number of explanatory variables is large. Fortunately, a recently developed computer routine-Multiple Classification Analysis (MAC III)-has been designed precisely to do this job: To compute the amount of variance in the dependent variable owing to variation in a given independent variable while controlling for all other variables under analysis.** Moreover MAC III is able to handle a large number of predictors without making any assumptions about the statis*See R. C. Sarri, W. Grichting, and J. E. Tropman, "Individual and Family Characteristics of Child Welfare Wards," unpublished paper, The University of Michigan, School of Social Work, 1969. **F. M. Andrews, J. N. Morgan, and J. A. Sonquist, Multiple Classification Analysis. Ann Arbor: Survey Research Center, The University of Michigan, 1967. 54

tical strength of the scales underlying the explanatory variables, i.e., they may be of a nominal or ordinal nature. Nevertheless, the dependent or predicted variable is assumed to be of at least interval strength, a condition met by number of months of commitmento Moreover, the independent variables are assumed to be mutually not interdependent. The mathematical model underlying the MAC III is the following: Y.., = Y + a. + b + Ck + d +. o+ ijklo.. i j k 1 ijklo... where Y = The score of social unit al that falls in category i of background factor a, category j of background factor b, category k of background factor c, category 1 of background factor d, etc. Y = The grand mean of all social units considered in any given group to be analyzed. a. = The fixed deviation from the grand mean associated with category i of background factor a. b.,ck,d = Same as a J' k1 i ~iekl = Error term for social unit ijkl...a Using the above mathematical model, MAC III computes the total and explained amounts of variance, respectively, in the dependent variable owing to all independent variables included in the analysis. It then drops one independent variable at a time and recomputes the reduced amount of variance in the dependent variableo This procedure is repeated until only one independent variable is left over. The difference in the amount of explained variation obviously is attributable to the predictor variable which was dropped between any two runs of the program. Using the squared multiple correlation coefficient prior to making allowance for the degree(s) of freedom of the explanatory variable is left over. The difference in the amount of explained variation obviously is attributable to the predictor variable which was dropped between any two runs of the program. Using the quared multiple correlation coefficient prior to making allowance for the degree(s) of freedom of the explanatory variable, r2 was computed as the explained variance over the total variance. Table 1.3 gives the percent distribution and rank of predictive efficiency of selected background variables for length of commitment. 55

TABLE 1. 3 PERCENT DISTRIBUTION AND RANK OF PREDICTIVE EFFICIENCY OF SELECTED BACKGROUND FOR LENGTH OF COMMITMENT Percent Variables Explained Alone Cumulative 1. Age at commitment 9.14 9. 14 2 2. Race 4.534 15.48 3 35. Type of institution 21.99 35.47 1 4. Residence.65 56.12 5 5. Reason for service.48 36.60 6 6. Marital status of parents.16 36.76 8 7. Living arrangement of ward 3.64 40.40 4 8. Religion.26 40.66 7 Notes: 1. Age is coded in 7 categories ranging from "less than one year" to "18 years and older." 2. The race codes are "White," "Black," "other" (mostly Indian). 3. There were 9 "types of institution" used, namely, "after care-home," "after care-other," "custodial," "maximum security," "treatment," "training," "adoption," "day care," "other." 4. By residence reference is made to "metropolitan," "urban" (= counties with 100,000 population and 60o urban (census definition)), and "rural." 5. Reason for service refers to East Central's manual according to which differential treatment is accorded to "neglected children," "children whose parents have a problem," "children with their own problem," and "other." 6. Marital status of parents coded as follows: "married," "remarried," "father dead," "mother dead," "both parents dead," "divorced," "separated or deserted," "single," "putative father." 7. Living arrangement ranges from "lives with both parents," "lives with mother 56

only," "lives with father only," "lives with mother and stepfather," "lives with father and stepmother," "lives with relatives," "lives with foster parents," "hospitalized," to "institutionalized" (not hospital). 80 Religion is coded "Catholic," "Protestant," "Jew," "none," and "other." Findings and Discussion Because of the large number of records at our disposal almost all statistical differences are beyond chance and reflect an actual difference in the population provided one is willing to view our sample as representative of a population trend over timeo There is statistically no way of testing for this since it is impossible to compute an intercluster variance having only one cluster. But, at any rate, statistical significance tests would be of little value given that nature of our datao It appears that social facts in the Durkheimian sense (external and constraining) best account for the differential treatment-in terms of length of commitment as wards of the state-welfare organizations accord their clients. Some of these factors at first glance may be seen as strictly individual characteristics; however, upon closer inspection one realized that many of these variables ought to be seen within a social system's frame of reference as they prompt organizations to react in certain socially defined modes. What then are the best predictors for length of commitment given the information which we have available? The total amount of variance "explained" in Table 1.3 equals 40.66 percent, more than half of which is accounted for by the type of institution to which a ward is sent, namely, 21o 99 percento This latter percentage looks rather suspicious, especially because "reason for service" which one would expect to be highly correlated with "type of institution" does not account for any substantial amount (.48%) of variance. As we pointed out above, Multiple Classification Analysis assumes orthogonality among all independent variables. If these explanatory factors represent a causal chain, the amount of variance "explained" by any given factor is a function of the sequence in which this factor is fed into the analysiso In other words in the MCA a given portion of variance which can be accounted for by two or more overlapping independent variables is attributed to the first variable under analysis and the remaining ones appear to be unrelated to the explicandum. Quite clearly, this is vividly opposed to zero-order correlations, where for instance both education and occupational status correlate with income and thus produce the by now famous (or infamous?) booster effect. The booster effect refers to the fact that if two (or more) independent variables are positively related to each other, that is, if they overlap, the sum of the two, each considered separately, explains more of the variation in the dependent variable than both together as a unit factor. 57

The multiple classification analysis overcomes this problem by "keeping track" of how much variance has been accounted for as new variables are introduced in the analysis. The following hypothetical situation is helpful in shedding light on the two methodological problems: (1) It shows the difference between zero-order correlations and the MAC in computing the strength of association between two or more independent overlapping variables. (2) It illustrates the increase or decrease of relational fertility of independent overlapping variables as a function of the actual input order. A AB B C = Dependent variable (= total variance to be explained) AB = Independent overlapping variables AB = Area of overlap (= explained by both A and B in customary zero-order correlation, and by A or B in MAC III contingent upon input order). One way of finding out whether one's suspicious with respect to the order of input generally, as well as with respect to "type of institution" specifically, are justified is to order the input variables in the most logical sequence. By logical sequence, of course, is meant the time series that most likely underlies a cause effect scheme of these eight explanatory factors. Table 1. 4 is based on what we believe to be the most logical causal path. Even though effectively the same amount of variance has been "explained" in runs 1 and 2 of MAC III (40L66% vSo 41o34%), the allocation of explained variance to different variables has changed considerably. Indeed the rank order correlation coefficient for the relative importance of each variable in the two runs equals only +~619. Quite clearly, the relative importance of any intervening variable analyzed in MAC III is a function of the input order. This is crucial to realize as many misleading results may be generated, mostly due to inattention. Overlooking this peculiarity of MAC III and possibly other techniques of regression analysis may well result in highly distorted claims of relational fertility of certain variables and turn out to be a serious impediment to realistic model building and theory construction. 58

TABLE 1. 4 PERCENT DISTRIBUTION AND RANK OF PREDICTIVE EFFICIENCY OF SELECTED BACKGROUND VARIABLES FOR LENGTH OF COMMITMENT Percent Variables Explained Rank Alone Cumulative lo Race 6 71 6.71 4 2. Age at commitment 60 52 135 23 5 35 Residence.77 14o00 6 4. Religion.57 14.57 8 5. Marital status of parents o 71 15.28 7 60 Living arrangement 9~ 94 250 22 1 7. Reason for service 7 68 32.90 3 8. Type of institution 80 44 41. 34 2 Notes: See Table 1o 5. In comparing the two tables several interesting observations of a more specific nature are worth making. First, by putting "type of institution" at the end of the analysis "reason for service" now appears to be a rather powerful predictor for length of commitment simply because "type of institution" cannot "absorb" the variance accounted for by "reason for service" which precedes "types of institution" in the analysis. What is even more interesting is the fact that "living arrangement" loads up to such a degree that it turns out to be the most potent predictor among all explanatory variables. Moreover, a comparison of the two tables indicates that age at commitment depends on race inasmuch as race and age overlap with respect to length of commitment. However, since race temporarily precedes age at commitment race must be analyzed prior to age of commitment. The variable sequence in Table 104 indeed takes the concatenation cf these variables into consideration by following a path as it is believed to evolve from "race" to "type of institution." In the total sample 41l.534 percent variance is accounted for by 8 independent variables. However, three of them, namely residence, religion, and marital status of parents collectively account for only some 2 percent variance and thus can safely be omitted from our further analysis. As can readily be observed the five remaining variables of major impact are of varying importance not only for the total sample but also for the different subsamples which are based on the 59

various categories of the "reason for service" variable (see Figures 1.2-1.4) as Table 1.5 indicates. TABLE 1. 5 PERCENT DISTRIBUTION OF RELATIONAL FERTILITY OF SELECTED VARIABLES FOR PREDICTING LENGTH OF COMMITMENT, BY REASON FOR SERVICE Percent Variables Accounted For Type of Living Reason Race Age for Total Institution Arrangement Service Total sample 6.71 6.52 8.44 9.94 7.68 59.29 Neglect 4.95 4.86 10.4-8 8.65 - 28.94 Parents' Problem.46 5.60.12.72 - 6.78 Child's Problem 2.77 13.63 2.04 20.81 - 39.25 Table 1.5 can best be studied in conjunction with Figures 1.1-1. 4 which graphically display and elucidate Table 1. 5. Looking at the race variable it appears that Whites generally remained wards of the state for a longer period of time than Blacks (23.3 as 20.2 months, see Figure 1.1). On the other hand, Figures 1.2-1.4 make abundantly clear that Blacks rather than Whites are kept for a longer time on every single program. How is this possible? Obviously the mean values of the total sample are statistical artifacts and do not represent true life situations as they are the weighted means of highly skewed distributions in all individual programs. A disproportionately large number of Blacks are diagnosed as having a problem by themselves (delinquency) or as being the result of problematic parents, whereas the Whites are more likely to be given treatment as neglected children. To put it bluntly: if a White boy and a Black boy get caught shoplifting, chances are good that the White boy will be treated as a neglected child and thus come to the attention of the local department of social services, whereas the Black one will be defined as a delinquent and probably go before the juvenile court. The seriousness of this situation is a compound effect of (1) the well-known fact that a disproportionately large number of non-Whites become wards of the state, i.e., about 10 percent of West Central's population is non-White, but some 59 percent of its wards of the state are recruited from the non-White population contingent; and (2) the proclivity of the system to define non-Whites as delinquents as Table 1.6 makes clear. 60

76 74 72 70 68 66 64 62 c 60 58 56 ~ 54,t52 / 50 50 __* _ *\ _ *48 ) 40 * * N / \. * 44 ** * * * / \ * 46 * 34 *4 432 * * 40.3 * 28 * K / " 326.* * \ 3 22Al 1 4.. T* sm\' 31^ ~ ~ ~ ~ 6.. 12 * / \ 10 *6 8 * \ 24's )- C le Ra- Qce, 2f,r, Co Co CI " SP co.-N gNN {^-^ C<1IN ^ ->v C^C^ji co I C-l 0^ CO I 3 CYc ( i0 16.....' Age C^ Q;~ z';^Jl Q Co I N C *.1 ~ t C co C i'Co Co q ~a'-' 10 ~ 4C -~0 S 0Ci^-S 0a-i Ci'1 0^ C 3 9 CoC~~ 0 ~ - Living Arran gem ents m^i^-~~~~~~~L PL4 - Reason for Service Figure 1.1. Total sample.,.%,,.~, %.,.*,...61 nt

< loo + + * ~100 ~ * * 95. *: 90 * * t 85 * *,. * * t 75 * * | 70 * * *. *. * 65 * * *. * I 60 F* -. * * X= *, t 350.'^ *. " * * * * -I4 /.* * ~ * ^ ^f\** * * ~l)3 *.** * 130 *' *' 55 - 40/ 0 r0 ^~'~^ ~2 > c, Race,' *. ~L ) -jI 9' s co. _2_v_0 -— ~ —---—, —-0 Z, L l) -,,'-?C <?^0 < I l I cType Inst. * * * *, * * * ~* * * * * * * * * "* m* * * 8 *; * * L* * * * *1 * * * *j *.... i- 1.-1 i _^ ~ iT~ ^i i I.'^ I' |6'2

1 \ 1 \ - 1: \ K 1 \ oy \ *g 11 \ 0 \ &S ~ ~ Fgr C3.ide whs\aet aeapolm — _.... \'*~.V." i~ * *.* Race. - ~.,,,,,1.., T-I'^ i^ |I- - I,~ Q,.Aae ~ E AlI are on "day care" X = X= 8.33 Type Inst. *4^^1/ 4^0c 4^C 4S- t-A CO - II — R1'~- Living Arrangements I I I i I I I Figure 1.5. Children whose parents have a problem.

206 months / \ 60 / c /o - 55 \ 50 \ - 45 \ I \ 35 340 205. _ _'* = 18.57 25 **.. 10 * *** =8 c' 5'' P \ 1 ~ < c Race 10 *.,... co co co * C ) Ag e 7,l 1,<.... 1 ~ I I o, 1I I 1 IQ~ r ^ C Co Q Col- V*i C C' Co,-, C- - -- - -/ - - * - -?0OS. S^ Cj'-i ___0_____-_______I -B' oi'~ O a Figure.. Children with their' own problem.

TABLE 1.6 DISTRIBUTION OF WARDS IN WEST CENTRAL STATE, BY RACE AND "PROBLEM LABEL" Neglected Child Parentst Problem Child' s Problem Total N % N % N % N White 693 24 337 12 1839 64 2869 Non-White 211 12 171 9 14o4 79 1786 904 508 3243 4655* X2 = 124o06; 2 do fo p < OO1 *The total number of cases analyzed dropped to 4655 because the "other" category in the control variable was eliminated from the analysis. Quite clearly, Blacks remain wards of the state for a longer period of time than do Whites notwithstanding the overall statistical average which seems to gainsay such a statement. Apparently Blacks do not have as ready access to legal redress as do Whites once they have been declared wards of the state. Yet the difference between Whites and Blacks is rather negligible as soon as the Indians' plight is introducedo Indians, roughly speaking, remain on wardship about twice as long as do non-Indians. If access to courts and justice is seen as a major determinant for length of commitment this is certainly not surprisingo Nevertheless it should be pointed out the Indians are most likely to be treated as neglected children. It is fair to say that Indians thus far have been seen and treated as objects of mercy rather than as public dissenters who are to be kept under tight supervision and control, eogo, Negroes, who have come to be socially acceptable scapegoatso The effect of the ward's age at commitment on length of commitment is rather blurred in Figure 1 1 but it proves a valuable predictor for length of stay because age is a basic determinant for program selection. Thus wards below age 13 are most likely to be treated as neglected children or as children of problematic parents, whereas wards above age 13 are more likely judged responsible for their own misbehavior. Nevertheless, there is a slight tendency to accord "adult treatment" to Blacks at an earlier age than to Whites, i.e., among Black neglected children only 5 out of 10 are 15 years and older whereas there are 4 out of 10 in the White group who are above 13 and still treated as neglected childreno Race and age at commitment are characteristics of the clients themselves. Both are viable predictors of how long a ward will remain committed, but the organizational properties of those structures which prepare and process a given individual are more powerful predictors of the tenure of a career than are psychological or personal characteristics~ With regard to living arrangement of 65

the client prior to commitment several interesting comments can be made. First of all, it is that variable in Table 1.4 which explains the largest amount of variance (9.94%). This is probably due to the fact physical location of the child constrains decision-making significantly. If he is already institutionalized and does not have a family situation to which he can return, then continued institutionalization or foster care may be the only possible alternatives. Depending upon his age and the availability of foster homes, continued institutionalization may be viewed as the most likely alternative irrespective of the client's personal characteristics or the desirability of this decision. These findings thus provide substantial support for our early assumption that client careers are a function of the interaction of personal and social characteristics of clients, and of the organizational characteristics of those agencies established to be responsible for child welfare wards. It becomes increasingly apparent that future policy planning should take such factors into consideration. Until now far greater reliance in decision-making has been placed on individual attributes probably because little was known systematically about organizational behavior vis-a-vis clients. 66

PART II SURVEY OF LOCAL WELFARE ADMINISTRATORS 67

2.1. INTRODUCTION A second part of this research involved a survey of local public welfare administrators. The county (or district) welfare office was selected for study because it is the principal transactional organization between the federal-state structures and the client, and because of the crucial role it was believed to play in the initial nomination of child welfare wards in client careerhood. The survey was originally planned for four states from which we obtained child welfare career data, but in order to obtain a sense of the national picture, a sample of counties across the nation was drawn, and questionnaires were sent to each of the offices serving those counties. Some surprising complexities were encountered in carrying out this task, but these problems came to be findings in themselves because they provided us with new knowledge about the structure of state and local welfare organizations. The purpose of this section is to report some of our findings to date about the structure of local welfare agencies across the country and also some' findings from the survey of administrators in the four states where all of the counties were sampled. This section will be divided as follows: (1) Preliminary findings from the survey of administrators in four states. (2) Research design and methodology for the survey in the four states and for the national sample, including sampling and coding procedures. (5) Structural complexity of state public welfare systems as revealed from the national survey of local welfare administrators. Because procedures for development of the survey instrument were reported in an earlier progress report submitted in March, 1969, these will not be described here. It should be mentioned, however, that as the study progressed, we came to recognize some of the shortcomings of our original instrument; but without extensive prior field work, most of these findings could not have been anticipated. Hopefully, this study will encourage others to do more extensive and intensive research about local public welfare agencies. As we have suggested before, this is a sorely needed area of research on human service organizations. We wish to acknowledge the excellent cooperation received from state and local agency staff, for without this, the study would have been impossible. Completion of the questionnaire was voluntary on the part of the local welfare administrators, and it required time from their already very busy schedules. In addition, we needed the assistance of the state welfare staff in several states where it was impossible to obtain from the Public Welfare Directory sufficient information about the local structural patterns. Many persons wrote us long letters to clarify the situation; others commented freely on the questionnaire 69

to clarify their particular response or to indicate additional areas of concern that they thought we should know about. That the questionnaire and its findings were of interest to local welfare administrators is indicated by the high rate of response for a mailed questionnaire and by the large number of requests which were received for a report of the study findings. These data are presented and discussed in Appendix 2. D, "Survey Respondent and Request Rates." 2.2. SURVEY FINDINGS The following section presents some preliminary findings from our analysis of the survey data from four states. The national sample data are still being processed and are excluded from this presentation. The majority of respondents from the national sample who are expected to participate have already done so. However, the New England States have just been contacted for the first time. The delay in surveying the New England States has been due to the difficulty in obtaining information about their local welfare structures. The information on these states in the Public Welfare Directory was particularly inadequate for our purposes and consequently it was necessary to seek the assistance of the state welfare departments. In three of the four states, the agencies studied were public welfare agencies handling various programs, including AFDC, old age assistance, aid to the blind and the disabled, medical assistance, general assistance, and child welfare. In Eastern State, welfare programs were functionally differentiated into two agencies, public assistance and child welfare. The latter agencies handled child welfare programs only and the remaining programs were allocated to the public assistance agencies. Thus, for purposes of clarity and comparability, child welfare agencies in Eastern State are excluded from the tables which compare the states on different structural and attitudinal variables. The functional differentiation of programs in Eastern State presents us with an excellent opportunity to compare the effects of different programs on agency organization. Thus, many of the findings are presented using public assistance agency vs. child welfare agency as an independent variable. The relative size of the agency, measured in terms of the total staff, including professional and nonprofessional employees, was used as an independent variable in most of the findings presented. Size is a rough indicator of an important dimension of the welfare system. Large agencies are usually in urban areas in which the need for public assistance is usually concentrated. Small agencies tend to be in rural areas in which both the quantity of service required is less and the quality of service is different. Urban agencies are faced with more clients to process and a clientele with a less stable relationship to the city or county being serviced by the agency. A smaller caseload based on more stable relationships is likely in rural areas. Thus, the problems 70

and the solutions for small rural and large urban agencies are going to be different. The urban-rural dimension will be introduced into the analysis later when the demographic data is merged with the organizational data for the ecological analysis. Two classes of dependent variables are introduced in the following subsections: (1) socio-structural variables, e.g., caseload size, turnover rates, growth rates, and the allocation of effort to various areas of service. The lack of availability of information requested in the survey reflected the impact of certain external constraints and influences upon the agency structure. The type of accountability required by governments constrains organizations to collect certain kinds of information. The professional training of the administrators is also a determinant of the adequacy of information processing in the agency; (2) attitudinal variables, e.g., the administrators' beliefs about the ideal allocation of effort to financial aid and to different areas of service. A description of the findings has been presented along with a table in each case. Where possible, some interpretation of the findings has also been presented. At the end of the written text on the survey findings is a series of additional tables presented without comment in order to give the reader a look at some of the available data which has yet to be fully analyzed and interpreted. Agency Size and Staff Workload In order to study the different administrative problems that may exist in the different states, a comparison was made of the distribution of the number of client applications in the various agencies in each of the states. Administrative problems may be different for a state that has many agencies with relatively few applications as opposed to a state that has many agencies with large numbers of client applications. Table 2.1 is designed to show the difference in the distributions of the total number of client applications in 1967 in the agencies of the four states. More than 60 percent of the agencies in Western and Eastern States report over 1500 client applications per year, while the two Central States have fewer than 10 percent of their agencies reporting as many applications per year. This suggests that the administrative task may be greater in Eastern and Western States than in the other two. 71

TABLE 2. 1 NUMBER OF CLIENT APPLICATIONS BY STATE (In Percentages) Number of Client Western, West Central, Central, Eastern, Applications o..% 1 - 250 0.0 31.7 24.1 8.1 251 - 1500 36.8 62.0 72.5 29.7 1501 or more 63.2 6.3 3.4 62.2 (N)a (19) (63) (29) (37) Number of agencies excluding nonresponses. A comparison of the distribution of the size of the agencies in the four states is presented in Table 2.2. This comparison is analogous to the comparison of client applications since large agencies can be expected to be large because there are many client applications each year. As the reader can see, the distributions in Tables 2.1 and 2.2 are very similar. While this in itself does not prove the connection between total staff and client applications, it does show that states with small agencies also have agencies with relatively few applications per year. TABLE 2. 2 SIZE OF STAFFa BY STATE ( In Percentages) Number of Western, West Central, Central, Eastern, Staff o.. %o 1 - 10 4.3 49.3 25.0 12.5 11 - 25 13.0 34.8 52.8 30.0 26 or more 82.7 15.9 22.2 57.5 (N)b (23) (69) (36) (40) aThis is the total staff, including professional, secretarial and clerical, and indigenous workers. All respondents gave this information for their agencies. 72

More than half of the Eastern agencies and more than four-fifths of the Western agencies have over 25 staff members, includirg professional, secretarial and clerical, and indigenous workers. Less than one-fourth of the agencies in the two Central States are as large. Tables 2.1 and 2. 2 show West Central to be in the best position regarding the si2ze of staff and the number of applications per year in their agencies. This state has more small sized agencies and more agencies with small numbers of applications than any of the other states. Western State seems to be in the worst position, with more agencies with large staffs and large numbers of applications. The reason for the same distribution in client applications and total staff in the agencies of the four states is that more clients means more staff to handle those clients. This rather straightforward point is shown in Table 2.3. In West Central State, 80 percent of the agencies with the fewest number of applications (250 per year or less) have the smallest number of staff (10 or less). At the other extreme, all of the agencies with the largest number of applications (over 1500) are in the largest size category (over 25 staff members). The middle group of agencies, with between 251 and 1500 applications in 1967, are more likely to be agencies with between 11 and 25 staff members. In other words, there seems to be a direct correspondence between the number of client applications in the agency and its size. TABLE 2.3 SIZE OF STAFF BY NUMBER OF CLIENT APPLICATIONS IN WEST CENTRAL STATE (In Percentages) Number of Number of Client Applications Staff 1-250 251-1500 1501 or More 1 - 10 80.0 33.3 0.0 11 - 25 20.0 48.7 0.0 26 or more 0.0 17.9 100.0 (N)a (20) (39) (4) aExcluding nonresponses. Eastern and Western States are likely to have large agencies as well as agencies with large numbers of applications. This is an attempt to equalize the workloads in the agencies. West Central and Central States have smaller agencies and smaller numbers of applications. Despite this positive relation73

ship between the size of the agency ard the number of applications, there is still a tendency for the states with smaller agencies to have more agencies with low client/staff ratios. This is true whether one considers the total staff or just professionals. For our purposes here we are using the term client to refer to the number of client applications reported by the respondent. Table 2.4 shows that more than 25 percent of the agencies in West Central and Central States have a client/staff ratio of 20:1 or less (twenty or fewer applications per staff member) while only about 5 percent of the Western and Eastern States agencies have ratios so low. We can also observe in Table 2.4 that only Central State has a substantial number of agencies (24.1%) with client/professional staff ratios of 20:1 or less. TABLE 2.4 CLIENTa/STAFFb RATIO AND CLIENTa/PROFESSIONAL STAFF RATIO, BY STATE ( In Percentages) Western, West Central, Central, Eastern, Ratio b % of <f Client/Staff 1 - 20c 5.3 27.0 37.9 5.4 21 - 50 57.9 60.3 48.3 27.0 51 or more 36.8 12.7 13.8 67.6 Client/Professional Staff 1 - 20 5.5 6.3 24.1 0.0 21 - 50 21.0 42.9 27.6 18.9 51 or more 73.7 50.8 48.3 81.1 (N)d (19) (63) (29) (37) aClient refers to number of client applications in 1967. bTotal staff, including professional, secretarial and clerical, and indigenous workers. cThis includes all agencies in which the client/staff ratio ranges from 1:1 to 20:1. dNumber of agencies, excluding those that did not respond. In addition, states like Eastern and Western, which have large agencies and large numbers of applications per year also are more likely to have high client/staff and client/professional staff ratios. Table 2.4 shows that Eastern and Western States are three to six times more likely than West Central or Central to have agencies with client/staff ratios of more than 50:1 (67.6 per74

cent and 36.8 percent vs. 12.7 and 15.8 percent). The same table shows Eastern and Western to have more agencies with high client/professional staff ratios. Three-fourths or more of the Western and Eastern agencies have client/ professional staff ratios of over 50:1, while only half the agencies in the other states are in this position. Table 2.4 suggests that Eastern State is in the worst position while Central is in the best position as far as client/staff and client/professional staff ratios are concerned. This is reflected in the client/staff and client/ professional staff ratio distributions. Central State has more agencies with 20 or less client applications per professional worker or per staff in general, while Eastern State is far more likely to have agencies with a client/staff or client/professional staff ratio greater than 50. The relevance of this difference in organizational size distributions in the states is that a preponderance of large organizations designed to handle large caseloads is a very different administrative problem for the states than where smaller organizations are more likely. The smaller agency is more likely to provide an opportunity for a collegial atmosphere among welfare workers, a condition which may promote greater professionalism and a service orientation toward clients. It can also be observed that those states which have a greater number of small agencies (as measured by staff.size) also have a smaller client/ staff ratio. They therefore may be able to promote a greater service orientation among workers. West Central and Central States show similarity in the distribution of client/staff and client/professional staff ratios when comparing agencies above and below a ratio of 50 (Table 2.2). It is interesting to note that the difference between these states is primarily in the smaller agencies. West Central in comparison with Central has a disproportionate number of small agencies handling relatively large numbers of client applications (Tables 2.1 and 2.2). The reason for this is that administrators reported that services were provided to a substantial number of clients who did not qualify for financial aid. As a result, the same staff must be stretched further than the equivalent staff in Central State which does not have such a practice. The very high client/staff ratios for Eastern State may be explained in part by the fact that these are public assistance agencies that do not have child welfare programs, while the other states do have these programs integrated with other public welfare programs. It could be expected that service oriented child welfare programs in general will have low client/staff and client/professional staff ratios. Excluding these programs from public welfare agencies would create an overall increase in the client/staff and client/professional staff ratios. However, this does not explain why Western State is more likely to have agencies with high client/staff ratios in comparison to West Central and Central. Since client/staff and client/professional staff ratios could be expected 75

to be lower for child welfare programs, a comparison of child welfare and public assistance agencies in Eastern State should show the former to have lower client/staff and client/professional staff ratios (Table 2.3). Table 2.5 does point out the existence of a lighter caseload burden in child welfare agencies as opposed to public assistance agencies in Eastern State. One-third of the child welfare agencies have client/staff (total staff) ratios of 20:1 or less, while only 5 percent of the public assistance agencies have such low ratios. Similarly, a client/professional staff ratio of 20:1 or less occurs in about 20 percent of the child welfare agencies, but none of the public assistance agencies. At the other extreme, two-thirds of the public assistance agencies have client/staff ratios of more than 50:1, while less than 20 percent of the child welfare agencies have such high ratios. At the same time, while four-fifths of the public assistance agencies have client/professional staff ratios of over 50:1, this is true for only one-fifth of the child welfare agencies. TABLE 2.5 CLIENT/STAFF RATIO AND CLIENT/PROFESSIONAL STAFF RATIO BY TYPE OF AGENCY IN EASTERN STATE (In Percentages) Public Assistance, Child Welfare, Ratio Client/Staff 1 - 20 5.4 34.5 21 - 50 27.0 48.3 51 or more 67.6 17.2 Client/Professional Staff 1 - 20 0.0 20.7 21 - 50 18.9 58.6 51 or more 81.1 20.7 (N)a (37) (29) aExcluding nonresponses. While child welfare agencies in Eastern State show expectedly low client/staff and client/professional staff ratios compared with public assistance agencies in Eastern State, there still are, unfortunately, many agencies with ratios above 50. In addition, the distribution of client/staff ratios in child 76

welfare agencies in Eastern State (Table 2.5) is no better than the distribution in Central State for agencies that are responsible for public assistance programs as well (Table 2.4). A higher proportion of the staff in Eastern's child welfare agencies are professionals. Moreover, in Eastern State, the child welfare caseload per worker is the highest of the four states with a standard of 60 (see page 93 in the Methodology section). Thus, it seems that professionals in Eastern's child welfare agencies have an unusually great burden. However, it should be pointed out that Eastern State is one in which many social services are contracted by the state to private agencies. In this case, it is easier for a welfare worker to supervise clients who are in private agencies than if they are in public agencies and completely under his supervision. The relationship between agency size and workload can be more directly examined by relating the average caseload per worker in selected categories of assistance. Table 2.6 shows that smaller agencies tend to have lower AFDC caseloads than larger agencies in West Central State. While more than 80 percent of the small agencies (those with ten or fewer total staff) have average caseloads of 40 or less, fewer than 40 percent of the larger agencies have such low caseloads per worker. The smaller agencies are more likely to have mixed caseloads, i.e., to combine AFDC cases with other categories of assistance, due to the efficiency of such an arrangement where clients are geographically dispersed. This would explain why there are fewer AFDC cases assigned to workers. In addition, Table 2.6 shows that there is still a difference in child welfare caseloads between smaller and larger agencies. 69.6 percent of the snaller agencies have caseloads of 40 or less, while 56.3 percent of the larger agencies have caseloads of 40 or less. This is significant because it is rare for child welfare cases to be mixed with other categories of assistance in any agency. Therefore, the effect of the type of caseload structure, mixed vs. specialized, has been excluded. TABLE 2.6 AFDC CASELOAD AND CHILD WELFARE CASELOAD, BY SIZE OF AGENCY (Percentage Caseload Per Worker of 40 or Less) Category of Small, large,a Assistance to % AFDC 83.3 36.4 (24) (335) Child Welfare 69.6 56.3 (23) (32) aThe distinction between large and small is a relative one; small indicates agencies with 10 or fewer total staff, including professional, clerical, secretarial and indigenous workers. 77

Turnover Rates A turnover rate is an indicator of dissatisfaction. In this study the turnover rate is measured by the number of full-time professional staff that left the agency in 1967 (Question 6, Appendix 2.A) over the total number of full-time professional staff in the agency at the time the respondent answered the questionnaire (based on Question 4, Appendix 2.A). This measure of the turnover rate in general will be an underestimate, in view of the fact that the actual number of professionals is reported for a later period than the number who left the agency. Given the tendency for agencies to grow over time, the denominator in the calculation will usually be too large. It is assumed that the more likely it is that a professional will leave the agency, the less satisfied he is with being in the agency. In Table 2.7, a 15 percent turnover rate has been used as the cut-off point between high and low turnover rates. This is reasonable in view of the fact that in industry and business, the average turnover rate for different kinds of professionals is less than 15 percent. Thus, 15 percent seems to be a good standard for evaluating whether the turnover rate is high or low.* TABLE 2.7 TURNOVER RATES BY STATE (In Percentages) Turnover Western, West Central, Central, Eastern, Rate o %..... Lowa 42.1 31.8 45.7 83.8 Highb 57.9 68.2 54.3 16.2 (N)c (19) (66) (355) (37) al51 turnover rate or less. Over 15% turnover rate. CExcluding nonresponses to at least one question used in constructing this measure of turnover. *A more extensive search of the literature is being undertaken to establish with certainty the appropriate standard for evaluating turnover rates. A study of scientists and engineers in Canada showed the average turnover rates to be under 15 percent. Government turnover was higher than turnover in industry. 78

Table 2.7 shows that while the vast majority of agencies (83.8%) in Eastern State have low turnover rates, more than half the agencies in the other three states have high turnover rates. This would seem to indicate that Eastern State has the most satisfied workers and other professionals (e.g., nurses, executives, etc.). This is hard to explain in view of high client/staff ratios and client/professional staff ratios indicated for Eastern State in Table 2.7. It seems that where the professionals are worked the hardest, they are most satisfied. Further exploration into the working environment of the Eastern State agencies is necessary to explain the low turnover rates. Is it more professionalism in the agency promoting work satisfaction by such means as delegating responsible decision-making to workers? Or is it just the opposite? Are bureaucratic agencies more satisfying to a civil service oriented public assistance worker? Both explanations are plausible and further analysis is required to discover the best explanation. Size of the agency does not seem to be related to the turnover rates. There is no difference in the proportion of small or large agencies in West Central State who have turnover rates of 15 percent or less. The distribution of turnover rates of child welfare agencies in the Eastern State seems to be more like the distributions in the other states than like the public assistance agencies in Eastern State. Table 2.8 shows 48.7 percent TABLE 2.8 TURNOVER RATES BY TYPE OF AGENCY IN EASTERN STATE (In Percentages) Turnover Public Assistance, Child Welfare, Rate % _o Lowa 83.8 48.7 Highb 16.2 51.53 (N)c (37) (39) al5 turnover rate or less. Over 15% turnover rate. cExcluding nonresponses to at least one question used in constructing this measure of turnover. of the child welfare agencies with turnover rates of 15 percent or less, while 83.8 percent of the public assistance agencies have low turnover rates (15* or less). The child welfare respondents show more of a service oriented ideology and the child welfare agencies are more likely to attract administrators with 79

professional degrees. This suggests that the more service oriented workers in child welfare agencies may find their work inherently less satisfying, and, blaming the agency, are more likely to seek another agency in hope of realizing their goals. The reason for the lack of satisfaction among service-oriented personnel may be that the ideals of service learned in the professional schools and motivating the workers to become service workers are frustrated in the current structure of welfare organizations. The public assistance worker, who is concerned with routine tasks such as determining eligibility using relatively bureaucratic criteria, is probably less service-oriented and less well trained in the first place. Having lower expectations, the demands of agency work would be less frustrating for the public assistance worker than for his child welfare counterpart. More study of the differences between child welfare and public assistance agencies is necessary to do more than speculate about the causes of the differences in turnover rates. Staff Growth More than half the agencies in the study report no growth during 1967. With the population of these states continuing to grow and the number of people in need of assistance also growing, the implied lack of expansion in budgetary allocations for professional personnel suggests that there is increasing pressure on the resources available to serve citizens in need of assistance. Table 2.9 describes the distribution of growth rates of the agencies in TABLE 2.9 GROWTH RATES BY STATE (In Percentages) Growth Western West Central, Central, Eastern, Rate o,o % %_ Noa 52.6 69.7 51.4 62.1 Lowb 56.8 19.7 37.1 37.8 High~ 10.5 10.6 11.4 0.0 (N)d (19) (66) (35) (37) This category includes a small percentage of agencies which lost professionals during 1967. 1-150 increase in professionals during 1967. COver 15% increase in professionals during 1967. Excluding nonresponse to at least one question used a a basis for constructing this measure of growth. 8o

the four states. The growth rate of an agency is defined by the difference between the number of professionals (caseworkers, administrators, nurses, etc.) gained and lost during 1967 over the total number of professionals at the time of the administration of the questionnarie. Because of the frequency of "no growth" in which the number of professionals hired in a year equalled the number who left, this was made a separate category. "Low growth" and "high growth" were arbitrarily distinguished at below, and above 15 percent, respectively. A comparison of the four states shows that there is basically the same problem of lack of growth in all the states. Between 50 and 70 percent of the agencies show no growth in 1967. West Central State shows the highest proportion of agencies which have not grown in 1967. This is also the state with the highest proportion of small agencies (Table 2.2). However, it would be misleading to infer that there is some connection between initially having small agencies and the tenddecy for these agencies to remain small. Eastern State, with a high proportion of agencies with total staff over 25 (Table 2.2) is nearly as likely to have agencies with no growth as West Central. Moreover, Western State, which has the highest proportion of large agencies (Table 2.2) has as many agencies (10%) as West Central with high growth rates. The basic pattern of distribution of growth rates of agencies in the four states is the same and the small differences that do appear do not seem to be related systematically to other patterns of distributions discussed thus far. A comparison of child welfare agencies and public assistance agencies in Eastern State shows the same percentage of agencies with no growth (62%) (Table 2.10). However, there seems to be a differentiation among those agencies that TABLE 2. 10 GROWTH RATES BY TYPE OF AGENCY IN EASTERN STATE (In Percentages) Growth Public Assistance, Child Welfare, Ratea % %_ No 62.1 61.5 Low 37.8 15.4 High 0.0 23.1 (N)b (37) (39) aSee Table 2.9 for the definitions of the categories of growth. Excluding nonresponses. 81

have grown relative to the degree of growth. Twenty-three percent of the remaining child welfare agencies report high growth rates while none of the remaining public assistance agencies have high growth. This difference may be explained by the relative size of child welfare agencies; i.e., they tend to be smaller than public assistance agencies. A small increment in staff in terms of the absolute numbers can create a relatively large percentage gain for small agencies. Table 2.10 shows the proportion of child welfare and public assistance agencies which have either not grown, grown a little, or grown a great deal during 1967. An examination of the relationship between the size of the agency and the growth rate of the agency in West Central State tends to cupport the initial impression gained from Table 2.9 that small agencies tend not to grow. While more than four-fifths of the smaller agencies did not increase the proportion of professionals on their staff, only 56 percent of the larger agencies did not report growth (Table 2.11). Of the very large agencies (those with 26 or more total staff), only one out of nine did not grow. TABLE 2.11 GROWTH RATES BY SIZE OF AGENCY IN WEST CENTRAL STATE (In Percentages) Growth Small,b Large,c Ratea,o. No 82.4 56.53 Low 0.0 40.6 High 17.6 3.1 (N) (34) (32) aSee Table 3.9 for definitions of these categories. bl-10 total staff cOver 10 total staff Both of these terms are relative. dExcluding nonresponses. Table 2.11 also gives us an opportunity to see the effects of size on the growth rate when there is some growth. The difference between public assistance agencies and child welfare agencies in Eastern State was hypothesized to be due to the smaller agencies in the latter group. That this may be so is supported by the fact that where there is growth among the West Central agencies, the small agencies report only large growth while only 5 percent of the large agencies show large amount of growth. Again this is an artifact of measurement. The addition of any workers or other professionals to a small agency 82

will create a large relative percentage gain, while this is not so for large agencies. A probable explanation for the fact that small agencies are not as likely as large agencies to grow is that the larger agencies are in more urban areas in which most of the population growth as well as the growth in the numbers of people in need of assistance is concentrated. This is an important point to keep in mind when analyzing the relationship of growth to various organizational patterns and strains, as well as orientations of workers and administrators in the agencies. The effects of the strains caused by growth or its lack must be carefully distinguished from the effects of urbanization on the same variables. Bureaucratic vs. Professional Promotion Criteria The respondents were asked to rate their agencies on a set of criteria for promotion in the agency (Question 7, Appendix 2.A). Each of the criteria were rated on a six-point scale from unimportant to very important. The following discussion examines the responses to two of these criterion, "securing M.S.W." and "competitive examinations." The former criteria is taken to be an indicator of the professionalism of the agency; i.e., the extent to which professional ideals embodied in the training secured in seeking an M.S.W., are considered important in the agency. The latter criterion, competitive examinations,reflect adherence to objective performance criteria for promotion. States whose agencies are more likely to consider securing an M.S.W. as important are also less likely to consider competitive examinations important. Smaller agencies are less likely to rely on M. S.W.'s, but they are not more likely to rely on competitive examinations. While small agencies are quite likely to rely on competitive examinations, the largest agencies do so as well. Child welfare agencies are more likely than public assistance agencies in Eastern State to require M.S.W.'s for promotion, while the public assistance agencies are more likely to require success in competitive examinations. Table 2.12 shows that few of the agencies in West Central State (56.4%) consider securing an M.S.W. to be important, while at the other extreme nearly all the Western agencies (91. 4%) consider securing an MS.W. to be important. At the same time, competitive examinations are important in most of West Central's agencies (80. 5) while Central's agencies are the least likely to consider competitive examinations important (61.7%) and Eastern's public assistance agencies are the most likely to consider competitive examinations important (86.8%). 83

TABLE 2. 12 IMPORTANCE OF PROMOTION CRITERIA BY STATE (In Percentagesa) Promotion Western, West Central, Central, Eastern, Criteria _ go o % Securing M.S.W. 91.4 36.4 70.6 56.7 Competitive Examinations 78.2 80.5 61.7 86.8 (N)b (25) (69) (36) (40) aThese percentages refer to those agencies that scored 4-6 on the six-point scale rating criteria from unimportant (1) to very important (6). bWhile the total number of agencies in the survey is presented, the percentages are based only on the number who responded to each of these items; no more than three respondents failed to answer either of these items in each of the states. From the comments on their questionnaires, many of the administrators in West Central State show their awareness of the unusual tendency to rate securing an M.S.W. as unimportant for promotions in their agencies. The reason given or indicated is that their agency is small and provides little opportunity for advancement within the agency. One respondent simply indicated that M.S.W.'s were simply irrelevant. It may well be the case that the reward for securing a degree in West Central is promotion within the state welfare system rather than the local agency. Information obtained independently from staff members in West Central supports this explanation that upon securing their M.S.W.'s welfare workers return to the state office rather than to the local agency from which they have come. The relevance of the size of the agency to the use of M.S.W.'s as a criterion in promotion can be directly tested in West Central State. Table 2.13 indicates that smaller agencies are far less likely to consider securing an M.S.W. as a criterion in promotion. Only 12 percent of the smaller agencies and as many as 60 percent of the larger agencies consider M.S.W.'s as important in promotions. 84

TABLE 2.13 IMPORTANCE OF PROMOTION CRITERIA BY SIZE OF AGENCY (In Percentages) Promotion Small,a Large,a Criteria.% Securing M.S.W. 12.1 60.6 Competitive Examinations 73.5 87.9 (N)1b (34) (35) Small indicateles ten or less total staff; large indicates 11 or more total staff. bNonresponses have been excluded from the calculation of percentages, but not from the numbers presented. Two respondents did not answer the competitive examination item and three did not answer the M.S.W. item. Table 2.15 also indicates a slight tendency for larger agencies to be more likely to consider competitive examinations as important in determining promotions. However, a further analysis presented in Table 2.14, distinguishing the TABLE 2.14 IMPORTANCE OF COMPETITIVE EXAMINATIONS BY SIZE OF AGENCY WITH MIDDLE GROUP CONSIDERED SEPARATELY (In Percentages) Smallest,a Middle,a Largest,a Importance of Competitive Examinations 100.0 74.5 100.0 (N)b (6) (51) (10) aThe smallest agencies have 1-5 total staff; the middle sized agencies have 6-25 total staff; and the largest agencies have 26 or more total staff. bExcluding nonresponses. 85

middle group from the rest, shows both the very large and the very small agencies all considering competitive examinations to be important. Only the middle group included agencies which considered this unimportant. The extreme groups may be more "bureaucratic" for different reasons. Large agencies dealing with a large number of cases may become bureaucratic as a means for handling these large numbers. Small agencies which cannot use professional criteria such as securing an M.S.W. as the basis for promotion must resort to more bureaucratic measures of performance. In Eastern State, child welfare agencies are more likely to consider securing an M.S.W. as important for promotions than public assistance agencies. In fact, welfare administrators are far more likely to actually have M.S.W.'s if they are in child welfare agencies as opposed to public assistance agencies. Table 2.15 shows 70 percent of the child welfare agencies to consider M.S.W.'s as an important criterion in promotion and only 56 percent of the public assistance agencies to consider this important. At the same time, only 17 percent of the public assistance agencies have people with M.S.W.'s in positions of administrative responsibility and 46 percent of the child welfare agencies have M.S.W.'s as administrators. Apparently, not only are child welfare agencies more likely to state that the professional is the ideal worker but they are more likely to realize this ideal. TABLE 2.15 IMPORTANCE OF PROMOTION CRITERIA BY TYPE OF AGENCY IN EASTERN STATE (In Percentages) Promotion Public Assistance, Child Welfare, Criteria %. Securing M.S.W.a 56.7 70.5 Administrators with M.S.W.b 17.5 46.3 Competitive Examinations 86.5 55.6 ~~~~~a~~~~~~~~~~~~~5. (N)c (40) (42) From Question 7, Appendix 2.A. bFrom actual report of whether respondents have M.S.W. or not in Question 67, Appendix 2.A. Nonresponses are excluded from percentages, but not from totals presented. Nonresponses vary from 1-6 in child welfare. (Only one respondent in child welfare did not report his educational background.) 86

The higher degree of professionalism in child welfare agencies is accentuated by the fact that public assistance agencies are more likely to consider competitive examinations as important. Table 2.15 shows 86 percent of the public assistance agencies to indicate the importance of competitive examinations while only 55 percent of the child welfare agencies indicate competitive examinations to be an important criterion for promotion. Program Emphases MENTAL HEALTH AND ADDICTION Respondents were asked to indicate the percentage of the agency's effort spent on mental health and addiction as well as other areas of service (Question 9, Appendix 2.A). They were also asked what percentage of the agency's effort should ideally be spent in this area. West Central agencies were more likely to spend time and personnel on this area of service than any of the other states. Table 2.16 shows that almost half of the West Central agencies TABLE 2.16 ACTUAL VS. IDEAL AGENCY EFFORT ALLOCATED TO MENTAL HEALTH AND ADDICTION BY STATE (In Percentages) Western, West Central, Central, Eastern, Over 5% of Agency Effort on Mental Health axnd Addiction Actual 9.6 43.9 20.0 18.9 Ideal 38.1 77.1 50.0 29.7 (N)a (23) (69) (36) (40) aNumber of agencies responding; percentages exclude nonresponses to question about actual or ideal effort; there were a total of 9 nonresponses to the actual section of the question and 19 nonresponses to the ideal section. spend 5 or more percent of their effort in this area, and less than 30 percent of the Eastern public assistance agencies' respondents thought this to be ideal. One interesting finding is that a comparison of the distributions of actual and 87

ideal responses in all four states indicates that more administrators want to put a greater effort into mental health and addiction programs. This preference is important when one considers that it must be in addition to familycentered problems, child welfare services, financial aid, and all the other programs for which the agency may be responsible. CHILD WELFARE SERVICES VS. FAMILY-CENTERED PROBLEMS IN CHILD WELFARE AGENCIES The pattern of responses to questions about the amount of actual and ideal effort spent on family problems and child welfare service from Eastern State's child welfare agencies shows that they are doing relatively little on family problems and the administrators want to do even less. And at the same time they are doing a great deal on child welfare services and desire to do even more. Table 2.17 shows that 56.1 percent of the agencies reported no effort in servicing families and 48.4 percent expressed no preference for spending time in this area. On the other hand, 51.4 percent of the agencies indicated that they spend nearly all or all their effort (over 75%) on child welfare services and 75 percent indicated that spending all or nearly all their effort (over 75%) on child welfare services was ideal. TABLE 2. 17 ACTUAL VS. IDEAL AGENCY EFFORT ALLOCATED TO FAMILY-CENTERED PROBLEMS AND CHILD WELFARE SERVICES IN EASTERN STATE' S CHILD WELFARE AGENCIES (In Percentages) Amount of Effort Family-Centered Child Welfare Amount of Effort Problems, % o Services, % Actual None 36.1 0.0 Over 75% 5.6 51.4 Ideal None 48.4 0.0 Over 75% 12.9 75.0 Note: Percentages represent the proportion of respondents excluding nonresponses, indicating the amount of effort actually or ideally allocated to the designated program; e.g., 56.1% of the child welfare agencies allocate no effort to family-centered problems and none of these agencies allocate no effort to child welfare services. 88

Both the actual efforts and the ideals of the administrators contradict sound principles of treatment for children with difficulties. The child does not exist in a vacuum, but in a family that needs help. In fact, a child in need of help is often symptomatic of difficulties in the family unit. What makes these results particularly surprising is that the state manual instructs workers to consider the child's problems in the family context. Thus, state regulations and principles for effective child care are not realized by the agency, both in practice and in the ideals of the administrators. It, is possible, too, that ideology and belief systems of workers have come to play an important part in their expression of preferences. Problems in Analysis The findings to be presented and discussed here might as well be included under the section on methodology. These findings, however, provided us with substantive information about local public welfare organizations so it was decided to include them with the other survey findings. MISSING DATA The failure of a respondent to answer a question is considered to be "missing data." Missing data can occur because of the unwillingness to respond to a given question or the inability of the respondent to do so. Estimating the proportion of respondents who do not answer a particular question is a useful tool for evaluating the question both in terms of the inherent difficulty of responding and the situational factors that make responding difficult. Refusal to answer an attitudinal question may occur when a question taps a sensitive area. The inability to answer a straightforward question about organizational size and inputs reflects the lack of availability of such information. This lack of information may reflect inadequate record keeping, or more likely the unavailability of usability of information which has been recorded (buried in the files someplace). Thus, missing data rates become an excellent estimate of the adequacy of information processing in welfare agencies. It should be pointed out that these are underestimates and that there are probably more agencies whose records are poorly kept among those who refused to participate at all. In the four states studied as a whole, nearly all of the respondents were able to describe the size of their staff in detail (Question 4, Appendix 2A). However, there were many welfare administrators who did not have information on how many client applications were made in 1967 and even more (as much as three times more in one state) who did not have information on first time applications (Questions 2 and 3a, Appendix 2.A). Table 2.18 gives the missing data rates for the four states for total client applications and first time applications. 89

TABLE 2. 18 PERCENTAGE MISSING DATA FOR CLIENT APPLICATIONS AND FIRST TIME APPLICATIONS BY STATE (AND PROGRAM) ( In Percentages) Client Western, West Central.-, Central, Eastern Application % PA.,a % CW,b % Total 7.5 51.0 17.4 8.7 19.4 19.5 3 First Time 55.0 50.0 52.2 21.7 27.8 51.8 (N) (23) (69) (36) (83)~ (40) (42) apublic assistance agencies. bChild welfare agencies. COne agency could not be identified. West Central State stands out as having the best information available among the four states. This is particularly interesting because the same pattern was observed in the study of client careers (see Part I). Comparing public assistance with child welfare in Eastern State, it seems that record, keeping is better in the former. This can be explained by the need of eligibility workers to use explicit criteria in evaluating the needs of clients. This facilitates the development of accurate records. Service workers in child welfare tend to use less explicit criteria in evaluating the need for service, making record keeping a less necessary function. It should be pointed out here that these particular items were more appropriate for public assistance agencies than for child welfare because the latter may have multiple definitions of clients and thus may find it difficult to record the number of applications as we requested this information. Furthermore, public assistance agencies are subject to greater demands for explicit fiscal accountability vis-a-vis clients than child welfare agencies which are evaluated with multiple criteria regarding accountability. The child welfare agencies in Eastern State provide an excellent opportunity to test the effects of professionalism on the adequacy of information in welfare agencies. This type of agency is much more likely to have directors and/or supervisors with M.S.W. s and M.A.'s (the criterion for labelling the administrator as "professional"). Table 2.19 shows the effect of having well trained people in positions of responsibility. Professionals are far more likely to have the requested information than the nonprofessional administrators. This is perhaps because they realize that in order for clients to be properly served, information about clients must be gathered and maintained 9o

TABLE 2. 19 PERCENTAGE MISSING DATA FOR TOTAL CLIENT APPLICATIONS AND FIRST TIME APPLICATIONS BY PROFESSIONALISM AMONG CHILD WELFARE ADMINISTRATORS IN EASTERN STATE ( In Percentages) Client Professionals, Nonprofessionals, Total, Application.... Total 12.0 56.3 31.0 First Time 36.o 68.8 50.0 (N) (25) (16) (42)a aOne respondent did not report his educational background. over time. With an absence of the demand for explicit fiscal accountability vis-a-vis clients and the absence of professionalism of people in positions of responsibility, the information gathering and record keeping of an agency is very likely to be inadequate. In order to process and serve clients more effectively, the intelligent manipulation of the environment requires information about that environment, both in terms of resources and demands upon those resources. CASELOAD SIZE AND STRUCTURE Welfare administrators were asked about caseload size for various programs administered. This question provided insights into organizational differences among welfare agencies. Some agencies have "mixed caseloads" and others do not. In the former situation, a worker may be assigned some cases in more than one category of assistance. In the latter agency, there is total specialization among welfare workers by program. Although we do not have this information in detail, some respondents state that there is further specialization in many agencies, in that eligibility determination is separated from "service." Agencies differ between states and within states as to form of occupational structure within the agency. This makes the comparison of the caseload size effects on other aspects of agency organization and service difficult although a comparison of the effects of the different types of caseload structure is possible. In addition, it is possible to look at the effects of caseload size within each type of caseload structure. Table 2.20 describes the frequency distribution of mised caseloads in the four states. 91

TABLE 2.20 MIXED CASELOAD BY STATE ( In Percentages) Western, West Central, Central, Eastern,a Mixed Caseload 41.2 25.5 29.0 57.1 (N)b (17) (55) (31) (35) bExcluding nonresponses. The maximum allowable caseload for AFDC in order to receive matching federal funds is 60 per worker. While the demand for this type of service is inevitably variable, Western State agencies are very likely to report exactly 60 cases per worker. This suggests that the demand may be greater than the service available. Although many agencies report less than the federal maximum caseload per worker, these agencies may have mixed caseloads which would account for the lower number of AFDC cases per worker (Table 2.21). TABLE 2.21 AFDC CASELOAD SIZE BY STATE (In Percentages) AFDC C-aselad Western, West Central, Central, Eastern, Caseload % %' % Size < 60 13.0 68.4 80.7 74.3 60 87.0 8.8 16.1 14.3 > 60 0.0 22.8 3.2 11.4 (N)a (23) (57) (31) (35) aExcluding missing data cases. From the researcher's point of view, one becomes skeptical when respondents are too consistent in their responses as they are in Western. This may mean a failure for the questions in the questionnaire to communicate to the respondents that the actual rather than the ideal or standard caseloads were desired. This 92

could hardly have been the case since the same questionnaire was sent to all the states and the other states' respondents were not as consistent. Thus it seems that some other factor is operating to account for the unusual pattern of responses in Western State. It is clear that the questionnaire has failed to elicit the real needs of the Western agencies as far as AFDC is concerned. Federal regulations through 1968 limited the size of AFDC caseloads to no more than 60 clients per worker if matching federal funds were to be obtained. The data in Table 2.4 suggest that this federal standard has had a substantial impact on reported caseload size. Federal regulations in 1969 removed the formal size definition providing quality of service is maintained. It will be interesting to observe the future impact of this change. As far as child welfare programs are concerned, standards varied from state-to-state (Table 2.22). The modal response to caseload size in Western was 50 per worker, in West Central it was 30 per worker, and in Eastern it was 60 per worker. These modal responses suggest different standards in the states. Central State has no clear modal response which suggests that standards are not enforced here. The majority of respondents in each of the other states who do not report the exact standard report above the standard number of cases. This suggests that need for service is above the standard of need that can be met by the resources available. Although the same federal regulations regarding caseload size applied to all the states, it was interesting to observe that three of the four states had consistently smaller caseloads for child welfare services than for AFDC or other programs. TABLE 2.22 CHILD WELFARE CASELOAD BY STATE (In Percentages) Child WeChild Western, West Central, Eastern, Welfare % Caseload % I Below Standard 9.1 21.8o 14.3 At Standard 63.6 21.8 57.1 Above Standard 27.3 56.4 28.6 (N)a (22) (55) (42) aExcluding nonresponses. The different standards in the states point to the differential needs and the difficulties which can be expected when uniform national standards exist for any category of assistance. Standards are usually lower than actual caseloads indicating that agencies are not equipped with sufficient resources both 95

to serve clients well and serve all those that need help. It is also apparent under these circumstances that it may be difficult to obtain accurate reporting of average caseloads. Supplementary Data This concludes the written text describing and interpreting some of the preliminary findings. The following pages contain tables describing some additional data which has yet to be fully analyzed and interpreted. We have attempted to select variables which are likely to be more interesting to the reader and more productive when the analysis is complete. 2.3. METHODOLOGY It is not our intention here to detail the technical problems of conducting a mailed questionnaire survey, but by examining various problems and their solutions in conducting this research, it will be possible to illustrate the complexity of the structure of the local welfare system in this country. This part of the report will be divided into three sections: (1) the problem of deciding what to study; (2) the problem of collecting the data; and (3) the problem of processing the data. Survey Foci The structure of welfare organizations is so diversified in the different states and even occasionally within states, that there is a major problem in designing research that will permit comparative analysis. Local welfare agencies differ in the types of programs administered as well as in their jurisdictional responsibilities. Regional differences in resources, urbanization and industrialization, and attitudes toward welfare create environmental differences that must be taken into account in selecting a national sample of agencies. Individual differences among welfare administrators are demonstrated by questions which allow persons to express themselves freely. As a result much effort is required to create standard categories of responses to permit comparison of attitudes and orientations of welfare administrators. Problems in processing the data highlight two major structural differences in welfare agencies. There is tremendous diversity in their ability to construct adequate information processing systems and to structure caseloads for 94

0 0 (U CO ) — I 0 O' — CM O'I.. i~- - 0 o P Pk r! (C F0CC) co CM CM;E W 0 8 c O t- CC a) ) H -1) 0 o o z (U " OkO O l r O Q) (U CM (U 4'CO'0 *H +) 4 4- - I., o 00j c\ ON CC) P- CM (cc < ) P') r.-C Pn H O C _- 0 -- rMi o o c ( U. t~ o C o \ 0q r; P ~ -. > LN.N o - ~~~~~~~~~~~~~~~~~~< r-lH~~~~~~~~~~~~~~~~~~~~~~~~~~ 0 C O a ~ opr l \0 ^ 0 4 0 d:!U O4 o0 aat H 0 (DOH ~ (, -i - Co -, -, IO 0 - 0 0.) H 4-t 1) 0 0a) a) a tH -a Ca 0 a)(aH H ) #> toa)) Cd 0 4C q 4 H) b 0O *H tO -I +'C a)+ a) co.H rd H 0 I > -P CHC Cco O CH o O HO Ha) 0%i0 cctOO a~tocO O)+'+' X PO H-0 t H a) a* ) c0En z w to c cOO - d c c - C C) ) dO X) H 0o cd4a)d Q O ba) to N o 0) * H to H o H o t Al.H tO -Hd((I)Pt - a).H- C d 0bI VQ) C;rXa 0.H.H C U 0 - O) * 0+' O h to *H * H aP a)' r o d "C NH (Q o u$ a) *rp T- dag hH ObdOh 0 ) *-4 arl -P! a)H OH(U 4)td a) h -.HO d O O 4 HH (OH Od OQ O 0( O *H O4>a (-Q 0 tO P O O O O X F +1 E- 40CH E-1H*P H HPCS- E-4O to H0$ Oh O tO

TABLE 2.24 ADMINISTRATIONS' REPORT OF PROPORTION OF CLIENTS "COMPLETELY HELPED" BY STATE (In Percentages) Proportion Western, West Central, Central, Eastern Helped, %. %..o PA, % CW, % 0 0 2 4 7 8 1- 25 37 13 21 13 42 26 - 50 5 20 21 3 31 51 - 75 21 33 14 33 15 76- 100 37 33 39 43 4 (N)a (23) (69) (36) (40) (42) aNonresponses are excluded from percentages. TABLE 2.25 PROBLEM PRODUCING SITUATION FOR AFDC CLIENTS BY STATE (In Percentages) Western, West Central, Central, Eastern (PA), problem _ %______%__% Caseloads Too Large 65 60 53 83 Too Much Red Tape 100 90 89 90 Inadequate Training 87 58 58 75 Conflicting Directives 52 29 64 47 (N)a (23) (69) (36) (40) aNonresponses are excluded from percentages. 96

TABLE 2.26 IMPORTANCE OF PROMOTION CRITERIAa (Percentages by State) Promotion Western, West Central, Central, Eastern, % Criteria % PA CW Seniority 48 55 35 51 51 Securing M.S.W. 91 36 71 57 70 Competitive Examinations 78 81 62 87 56 Political Considerations 0 1 6 9 6 Demonstrated Competence 96 99 89 92 97 Colleague Judgment 76 55 37 43 28 (N)b (23) (69) (36) (40) (42) aFigures represent percentages of respondents who feel the criteria are important as opposed to unimportant. bNonresponses are excluded from percentages. TABLE 2.27 PEOPLE'S BELIEFS ABOUT AFDC MOTHERS BY STATE (Percent Who Answered "Yes" to Statement) People Believe Western, West Central, Central, Eastern, % AFDC Mothers Are: % % % PA CW Promiscuous 87 68 81 84 91 Responsible 13 36 22 29 12 Lazy 86 77 81 87 91 Exploiting Taxpayers 83 69 78 87 91 Deserving 26 44 19 34 12 Family-Minded 48 64 42 52 27 (N)a (23) (69) (36) (40) (42) Nonresponses are excluded from percentages. 97

TABLE 2.28 PERCENT OF ADMINISTRATORS WITH M.S.W. DEGREE BY STATE (In Percentages) Western West Central Central Eastern, o f% _ _ __ _ %_ %__ PA CW 26 7 22 17 46 (N)a (23) (69) (36) (40) (42) aNonresponses are excluded in percentages. TABLE 2.29 SOCIAL CLASS DISTRIBUTION OF ADMINISTRATORS' FATHERSa (Percent in Each Class by State) Western, West Central, Central, Eastern, o Social Class PA CW Lower 29 40 51b 26 31 Lower Middle 9 16 23 26 9 Middle 43 24 11 23 37 Upper Middle 19 15 14 14 14 Upper 0 7 0 11 9 (N)c (23) (69) (36) (40) (42) Social class determined by Duncan socioeconomic index for occupations. These figures are high because the Duncan index includes farm owners as well as tenants in the lower class. Nonresponses are excluded from percentages. 98

I 0- |\ ON Hr 0 C+ ) 0 H K' r1 OJ OJL\ OH prd - O ~S]~d U) rl *) LC 0' 00 LC\d..::4 < 0 H H S H H co 4 a- o D > F X uzua U)CO- Q z i < a) Ut ( ) H a) O] H O ( r-C d H H h- 0 thD <C g*H r V> H H IrN HOO Xi 0 H E +)4 CO U) *r 0 H- 0 CO VN 4-H U) > 4 U) ) U) Ut C) H ~:~.! O HCD *H # r d1 co ) *rla U) - P- ON) 4-) + -) A- CQ a) U' > C *H U *H ri Ura H - P O a) V W a C ) r O (O 0 U) PH d 0-l M ~ r-l cu oj 0'~ r- 1 -C tr\ ^0 QD H H 0 4) X *r Q H m c U)H * C C < aD *0 a * H o U a) ar a) HH a) 0 O v IlA C d W O z z @ # 0 ~ eD HOe W r M P)) ( U)r C) 0 O Oa )T3 T3 H > *H:la 0:Hrd 4*H o - O 3 q ~ O Q *Hi oP u, a) o( U a)

welfare workers. The first point has direct bearing on the adequacy of service to clients and the second directly affects the ability of welfare workers to do the best job possible by minimizing structural conflicts within the organization and thus serve clients better. Selecting the States for Intensive Study Deciding what to study first presents itself in selecting the unit of analysis for study. Since some local welfare agencies are organized on a county basis and others on a district basis, the decision had to be made as to whether to make the basic unit of analysis the county as such or the welfare agency as such. Whereas some states group counties into welfare districts administered by a single agency, others may contain some counties which have more than one district; e.g., King County, Washington (Seattle) is divided into four separate districts. It was decided to examine the problem from different angles for the intensive analysis of four states and for the national sample. In four states, ranging geographically from the eastern to the western part of the United States, all agencies in the state were surveyed. This fact avoids the problem of making an a priori decision on the unit of analysis here. In fact, it becomes possible to shift units of analysis as the needs of research arise. In the section discussing the processing of data, the method of moving from the organizational mode of analysis to the ecological mode is described. The former uses the local agency as the unit being studied, while the latter uses the county as the basic unit. While an organizational analysis is preferred when doing an internal analysis of the welfare structure in each state, ecological analysis is more suitable when the researcher is primarily interested in the various conditions of poverty and dependency in the counties and the welfare organizations available as an environmental factor to handle these problems. In other words, each county is faced with an organizational environment, i.e., a single welfare agency or many welfare agencies, designed to solve its problems whether or not that agency is also responsible for other counties as well. Problems such as motivating workers or eliminating "red tape" are organizational problems; problems such as reducing the poverty rolls and ensuring decent standards of life are ecological problems. An important part of the methodology of welfare research is the clear understanding of the type of problem the researcher wants to solve in order to determine how that research shall be done. Selecting the National Sample As far as the national sample is concerned, certain constraints introduced themselves in determining what the basic unit of analysis would be. While the 100

four states were selected for their variations in structural and operational patterns, it was decided that a stratified and representative national sample of local welfare administration should be obtained. This sample would permit us to comprehend fully the variable nature of the administration of welfare in all parts of the country. To study every agency in the country would be an impractical undertaking, at least at this stage of the analysis. Thus a random sample had to be drawn. But, welfare agencies are not truly comparable units throughout the country. Some administer single counties, while others administer part of a county or more than one county. Some agencies handle general assistance while others do not. In some states, child welfare is administered by a separate agency and in others this program is combined with others such as AFDC. Adequate sampling procedures require a clear and relatively fixed unit of analysis of selection. The structure of welfare agencies is itself a variable and in no way could be conceived as a stable unit of selection. This fact constrains the researcher who wishes to study welfare administration to seek some other basic unit for selection: the county. Except for New England, this is a basic unit of government and administration in this country. This fact makes studying organizational problems per se in welfare agencies more difficult than studying the effects of those organizations on the poverty and dependency structures of the various counties. It is impossible to take into account the number of counties represented by each local agency and thus through statistical means convert the unit of analysis of collection, the county, to the local agency itself. Thus, both organizational analysis and ecological analysis of the national sample are possible. Stratification of the sample was done according to the two major dimensions of difference in welfare administration in the United States: urbanrural, since total caseload and poverty levels may be quite different in the two types of counties; and regions, since values East and West, North and South are quite variable as well as resources and the nature of poverty and dependency are quite different in these different regions. Stratification insures that no regional group, whether urban or rural, shall be excluded from consideration. Each stratum, e.g., Mid-Atlantic urban, has an equal number of counties selected from it. The chance of selection between strata is different, but can be calculated. This calculation is a sample probability and allows the responses in the questionnaire to be "weighted" according to the chance of selection to gain a truly representative picture of welfare agencies in the United States. Initially, 180 counties were selected. Twenty were chosen from each of 9 regions of the United States listed in the County-City Data Book of the U. S. Census for 1962. Ten urban and ten rural counties were selected in each region. Thus, a total of 90 urban and 90 rural counties were selected. (A failure to stratify according to the urban-rural dimension would have produced only about 20 urban counties in the sample.) The definition of urban employed was that the county should be within a' SMSA as indicated in the County-City Data Book. Two problems arose with this definition, the first was in New England. Here, counties are defined as urban if they contain an SMSA, since in contrast to the rest of the nation, counties are larger than the town or city used as the unit for defining SMSA's, not smaller than SMSA's. The second problem is in New York, in which 101

four New York City counties are not listed as urban according to the definition. These were then included as urban. (The problem arises due to the fact that only one city contains five counties, an SMSA listing being limited to locations of cities.) Two-stage random selection was employed. The order of states within each region was randomized; then each state's urban counties were listed in that order. Next, a random number was selected to start the selection and 10 counties were then selected from the list, choosing every (N/10)th one. It was decided that if any (X +.5)th county were selected, the (X + l)th should be chosen. This process was repeated 18 times for urban and rural counties in each region. It was decided to select the next in the list should it turn out that a selected county did not provide welfare; e.g., Yellowstone National Park. It was also decided to include the response of a welfare director twice if his de-"partment were selected in the sample twice because he was responsible for two counties which happened to be selected; e.g., Richmond and Kings County for New York City. Occasionally, there are several counties organized into a larger unit such as a district welfare office, similar to the New York City situation, but where only one country is selected. In this case, there is only single weight given the response which is taken to represent the county selected only. In New England, where welfare is organized in a unit smaller than a county in some cases and different than a county in others, a third step was necessary in selecting a welfare agency for response. One agency was selected from each county selected in the sample where the town, which is smaller than the county, is the basis of local welfare organization; the largest town in each county was chosen to represent the county in Massachusetts. The reason for this was that the town is essentially the equivalent socio-political unit to the county in other states. By always picking the largest town, the probability of selection in the sample is unaltered and the greatest number of people are chosen to represent the county which is one way in which the demographic-ecological data is available. Matching districts with counties most closely approximated by the district was done in Connecticut. Alaska and Hawaii are excluded from the list of counties and have no chance of being selected. It was later decided to increase the size of the sample from 180 to 560 counties. Cluster sampling was used to draw the additional sample. The county immediately following a selected county was also selected, keeping in mind that if a given county was urban, the next chosen one should also be urban, and if the selected county were rural, the next chosen in the list should be rural. Since there are fewer than 20 urban Mountain states, only 10 were selected. It was decided to double weight the responses from these counties. Table 2.31 102

TABLE 2.531 SAMPLING RATES Strata No. of Sampling Rate Counties Original Adjusted New England-urban 26.3846.7592 New England-rural 41.2459.4878 Middle Atlantic-urban 57.1754.5508 Middle Atlantic-rural 93.1075.2150 South Atlantic —urban 71.1408.2816 South Atlantic-rural 514.0194.0589 East North Central-urban 65.1538.3076 East North Central-rural 371.0269.0539 East South Central-urban 20.5000 1.0000 East South Central-rural 344.0290.0581 West North Central-urban 31.3225.6451 West North Central-rural 588.0170.0340 West South Central-urban 47.2127.4255 West South Central-rural 423.0236.0472 Mountain-urban 17.5882.5882a Mountain-rural 261b.0383.0766 Pacific —urban 26.3846.7692 Pacific-rural 107.0934.1869 Total-urban 360.2500.4722a Total-rural 2742b.0328.o656 Total 3102b.0580.1128 The original sample was not doubled through clustering. Three "counties," all Yellowstone National Park, were excluded from having a chance at being selected. 103

presents the sampling rates for the eighteen different strata used as a basis for selection. Collecting the Data There are two alternative approaches to conducting a questionnaire survey. The first is to conduct face-to-face interviews, which is practical only when the respondents are located near one another. While this procedure tends to produce the higher number of returns, the impracticality of such an approach is evident when studying agencies distributed across a whole continent. This research problem like so many other research problems has its analogy in problems of administration. Just as it is difficult to administer a questionnaire to geographically dispersed individuals and organizations, it is difficult to administer the welfare system itself for the same reason. The problem of administration and research on public agencies is not unique to welfare, but is the case for the courts, police departments, and the schools. (For an extreme example, the failure of the government to enforce desegregation f the schools is an administrative problem not in small measure due to the difficulty of administering an educational system that is fragmented and dispersed geographically.) The more difficult it is for a central agency to be physically present in its local branch, the less able it is to regulate its behavior. The second approach, and the one followed in this survey, is the mailed questionnaire. The result was that a remarkably high proportion of respondents were willing to participate in the study. While the national survey is not complete, the expected return rate once completed (based on a projection of the almost complete returns thus far) is above 60 percent. As far as the four states are concerned as many as 79 percent have responded in one state, while the lowest number to respond was only 50 percent. The return rate for this survey is well above the usual return rate for mailed questionnaire and is similar to that of the average interview survey. This fact, and the fact that 74 percent of the respondents wished to obtain feedback from the study indicates a strong interest among welfare administrators in contributing to and learning about organizational and administrative problems in public welfare. There is a deep rooted concern and need for information in public welfare agencies.* *See Appendix 2.D for a discussion of respondent rates and requests for information. The high rate of participation by welfare administrators is contrasted by many complaints by the same people who participated about the time it took to do it. Yet, it was not uncommon for a respondent to mention that although the questionnaire took a great deal of time and effort to complete, he hoped that his doing so would be of help. In many cases, particularly in the metropolitan agencies, the directors themselves were unable to find the time to answer the questions and instead an assistant or supervisor did it. It is clear from the comments and returns that the majority of administrators are concerned individuals trying to their best with limited resources. 104

Differences in return rates between states and regions of the country become important data on levels of professional concern for the quality of administration and the service of clients. It is hardly coincidental that in conservative states there were more refusals to participate while states with more progressive welfare laws and professionally trained individuals administrating welfare were more likely to respond to our request for answering the questions. STRUCTURAL COMPLEXITY AND SUPPLEMENTARY INFORMATION The complexity of welfare administration in America became evident when it was necessary to devise a mechanism for collecting data in counties administered on a regional or district basis rather than on a county basis. We wished to obtain information for each of the counties selected in the national sample with respect to clients and staff even if they were a part of a larger welfare district, so that comparisons could be made with demographic data from the Census data, the City-County Data Book, and so forth. In many cases, the agencies did not have available information for the individual county but only for the district as a whole. A device soliciting information about data processing in the agency was designed to meet this problem (Appendix 2. B). Welfare administrators were asked first whether they had available information for the individual counties selected or just for the district as a whole. Of the counties selected in the national sample (excluding New England which is still in process) 21 percent are combined with some other county in a district or regional welfare office.* Of those counties which are districted, at least half do not have information on individual county clients and staff. Th'is fact suggests great organizational problems for welfare agencies so organized. The lack of information on clients by county creates both internal structural problems and problems in relating to the political environment because resources are usually allocated on the basis of differential needs in different communities in order to maintain a balance between the needs of the locality and its requirement to meet those needs. A series of supplements were designed to handle structural complexity. In those counties where programs were administered by more than one agency, say Child Welfare vs. Public Assistance, both agencies were studied separately. This division into a multiple agency system in some parts of the United States provides an opportunity for two types of analysis. The first is a comparison of the effects on welfare administration of treating a different set of clients with different needs and different resources. The most notable difference that appears from preliminary study of some of the comments is the hostility of child welfare workers toward "welfare" as a concept and toward welfare workers as such. Over and over again, child welfare workers refused to acknowledge that they were providing a "welfare" service. This battle in semantics is simply *Twenty percent of these are child welfare agencies only. 105

the iceberg above the surface reflecting more serious problems beneath. Use of supplements also made it possible to compare the effectiveness and efficiency of single and multiple agency structures in servicing several programs. Because of the problems we encountered in attempting to comprehend local welfare structures, the problem was studied further and a working paper prepared which is included as the third part of this section. An important source book for ascertaining local welfare structure was the 1969 American Public Welfare Directory. It, however, was not always current because of state reorganization, or because incomplete information was provided about local welfare structures and staff. The difficulties which we encountered in designing procedures for collecting and processing the questionnaire data reflects at least some of the difficulties which can be expected in the designing of new or modified welfare programs when information systems are incomplete. Obviously there is an urgent need for up-to-date accurate information about operating patterns for decisionmaking at all levels of government. It appears to be particularly problematic at the federal level today where major substantive changes in welfare policies and programs are being considered. Operational Problems in Processing Data The effects of a complex welfare structure on the difficulties of deciding what to study and how to study it have been discussed. In this section, the actual technical procedures are described in detail. The purpose of this is not to bore the reader with substantively irrelevant material, but to show that a complex logically consistent information processing scheme is possible to construct in order to handle a complex environment. The first section deals with the structural complexity of welfare organizations and its relationship to organizational and ecological analysis. The second section deals with the functional differentiation of the welfare system, i.e., the division into program related agencies like Child Welfare and Public Assistance agencies in the same geopolitical area. Twenty-five percent of the counties in the national sample have separate child welfare agencies and all the counties in one of the four states studied have a separate child welfare agency. STRUCTURAL COMPLEXITY In the section on selecting the sample, the difference between organiza106

tional and ecological analysis was discussed. Through processing the questionnaires in order to represent fully the organizational complexity of the system, ecological analysis becomes possible. In other words, it becomes possible to compare the responses of welfare administrators who are responsible for units both larger and smaller than the basic demographic-ecological unit, the county, with data from that demographic-ecological unit. A coding procedure was devised such that the original questionnaire received an identifying number. Then, for every additional county represented by a questionnaire, a duplicate set of responses was produced with a number identifying it as a duplicate. In this way organizational data for every county was made available to correlate with demographic data for that county. In organizational analysis only the original responses need be considered without regard to whether the agency is responsible for one or more counties. When considering the relationship between the organizational data and the demographic data, both originals and duplicates must be considered. In other words, the organizational structure of the welfare agency in a county is part of the organizational environment of the county and enters into the ecological analysis. Another coding procedure was devised to record the number of counties the data represented. Many of the respondents were unable to furnish data for individual counties, although they did have data for their welfare districts. The number of counties in those districts was recorded in order to facilitate adjustments in the data where county information was required, but district data was given. It was also necessary to account for agencies that serviced part of a county, e.g., one county had a county welfare office and a city welfare office for the county's largest city. The most complete set of responses from one of the agencies was used to represent the county for the ecological analysis. All responses were coded as "originals" for the organizational analysis. Appendix 2.C gives in detail the procedures for handling the different types of structural differences by reference to questionnaire processing procedures. FUNCTIONAL D IFFERENTIATION One functional (program) difference that exists in the welfare system is the separation in some states and/or counties within the state of child welfare services from other public welfare services. The processing of this information, iQe., whether a response was from a child welfare agency or public assistance with or without child welfare programs, was treated as a problem in "identification." In other words, rather than create a separate code for type of organization, the type of organization was built right into the identification number for the questionnaire. 107

All public assistance organizations followed the 1962 County-City Data Book (U. S. Census) identification numbers for counties. (This was the county of the principal office when an agency was responsible for more than one county.) This was true whether or not that agency contained child welfare as one of its service categories. The identification of child welfare agencies was accomplished by adding "500" to the identification number. Thus a public assistance agency in a county with a child welfare agency in it would have number "001" while the latter had "501." The above procedure provided a unique means of identifying each agency regardless of program. ANALYSIS OF SENTENCE COMPLETION ITEMS The bulk of the questions asked in the questionnaire were "closed" questions. Respondents were given a limited number of alternatives from which to choose. While these questions were pretested in order to insure that the range of choices would adequately represent the possible situations for all welfare administrators, the survey gave some opportunity for the respondent to express for himself what problems were foremost in his mind and thus the chief problems in his agency. Each respondent was free to comment about each response in order to fully explain his position. Room was left at the back of the questionnaire with a full page for comments to encourage the respondent to express himself. And many respondents have made useful and illuminating suggestions on this page. Question 12 is a more formal attempt (see Appendix 2.A) to elicit the basic concerns of welfare administrators in personnel, client, resource, and environment related problems. The form of the question is the "open-ended," sentence completion type, and was designed to obtain the respondent's primary concern without any a priori attempt to pigeon-hole him. The questions asked were: 12a. I can't understand why so many social workers _ 12b. It's a shame that many clients ___ 12c. I can't understand why our county agency 12d. It's a shame that our financial resources 12e. I wonder why the government 12f. It's strange how people Once the respondent has had a chance to express himself freely, it is necessary for the researcher to create a classification of "coding" scheme which will make the large variety of responses manageable. The creation of such a code requires the search for common themes in the responses that tell 108

us that although the problems may differ somewhat, they are of the same general type. There are two approaches to code construction. The first is the "analytic" method which basically involves the researcher stating what themes he expects to recur throughout the responses and then seeing if they do. This approach was first tried with the responses to this questionnaire and it was found to be unsuitable. The diversity of the welfare departments and their organizational problems made it impossible to adequately list in advance both the thematic nature of the responses and the specific manifestations of those responses in the individual case. The second approach and one that worked with remarkable success is the "empirical" approach. This approach worked in part because it is much more consistent with the original purpose of the open-ended form of question. It allows the respondents to tell the researcher what themes are important. In other words, responses are grouped as much as possible according to the basic similarity of the actual content of the statements. Labels reflecting the thematic nature of the responses are attached only after the responses are categorized. For example, the theme "Policy of the government is too liberal" was constructed as a theme of responses to Questions 12e only after a sufficient number of responses reflecting this feeling had been expressed (Appendix 2,E). This theme is one that surely would not have been expected from welfare administrators on an a priori analytic basis. Appendix 2.E describes the open-ended codes for Question 12. The codes are organized by levels of generality. The arabic numerals represent the most general level, e.g., for Question 12a, the profession, the workers, or the work may be of primary concern to administrators. The next level of specificity is represented by capital letters, e.go, the administrator may be concerned by his workers? fears, apprehensions, or difficulties, or his attitudes. The next level is represented by Roman numerals, e.g., fears, etc., may be either present or (less frequently) absent. Still further in the direction of specificity is the small letter category represented by the type of fear: ideas-oriented or problem-oriented. And again, the most specific level in the codes is represented by lower case Roman numerals and following through with the same example would distinguish between personnel ("psychological") fears and client related fears. How many levels of specificity in a coding series will be required, will vary according to how general or detailed the respondents are with respect to a given theme. The numbers of level of generality-specificity indicates the extent to which there is concern in an area. More levels of specificity is due to greater detail in the response and thus greater concern. Thus, personnel would seem to be an area of most concern to administrators (Question 12a), while attitudes of the public toward welfare, etc. (Question 12f), are of least concern, Responses are classed together at the most specific level. Then classes 109

are grouped into more general classes and so on. An example of similar responses classed together are the two responses in 2.A.I.b.i "hang-up" over problem is personal. Becoming "agitated over small problems" and "hesitancy in making decisions" show individual problems that make coping with problems difficult. These are clearly similar responses reflecting the same thematic concern of administrators for the ability or lack of ability of workers to deal with the problems of others given their own difficulty in dealing with problems. RELIABILITY OF OPEN-ENDED CODES If the categorization of open-ended responses is adequate, there should be a high rate of agreement between two or more different coders using the coding categories. This rate of agreement is a measure of reliability which proved to be high for all the questions. Table 2.32 describes the reliability of the six codes measured in terms of the rate of agreement between the original coder and a check coder. These rates are high despite the rather large choice of alternate categories and the tremendous variability of the manifest content of the responses. Reliability was estimated twice, once for the most specific level of coding, and the second time for the more general capital letter level. TABLE 2.32 RELIABILITY OF CODES Question Level of Categ6rization Number General Specific 12a.79.79 12b.90.76 12c.83.79 12d.86.79 12e.72.69 12f.76.72 Note: These reliability rates are based on a checkcode of 29 responses drawn at random on a stratified basis at a rate of one in seven from the four states studied as a whole. It is hardly coincidental that concerns about clients and financial resources show the highest reliability. These responses showed greater uniformity and commonality of perspective. The most difficult response set to create a code for was the concern about the government which showed the greatest ambiguity and variability of content. 110

Findings from Analysis of Sentence Completion Items The preparation of the coding procedures for the sentence completion items was a complex and time-consuming task. Therefore, actual analysis of the findings have just begun. Some preliminary work has been done to construct indices for the entire set of items in Question 12. The strategy was to search for common themes which recur throughout a respondent's answers in order to categorize his main concern. Guttman-scale patterns were found in two sets of responses, indicating two basic thematic dimensions along which respondents answer. The first theme concerns a "difficult working environment." This may be expressed in various ways. Half of the respondents considered some of the attitudes of workers as problematic in response to Question 12a: "I can't understand why so many social workers ___." Some of these also complained about the attitudes of clients (in response to Question 12b). These respondents seem to think that certain characteristics of the people involved made getting the job done difficult. Unprofessional workers and clients lacking confidence in the agency were two factors mentioned by the respondents which hinder the effectiveness of the agency. Some of the respondents who complained about how workers and clients make things difficult also complained about the poor physical facilities.(in response to Question 12c). Respondents could be ordered on scale from no complaints at all to complaints about all three types of difficulties: worker attitudes, client attitudes, and poor facilities. The second theme could be called "desire to dO more for clients." Respondents indicated a great deal of concern by complaining about the lack of financial resources in reply to Question 12d, the agency doing an inadequate job in response to Question 12b, and the need for the agency to expand its operations to serve more clients and serve them better in response to Question 12c. The respondent who shows less concern about doing more for clients seems to be concerned about who controls the allocation of funds, staff problems, and client attitudes. There seems to be a negative relationship between these two thematic dimensions. Respondents who express concern about doing more for clients are less likely to be concerned about improving the current working environment. This pattern suggests that there may be two types of organizations. The first type consists of those so beset with problems of varying kinds that there is little time to consider expanding operations to serve clients better. The second type is a service-oriented agency in which the overriding concern of the administrators is providing more service, giving more financial aid, and serving more clients. It seems likely that freedom from the problems that beset the first type of agency is a necessary condition for the second type to exist. Nonprofessional workers and poor relationships with clients and funding agencies of government may not permit the agency to do the best job possible. Concern for doing more for clients shows at least the willingness to try to do the best job possible. 111

2.4. STRUCTURAL COMPLEXITY IN STATE PUBLIC WELFARE SYSTEMS In large and small communities across the United States the hub of social welfare programs and services is the local public welfare agency. Observations of many students of welfare have pointed to the importance of the local public welfare agency if one is to gain any understanding and appreciation of the overall welfare system. Unfortunately, there has been very little study of the local agency as an organization. In contrast, the literature on other human service organizations is filled with reports of studies of mental hospitals, prisons, public schools, general hospitals, and more recently of OEO and other community action agencies. It is not unlikely that public welfare agencies were viewed largely as serving residual functions for society and as such were of less interest to social scientists studying human service organizations. Studies by Francis and Stone and by Blau and Scott were directed at the local agency but they tended to focus only on one or two dimensions of organizational behavior.* The recent studies of the New York City Welfare Department by Lawrence Podell and his associates are extensive and enhance our knowledge of one agency, but all of the usual limitations of the case study approach still apply.** We do not know the extent to which their findings can be generalized even to comparable agencies in large metropolitan communities. Many researchers in federal and state agencies have also studied aspects of public welfare, but their reports have focused heavily on personal and social characteristics of clients and financial information rather than on the operation of the local agencies. As a result one finds it difficult to use the above findings to obtain a general "verstehen" of the local agency as an organization, or to understand the social environment in which this agency exists. It is certainly obvious that the local welfare agency is undergoing a period of rapid and significant social change because of many environmental constraints. Legislative hearings, investigations by several bodies, and public media reports highlight the rapidly changing situation. The federal government has and continues to modify laws and grants-in-aid programs with dramatic effects on state and local agencies. Welfare recipients are organizing and demanding that the agency function in new and different ways than it has in the past. And, last but not least, professional organizations and other local agencies are taking actions which have both direct and indirect effects upon local public welfare organizations. *See R. G. Francis and R. C. Stone, Service and Procedure in Bureaucracy. Minneapolis: University of Minn. Press, 1956; and P. Blau and R. Scott, Formal Organizations: A Comparative Approach. San Francisco: Chandler Press, 1962. **See Lawrence Podell and associates, Studies in Public Welfare. New York: Center for the Study of Urban Problems, City University of New York, 1969. A series of five reports are available on various aspects of the New York City Public Welfare Agency. 112

A paper was prepared in conjunction with a survey of local welfare administrators in 360 randomly selected counties of the United States, and a study of career patterns of child welfare wards in five states. In the larger study we became aware, as we began to delineate career patterns of child welfare wards, that local welfare agencies played crucial roles in labelling and processing children initially into the welfare system. Because we wished to determine how organizational characteristics interact with individual attributes to produce different client career patterns, it was important to have greater knowledge of the local welfare agency. Time and other resource limitations made a comprehensive and in-depth study prohibitive even if one were to concentrate only on a limited sample of agencies. It was decided, therefore, to complete a mailed survey of local welfare administrators, and to obtain other demographic and organizational data about each county so that some conceptualization of the environment and its constraints was possible. The initial plan was to confine the study to the five states for which career cohort data were being obtained. -Because of the interest reflected in responses and because of the many changes occurring throughout the country in public welfare, we decided to extend the survey to a stratified random sample survey of 360 counties. This paper was prepared as an outgrowth of the national survey as the research staff gained greater appreciation of the complexity of the federalstate-local welfare system. Demographic data about the counties is being correlated with the questionnaire data and both are analyzed with reference to longitudinal data from the selected states about career patterns. The questionnaire was sent to the local administrator for his response or that of one of his immediate supervisors. This instrument was designed to obtain information about the following dimensions: (1) Patterns of agency structure (2) Volume of client input (5) Agency goal priorities (4) Actual and perceived ideal agency resource allocation (5) Role description and expectations of welfare workers (6) Staff time allocations (7) Criteria for assessing agency performance (8) Inventory of administrators' perception of the attitudes of community residents relative to their own attitudes re agency-relevant topics (9) Synopsis of past, present, and perceived ideal agency policies (10) Personal background information about the administration 113

Questionnaires were distributed in all counties in four selected states where the longitudinal data about clients were also being studied. This phase of the study has been completed with good results. Return rates thus far range between 50 and 79 percent. Questionnaires were then sent to the national sample counties and a return rate of nearly 60 percent has been obtained.* Preliminary study of the questionnaire response suggests that local administrators across the country view this as a period of change, and a time in which they are confronted with many dilemmas and contradictory expectations. Marked similarities as well as differences are also apparent. It is hoped that findings from analysis of the data will enable us to develop meaningful typologies of welfare departments. To the surprise of the research staff, 74 percent of the respondents requested that a copy of the survey findings be sent to them. In addition, many wrote comments on the questionnaire and sent additional letters requesting information. It appears that local welfare administrators are seeking information about colleagues but presently have few, if any, mechanisms to obtain such information. Many appeared to be seeking a reference group with which they could identify and from whom they could obtain support for certain practices or attitudes which they maintained regarding welfare policies, programs, and/or clients. When the decision was made by our research staff to draw a national sample of county welfare departments, none of us fully appreciated the complexity of structure at the local level across the country. States differ markedly in their organizational patterns, types of specialization, and division of labor, centralization-decentralization patterns, forms of authority, and so forth. The 1969 American Public Welfare Directory was used extensively and provided the best information about state and local structures. This directory, however, was not wholly adequate because of continuing reorganization in the states, and because certain programs which vary within states were not adequately identified. A chart was prepared to depict some of the complexities of the statelocal structures in the various states. These findings are summarized in Chart VII for all of the states. The states are classified by region into two principal patterns, integrated and bifurcated welfare structures. A total of eight structural patterns were identified among all the states, excluding Hawaii and Alaska. These were then grouped together into two general categories. In those classified as integrated** all programs administered on the local level *This return rate is relatively high for a mailed questionnaire in which a voluntary response is solicited. Many studies are reported where the response rate is between 35-45 percent. **We recognize that there may be several "branch" offices in a given county or district. We are identifying as our integrated structure those systems where primary authority is assigned to one administrator who may have one or more offices through which services are dispensed. 114

CHART VII State Structural Patterns for Local Welfare Program Administration by Regionsa Integ:gaetds Welfare Programs Structures faurcated Wlfare ProeaeM StrutuE A B C D 1 F- G Gt County County Town or District County County District- County office all office all District or. offices- 1 office for 1 each G.A. admin P.A. Prog. P.A. Prqg. office all Regional- each for P.A. & for C.W. separately & G.A. admin thr P.A. Prog. 1 Dir. C.W. & District & P.A. admin thru 1 office for 2 or other P.A. C.W. 1 office more office countiesREGION all P.A. MT Arizona_ _ Colorado. Idaho X X Montana X Nevada X _X New Mexico k X _ __ Wyoming " x X _ _ _ Iowa _ X Kansas X Minnesota X Missouri Xs Nebraska X... N. Dakota X___ _ _____ S. Dakota a n i i __X ENC Illinoisa X K a e he Michigan sa ts-i t md pten Ohio Sak e Wisconsin _ om PACIFIC California _ _.. x 1._< Oregon Washingto n Delaware SC - _?.- Florida X _ _ Georgia X S.. - SS XMaryland X H a d N. Carolm"a5 X S. Carola X —-. Virginia X _ _, W. Virginia X__ New Jerse. New York X | Pennylvania _Y_ _X Arkansas I | __ l Louisiana Oklahoma.., s A e Malanchsietts l l i_ x Maine.X.X (towns) N.-ggga; e ________________X lU_______________ X (towns) Rhode LimaA X towns Vermont _ X a. See Glossary Appendix 2. - for definitions of abbreviation of programs. c. King County, Washington has four separate and autonomous offices; Kentucky and New York each have d. Kentucky has variable county welfare structures for P.. A.-program other than C.W.., e. Mixed patterns for G. A. appear to prevail in several New England states-with the modal patterns appearing to be a "town" administered programs. f. Alaska, Hawaii, and U.S. Territories are excluded from this comparison.

are administered through a single welfare office. This office may be a town, county, or district office depending upon governmental organization. Bifurcated structures refer to those states where there are two or more local offices each of which administers only specified programs. There may or may not be any formal linking mechanisms between these offices; in practice, many respondents reported relatively little contact between offices. Furthermore, some respondents explicitly disclaimed relationship to the other programs.* Observation of the chart reveals that in 19 states there is a single county or district agency through which are administered most of the federalstate-local public welfare programs. In eight states all programs, except general assistance are administered through a single local office. District offices serving two or more counties were observed in eight states, most of which were in the Far West where the population is sparse. Three states in the Northeast have town-district agencies, and throughout New England the county is almost irrelevant as an administrative unit for welfare programs. In nine states there are two or more local offices administering welfare services. For example, in Illinois, there are separate offices for children's services, for general assistance, and one for other public assistance programs. Although there is a trend toward town integration of general assistance within a larger local welfare agency, it is apparent that this goal is far from being realized. In at least 18 states general assistance is still administered separately in at least some parts of the states. Changes, however, are underway in most states. Our findings about the complexities of the structure were not directly related to the principal project objectives. These had to be understood, however, if we were to identify the appropriate local administrators to whom the questionnaire should be sent to obtain the needed data. The remaining part of this report details some of the information which we obtained thus far about the various structural patterns in the states. *See the report on survey findings earlier in this section. In Eastern State we noted that child welfare respondents often made explicit spontaneous comments to disclaim relationships with public assistance workers. Furthermore, in administering the survey we found it necessary to send many letters of clarification to child welfare administrators to assure them that we wanted to include them in the survey. 116

Integrated Structures STATES IN WHICH ALL LOCAL PUBLIC WELFARE PROGRAMS ARE ADMINISTERED THROUGH A SINGLE COUNTY OFFICE There are 1-9 states with one county director in an agency responsible for the categorical assistance programs: GA, OAA, AB, AD, ADC, and CW.* Some of these states administer additional programs such as MAP, MAA, CCS, and VRB. In Table 2.553, these states and the programs are identified. In the case of Nebraska, Kansas, and Arkansas, the 1969 Public Welfare Directory does not indicate the directors responsible for Child Welfare. However, Kansas informed us that child welfare was administered by the same director as other public assistance programs; the questionnaires from Nebraska appear to indicate that child welfare is likewise administered through the same agency as other programs. We still are not clear about how child welfare programs are administered in Arkansas, but it is assumed that one director is responsible for all programs, as a returned questionnaire indicates such and in other states where the same kind of information was given in the Directory with no child welfare list provided, this was the case. STATES IN WHICH THERE IS A SINGLE COUNTY OFFICE FOR ALL PA PROGRAMS, USUALLY EXCLUDING GENERAL ASSISTANCE Column B (Chart VII) includes those states where one director administers all of the categorical assistance programs: OAA, AB, AD, ADC, and CW, and in some cases other programs, through one agency. As shown in Table 2.34, some of the directors in each of four states administer general assistance programs but in four states, general assistance is administered through a separate agency. In the case of Mississippi, general assistance programs were not mentioned in the Directory nor in returned questionnaires. Also, the Directory did not indicate the agencies administering CW programs in Oklahoma, but a letter verified that one director is responsible for all programs in his county. STATES IN WHICH THE DIRECTOR IS RESPONSIBLE FOR A DISTRICT OR TWO OR MORE COUNTIES Columns C and D on the Chart VII and Table 2.35 present those states which have a director responsible for two or more counties and who is also responsible for GA programs, except Idaho and Nevada. Also included are three New England states (Col. C), Connecticut, Maine, and New Hampshire. The latter states have town-district organizations. Some of the areas are smaller than counties and some are larger. *See glossary of Program Abbreviations, p. 124. 117

TABLE 2.5533 STATES WITH A SINGLE COUNTY AGENCY ADMINISTERING ALL MAJOR PROGRAMS _____State _Programs O State AA OAA AB AD ADC MAP CWOther Arizona x x x x x x x(BS) Colorado x P x x x x Wyoming x x x x x x Kansas x x x x x x x Missouri x x x x x x x x(BP) Nebraska x x x x x x x North Dakota x x x x x x x x(CCS) Michigan x x x x x x x x(BS,MR,VRB) California x x x x x x x Alabama x x x x x x Delaware x x x x x x Maryland x x x x x x North Carolina x x x x x A x South Carolina x x x x x x Virginia x x x x x x+A x West Virginia x x x x x x x(DP) New York x x x x x x x Arkansas x x x x x A x Louisiana x x x x x x x 118

TABLE 2.534 STATES WITH SINGLE COUNTY AGENCIES FOR ALL PROGRAMS EXCEPT GENERAL ASSISTANCE Programs State GA OAA AB AD ADC MAP CW Other Some Iowa x x x x x x x Minnesota Some x x x x x x x(MA,MR) Indiana x x x x A x x(BS,DP,CCS) Some Wisconsin x x x x x x Mississippi x x x x x Tennessee x x x x x Some Georgia x x x x x x Oklahoma x x x x x TABLE 2.35 STATES WITH DISTRICT AND MULTIPLE-COUNTY ADMINISTERED PROGRAMS State Programs, State __ _____ GA OAA AB AD ADC CW MAPOther Multi-County Idaho x x x x x x BS,VRB Montana x x x x x x x Nevada x x x x x New Mexico x x x x x x x CCS Oregon x x x x x x A Utah x x x x x Washington x x x x x x x BS Town-District Connecticut x x x x x x Maine x x x x New Hampshire x x x x x 119

Column H is self-explanatory indicating where all or some GA programs are administered through separate agencies within the counties, districts, or towns in the different states. Three states have further structural raminfications other than those discussed elsewhere. King County in Seattle, Washington, has four separate agencies in different districts within the -county administering the same categorical assistance programs. Chichasaw County, Mississippi, likewise, has two separate welfare agencies, one is east and the other west. There are five counties (burroughs), which are included within the New York City Welfare Department. Also, in New York, there are five counties which have separate city offices responsible for the categorical assistance.programs. For example, included in our sample are Cayuga County and the city of Auburn. Bifurcated Structures STATES IN WHICH CHILD WELFARE IS ADMINISTERED SEPARATELY There are 10 states: South Dakota, Illinois, Ohio, Kentucky, New Jersey, Pennsylvania, Texas, Massachusetts, Rhode Island, and Vermont which have separate agencies solely for the administration of child welfare programs. The agency administering the other PA programs may or may not be responsible for general assistance (see Table 2.36). TABLE 2.36 STATES WITH BIFURCATED AGENCY STRUCTURES......ate__ _Programs Oh State _ - GA OAA AB AD ADC MAPOther South Dakota x x x x x Illinois x x x x x Ohio x x x x x No one Kentucky px x x x x pattern New Jersey x x x x A Pennsylvania x x x x x x BP Texas* Massachusetts x x x x x Rhode Island x x x x x Vermont x x x x x x *As the State of Texas is undergoing a major structural change, the programs and who administer them is still not completely clear. 120

Regarding child welfare, counties in two states, Ohio and Pennsylvania (Col. E), administer their programs on a single-county basis with an administrator responsible for programs only within that particular county. In Ohio 47 counties have separate child welfare agencies, and in the remaining counties the programs are integrated in a single county agency. In the other 5 states (Col. F) where there is a separate CW agency, one director is responsible for two or more counties. In those states which have the dual agency structure, the agency administering the other PA programs (except CW) within the county vary considerably. Table 2.36 illustrates the states which administer the different programs. There is only one state, Texas, which, in the majority of cases, has one director responsible for two or more counties in the administering of their PA programs. South Dakota and Kentucky vary to a great degree as to whether one director is responsible for only one county or for a number of them. In some counties programs are integrated in a single office in both of these states. To summarize, there are states which present little or no variation from the single county administrator who is responsible for all public welfare programs. From this point, the state structures vary from one director administering all programs, (but in a number of counties), to those states where a county has two or more agencies administering different programs in a number of counties. The jurisdiction of those agencies vary making it difficult for a client to receive all the services he might need, because there are multiple agencies in different geographic locations which serve similar or related client problems. Coordination among these several agencies is problematic at besto The prime example of the more complex welfare structure is Texas. Texas has been undergoing a radical structural change in the past year and a half, with some counties experiencing a change in August 1969. The present agency structure is ambiguous and the information we have received has been of little assistance in trying to understand the programs as they exist today. From the returned questionnaires it appears that some, but not all, of the regional boundaries have been changed from the list of regions used in sending out the questionnaires. Likewise, some regions suggest that the child welfare and public assistance programs are administered from the same office, and in other cases, they are administered separately. The Directory provided little information regarding the local welfare offices. However, using the only list available, the questionnaires were sent to the regional directors, one questionnaire to the CW director, and another to the director responsible for other PA programs. Chart VIII illustrates the counties within each regional office which were selected for our sample, plus the total number of counties within each region. Texas has a total of 250 counties. 121

CM CM 0 M -r \ ~0~ ~i o NO a0 C u W H H 0,4 PC) - - H ~~N0 N Mt N H I Tl T1I CM CM ri i H a ^ M N H 0 O?0 C r-0 H C V, i'o W 0 iy o N A c) E C Cu 0 (U 00 k _ _ _ _ _ _ _ _ _ __ _ _ _ _ _ _ to_120 ^ <U -- - - - - - - - - - - - - - - -- - -'-4 Mr5 ~~ ~~~ ~~ Ni CO__ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ __ ___ - f-i r-1 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ *P~~~~~~~~~~~~~ ~ NU N N

Observation of Chart VIII reveals that there are great variations in the number of counties per district. Furthermore, districts for child welfare services are not identical with the districts for other public assistance programs. Obviously this must create problems in planning and supervision within and between programs at state and regional levels as well as at local levels. Delivery of services to clients likewise cannot help being affected by these structural patterns for a mother who receives ADC may be assigned to one regional office for that program and to another regional office should one or more of her dependent children require child welfare services. As the national survey is completed and the data are studied, we hope to obtain findings which will enable us to develop a more complete and definite statement about structural complexity. We believe these findings are particularly important today where there are many proposals about partial federalization of the welfare system. Obviously any program which requires state and local governments to supplement federal grants must take into consideration extant local units. This could be finessed through the use of Social Security offices for administration of federal allotments, but it would mean that many clients would have still another office through which they must be processed and supervised. Problems in the delivery of services would also be affected in most situations. 123

Program Abbreviations The program abbreviations used in this paper are the ones most often used to identify the public welfare programs in the different states and localities throughout the United States and Canada. The descriptions of the programs are necessarily broad and general and there may be in the text several variations on each program name. As states and localities reach common agreement on new names to designate specific services, these will be added. AB - Aid to the blind* AD - Aid to the permanently and totally disabled* ADC - Aid to dependent children** BP - Blind pensions BS - Services for the blind CCS - Crippled children's services CP - Correctional program CW - Child welfare DP - Displaced persons ES - Employment service GA - General assistance MA - Mother's allowances MAA - Medical assistance for the aged* MAP - Medical assistance (Title XIX) MH - Mental health MR - Mental retardation MS - Merit system OAA - Old age assistance OAP - Old age pension Par - Parole service Pro - Probation service UC - Unemployment compensation (employment security) Vet - State veterans services VR - Vocational rehabilitation VRB - Vocational rehabilitation for the blind YA - Youth authority *AABD - Combined programs of assistance for aged, blind, and disabled. **AFDC - Aid and Services to Needy Families with Children. 124

PART III SOCIAL INDICATORS AND WELFARE STRUCTURES 125

3.1. INTRODUCTION To look at the careers of clients of public welfare systems, or the agencies which operate those systems is, in its own right, necessary research, and research which is being pursued within the context of this project. However, it is increasingly clear that persons and organizations do not exist "out there," but rather, exist within a framework of social structure, a social structure which influences and shapes, in both subtle and gross ways, the thrust, purpose, and behavior of units-persons and individuals-within the structure. Hence, it is desirable to reflect, and to attempt to detail aspects of the nature of the structure of that context as a means toward understanding how the organization functions in the environment. There are several ranges of contexts which are salient for organizations. Most obviously, there are the corporate and categoric interorganizational networks within which organizations are located, the former being the association of different parts of a subsystem, such as the manufacturers of autos and the makers of wheels, and the latter being the collectivity of similar organizations, such as an association of real estate brokers, etc.* Such aspects of the environment influence organizational behavior. It is important to note, though, that while these influences may originate within the physical territory served by an organization, they may also be extraterritorial, and in an age of modern communication and transportation, the transmission of influence is easy across long distance. More subtle than the influences exerted by other organizations is the quiet shaping of organizational purpose by the people who move in and out of it each day, whether as staff or, in the case of welfare organizations and other human service organizations, as clients. These influences tend to be located somewhat more concretely within the geographic locality in which the organizations exist. Perhaps the most subtle, and least well understood set of influences on the organization are those which stem almost directly from the local community context itself. Attitude of people in the community toward the mission and role of the organization are critical as are the resources of all types available to the organization within the context of its service area. This research project provides a unique opportunity to investigate the relationships between community characteristics, organizational structure, These concepts were first mentioned by Hawley. They are similar to Durkheim's mechanical and organic solidarity. Amos H. Hawley, Human Ecology (New York: Ronald, 1950). 127

and personal careers. The county will become, for operational purposes, the geosocial community, although we recognize at once that there are vertical inputs into this geosocial system from a variety of external sources. There are, of course, many sources of data on counties in the United States, and this richness represents a convenience in analysis. However, the fact that there is so much information on a county basis is not simply a collector's artifact. The 3,102 counties in this country are the social context within which much significant social life occurs. County institutions, close as they are to the grass roots, are among the first to experience changes in key aspects of local life, and become the first to feel the pressure of increasing social problems. In this respect, the county welfare department is a transactional organization, providing an output of money and services to local populations, and conversely, initiating many locals on people within and outside its boundaries. The purpose of the section, as follows, is to attempt to conceptualize the community structure. We shall suggest some ways of looking at and thinking about the community which are theoretically fruitful and salient in terms of the local social problem/social amelioration system. Two steps are required. First, it is necessary to identify crucial indicators of salient dimensions and operational measures. These clearly need to be variables about which we have valid and reliable information preferably at more than one point in time. The mere listing of key "factors" however, is not sufficient. We must carefully attempt to identify the social meaning of the variables, and not let that rest on "common understanding." For example, at the community level, the size of the community is always chosen, as we will choose it, and offered as important. Unfortunately, many analysts assume that the social meaning of size is not only clear, but also that it is uniformly understood. Neither of these two latter assumptions is valid. The social meaning of size must be investigated, and the implications of this investigation assessed in terms of the variable used in the analysis. After completing'these two tasks, we shall look at some of the approaches to community structure which stem from a more theoretical and functional perspective, and which, although thin on actual measures of an operational sort, do capture an important "sense" of the community, especially when one is considering it from the viewpoint of problem solving. Given a background about the geosocial matrix, we shall present a conceptualization of the formal organization, designed to permit the comparative analysis of human service organizations, among which is the local county welfare department. This scheme attempts to facilitate not only the systematic consideration of intraorganizational relationships, but, by breaking, or disaggregating, the organization from a holistic totality into subsystems with specified variables, permit a more detailed examination of the relations between the formal organization and the host community. 128

It is then appropriate to move to a specific consideration of the state systems under analysis, both for an appraisal of general social facts, and the presentation of specific welfare data. We conclude with an analysis of the interaction of rates and grants in welfare with selected social variables, as an illustration of the possible line of analysis. 3.2. COMMUNITY COMPOSITION-THE EMPIRICAL DEVELOPMENT OF COMMUNITY INDICATORS It has often been said that research alternates between empirical assessment on the one hand, and theoretical formulation on the other. Where one begins the process is largely a matter of choice and personal history, as long as one does not remain in that particular phase. Let us begin by considering the empirical development of community indicators, and then move to a discussion of the social context and meaning which each of them may have. The rather large amount of data developing on the local community throughout the 1950's and 1960's was providing a great source of richness for social investigators. Yet, it was also becoming increasingly difficult to use much of the data because of processing difficulties, slight differences in definition of units, and so forth. One of the first attempts to address this problem directly came from Professors Hadden and Borgatta in their volume, American Cities, Their Social Characteristics.*.From a variety of published sources, the authors selected 65 variables thought to be key indicators of various dimensions of community structure. These variables were factor analyzed, and considered by different size classes of city, as well as by "all cities," and "central cities." The results of these analyses were a set of 12 variables which Hadden and Borgatta felt described key dimensions of community structure. These were as follows: Total Population Median Income Percent Non-White Percent Foreign Born Density Percent Same House, 1955-60 Percent Population Increase, 1950-60 Percent Single Dwelling Units Median Age Percent Migrants Index #1: Deprivation Index Index 2: Educational Center Jeffrey K. Hadden and Edgar F. Borgatta, American Cities, Their Social Characteristics (Chicago: Rand McNally, 1965). 129

These variables provided a useful point of commencement for the development of an understanding of community structure. As reported in the book, however, they were simply an empirically derived list, and attached to no conceptual frame of reference. As such, it seems appropriate to consider them indicators of community composition, rather than social indicators, or aspects of social structure, because the relationship between the variables themselves and social structure needed to be detailed. A beginning attempt at this work is reported in a paper by Tropman, entitled "Critical Dimensions of Community Structure: A Re-Examination of the Hadden-Borgatta Findings*" (see Appendix 5.A). Four concepts are employed to account for the original 12 variables. The reorganization is shown in Figure 35.1. Indicator Variable (1) Community Size............ Total Population (2) Socio-Economic Class......... Median Income (3) Race................. Percent Non-White (4) Maturity/Growth........... Percent Foreign Born (a) Maturity Median Age Density Percent Same House, 1955-60 Percent Migrants (b) Growth............. Percent Population Growth 1950-60 Percent Single Dwelling Units Figure 3.1. Salient community characteristics (Tropman, Hadden-Borgatta). Basically, it seemed that Size, Class, Race, and Maturity/Growth were key concepts which could be used to describe the community. The variables clustered in a way as to suggest that there were several measures of maturity/growth, and a single measure of class. The four variables provided a point of departure. We had some confidence that these dimensions would prove to be salient aspects of community structure. Yet several cautions had to be borne in mind. First, the original work had an * John E. Tropman, "Critical Dimensions of Community Structure," Urban Affairs Quarterly (December, 1969). 130

empirical, not a conceptual thrust. Despite the "reexamination," there was work to be done in the area of understanding the social meaning of the variables. Secondly, our particular interest in welfare and welfare organizations, careers, rates and appropriations suggests that we should consider some data reflecting these concerns, something which Hadden and Borgatta did not include in their original 65. Thirdly, one must recognize that 65 variables are actually a very few. The decade of the 60's has within the ten-year period, brought a revolution in data processing and information systems. Limitations on computer processing (which probably dictated the number 65 for Hadden and Borgatta) are less pressing now. We have ourselves one data set with 1,045 variables for each county in the United States. Hence, one should keep in mind that the inclusion of more variables might produce quite different results. The first priority seemed to be the development of the "Social" aspects of these indicators. 3.3. THE SOCIAL MEANING OF SOCIAL INDICATORS One difficulty with modern social science, particularly given the new modes of data processing, is that we frequently have too much data, too much in the sense that we can comprehend meaningful analysis then. The process of developing a careful conceptual understanding of variables and their interrelation is often slighted, particularly when it is relatively easy to move to further and more extensive analysis. For these reasons, we have spent considerable time developing the conceptual meaning and understanding of the four variables or concepts about community structure, from a variety of perspectives. Work on two-size and class-has occupied most of the attention so far. Work on the others-race and maturity/growth-is in the hypothetical stage, and those beginning efforts will be reported. Social Stratification Let us define social stratification as the system of vertical differentiation of social units, in which each unit can be assigned a rank relative to all other units on the basis of one or more criterion. If two units have the same rank, then they are equal in the hierarchy of stratification. It seems that the position of one unit vis-4-vis another unit in the stratification system is of critical importance in understanding the ebb and flow of social action within the system. On an individual level at least, no social scientist today would not collect measures of "socio-economic status" or some analogue. Our thinking about the stratification variable suggests two hypotheses which are of some importance. First, there seems to be good reason to argue that the stratification system is multidimensional, rather than unidimensional. 131

Secondly, it seems reasonable and helpful to assume that social units other than persons-groups, organizations, and communities-have their own stratification system within which they can be assigned a rank. These two points lead to a third-that stratification is as important in interunit interaction as it is on an intraunit basis. For a number of years, social theorists have debated about what kinds of indicators are appropriate for social stratification. Various ones have been used, including education, income, and the like. These arguments have tended to conceal the fact that there were conceptual differences among the indicators, and that these indicators did not relate to- each others all that well.* Given this situation, it seemed reasonable to propose an alternative hypothesis, viz., that instead of there being a single system with different indicators, there was a system of several dimensions, which, to a degree, were interchangeable. It seemed that the salient dimensions were five- money, occupation, information, status, and power. These came from the original "class, status, and power," of Weber, plus the addition of an occupational variable (Blau and Duncan) and one of information (Svalastoga).** If the stratification system is multidimensional, and the dimensions are theoretically distinct, this explains why the various indicators account for 40-60 percent of the variance in each other. This thinking is reflected in the working paper, "The American Stratification System" (Appendix 3.B). We therefore hypothesize that the American Stratification system is multidimensional, with the amount of intersect between any two dimensions yet to be determined. The multidimensionality hypothesis also brings into new focus the work of Lenski on "status crystalization. "*** If one assumes that there are five dimensions, and a rank can be assigned to the appropriate unit on each dimension, then it is appropriate to ask about the degree to which the ranks are the same across dimensions (rank crystalization) or different (rank inconsistency). We have hypothesized that the degree of rank crystalization is inverse to the rank itself. Hence, a unit which was very low on one dimension would be likely low on another, but the converse would not be true. Therefore, one finds units characterized by the absence of money, information, status, power, and occupational position much more easily than one finds units characterized by the presence of high degrees of all of them. This situation, called stratification deprivation (or poverty'), has some important intervention implications, which are discussed in the working paper, "Social Stratification and Social Intervention," (see Appendix 35C). For example, median income and median education correlate.58 (N = 644 cities). Source: Hadden and Borgatta, op. cit. Peter Blau and Otis Dudley Duncan, The American Occupational Structure (New York: John Wiley, 1968) and Kaare Svalastoga, Social Differentiation (New York: McKay, 1965). Gerhard Lenski, Power and Privilege (New York: McGraw Hill, 1966). 132

Whatever the number of dimensions, from 1 to N, one needs also to consider which units are stratified. Typically, we think of individuals as stratified, and as thus having a locus within the stratification system. However, our observations suggest that groups, formal organizations and communities as social units have positions within the American stratification system as well. Further, as social units, they undergo mobility up or down within the social system of stratification. It is important in understanding the problems and capabilities of groups, organizations, and communities to understand their position and mobility within the stratification system. For example, private social welfare agencies are often condemned for "leaving the poor," but they have been following the accepted American course of "bettering oneself" under this interpretation (see Appendix 3.C). These few thoughts suggest at least the following: (1) Poverty, as stratification deprivation, is multidimensional. A program aimed at a single dimension is bound to fail. (2) Organization will seek to avoid lower class clientele. If they cannot "cool them out," they will seek to discourage them in a variety of ways. (53) As a community's position in the stratification system declines, its social problems increase, and its ability to deal with them decreases. Data from this project will permit examination of each of these hypotheses, plus some of their underlying assumptions. Race or Mobility Coagulation If stratification is important, then mobility within that system must be important as well. Our myths and ideologies suggest that there is an open system, that anyone may move up or down depending upon his abilities. Horatio Alger stories lend credance and support to the notion that you can start low and rise to the highest peaks. It is, however, becoming increasingly apparent that this is not so, that the social system is rather one of great fixity, with little chance for upward mobility. In the early history of the country, people were assumed to be able easily to do better than their fathers. Today, we hear much more discussion about the "Culture of Poverty." Implicit in this conceptual assertion is the assumption that one will not, and indeed cannot, do better than his father. There are, then, two competing myths-one suggesting mobility, and the other suggesting coagulation. 133

Provisionally, we shall argue that both myths are right, and that the explanation of this paradox lies in the different operation for different groups within the system, Blacks and Whites. Basically, we hypothesize that the mobility system does not operate well for the Black population. There is some evidence that this is so. Blau and Duncan, in their massive study on occupational mobility point out that there is great discrepancy between Negro educational achievement and job.* In fact, the better educated the Negro is, the greater the discrepancy between the average status for that educational level and his status. From a community structure viewpoint, then, the percent nonWhite or "Race" variable represents, we think, the approximate degree of mobility coagulation. This interpretation is felicitious because it provided a reasonable link between Race and Class variables which is always assumed, and never specified. Is it "race" or "class"? If one assumes that the system level meaning of race is mobility coagulation, then it is both. Coagulation does not mean that the mobility system is inoperative. It can, however, and does mean, we think, that the system works poorly. Some data to illustrate this point are displayed in Table 3.1. These data suggest that while there has been progress in the mobility of Negro sons, as compared with their fathers, contemporary Negro sons are about a generation behind, their own educational distribution being quite similar to that of the White fathers. It is clear that the system is not working as well for them as for others. If this assumption is correct, and, in certain communities there is a coagulation and perhaps strangulation of the mobility process, then one of the important aspects of social metabolism has failed to work, trapping people at the bottom end of the scale with very little hope of rising.** Such a situation not only presents the problems attendent to the blockage itself, but problems attendent to the lack of operation of the metabolic processes. Blau and Duncan, op. cit. Ch. 6. Strictly speaking, the process of coagulation could "trap" people at any level. One could discuss, with some merit, a coagulation at the top of the system, among certain elites. Nonetheless, our intent here is to call attention to the people stuck at the bottom of the social structure. 134

TABLE 3.1 PERCENT DISTRIBUTION BY EDUCATIONAL LEVEL OF MEN 20 TO 64 YEARS OLD AND THEIR FATHERS, BY COLOR: MARCH 1962* (Excludes cases with no report on education of the father) Men Fathers Difference Ratio of Years of School Completed and Color (1) to (2) (1) (2) (3) (4) White 100.0 100.0 - 1.0 Less than 8 years 12.6 36.5 -23.9 0.3 Elementary 8 to high school 3 years 32.2 38.8 - 6.6 0.8 High school 4 years or more 55.2 24.7 +3505 2.2 High school 4 years 29.6 14.4 +15.2 2.1 College 1 or more years 25.6 10.3 +15.3 2.5 Non-White 100.0 100.0 - 1.0 Less than 8 years 36.8 63.4 -26.6 0.6 Elementary 8 to high school 3 years 34.6 25.2 + 9.4 1.4 High school 4 years or more 28.5 11.4 +17.1 2.5 High school 4 years 18.1 7.0 +11.1 2.0 College 1 or more years 10.4 4.4 + 6.0 2.4 Table E, "Educational Change in a Generation," Current Population Reports, Series P-20, No. 132, September 22, 1964. Bureau of the Census, U.S. Department of Commerce. There is an important point here for planners and social ameliorators, as well as sociologists. We seem to spend a lot of time dealing with the fact that in a particular area there may be many low income or poor people or people, in our terms, suffering from stratification deprivation. Programs give them money, clothing, etc. We spend very little time asking why the normal process of metabolism is not working. Our contention is that both sets of problems must be addressed. A similar situation exists, in the area of the participation of the poor on advisory and governing boards. No one asked why special efforts for this group should be necessary in a democracy. This question is not more intended to argue against the participation of the poor than the former one was intended to argue against forms of ameliorative help. However, in each case, had the question been posed, somewhat different types of remedial action might be taken, and they can be taken now. One of the first steps is to identify the impediments to the mobility process, and attempt to remove them. 135

Work in the area of mobility here has just begun. In one paper, "Social Mobility and Marital Instability,"* we have suggested that the greater the degree of marital instability, the less will be the mobility. Put a little differently, marital instability and breakup become an impediment to the mobility process. To the extent this obtains in certain communities more than others, those communities will have a somewhat greater degree of mobility coagulation. A second factor, upon which we have just begun to do some exploration, is the problem of young age at marriage, or as it is defined here, adolescent parenthood. The issue is not one of illegitimacy. It appears that the young family is simply unable, in economic terms, to care for children. Often, the problems in young marriages are masked by reliance on parental resources, but this arrangement is without sanction in our culture as a permanent solution, and sooner or later the young family must "make it" on its own. The problems of young age at marriage show up in the positive relationship between early assumption of the marital role and subsequent family breakup. However, we believe that there is a serious, and unattended, social problem here, and thus frame the following hypotheses: The younger the person when he (or she) assumes the role of parent, the less intergenerational or intragenerational mobility will the person experience; the younger the person when he or she assumes the role of parent, the more likely is that person to experience some type of "social problem"; the younger a person is when he or she assumes the role of parent, the more likely it will be that the first and subsequent children will suffer from social problems. These hypotheses require testing in empirical research. Nonetheless, our preliminary bibliographic review, consultation with experienced family therapists and the relationship between marital instability and social mobility suggest strongly that findings will be in accord with the assumptions. If that is so, then a program to study and help the adolescent parent (and adolescent family) becomes imperative. The opportunity for mobility, then, is seen as key to the operation of the social system. As the operations of the processes which support and promote mobility, congeal and coagulate, system metabolism increasingly fails, and persons become trapped in low positions. Percent non-White is, perhaps, an operational indicator that the process of coagulation is occurring, but not the cause of that process. We must look elsewhere, including into the family structure, the degree of early marriage, etc.. John E. Tropman, "Social Mobility and Marital Instability," unpublished paper, The University of Michigan, School of Social Work, August, 1969. 136

We should also look at the institutions-schools, shops and agencieswhich sponsor and work with the mobility process. They may be failing in their function, something the Negro population refers to as institutional racism. Community Complexity One dimension of social affairs which is frequently discussed is the idea of complexity. Some persons and processes are alleged to be very complex, while others are thought to be relatively "simple." Often, what seems to be implied is the number of different dimensions along which variation occurs. Something which has relatively few dimensions, and limited variation along them, is simple. Something which has many dimensions, or many possibilities for variation or both, is complex. Wars, for example, are sometimes described this way. Where there is a clear line of demarcation, each side has its own distinct uniform, and there is little confusion about who is a "combatant" and who is not, the war is "simple." When there is no clear line, when the soldiers are indistinguishable from the civilian population, and when people flow back and forth between the status of civilian and soldier, then the war is complicated. Though it is not much discussed, the complexity of community systems seems to be an important variable, especially for the people and organizations who live in a particular community. They are the ones who have to deal with and process the complex series of interactions as regards their own personal life course. Then, at the level of the community or system itself, the question of complexity becomes important from a management viewpoint. Can the system be managed? At the community level, some people are saying that it is simply too complicated, that certain special systems, like New York City, for example, cannot be managed, either as a total system, or by the people within it. At the organizational level, complexity often manifests itself in complaints about "red tape," and "bureaucracy" or getting the "bureaucratic run around." The concept of complexity appears to be a shorthand way of referring to two important social processes, which might or might. not occur together-differentiation and integration. A complex system is a differentiated one, which as we noted, has many dimensions and/or many types of variation along the dimensions. Differentiation refers to the spread of the system, to its extensivity and scope, but does not imply that the system necessarily has any unity. That is provided by the integration of the differentiated parts and elements of the system. Without attendant integration, a system can differentiate only so far before it collapses. However, it appears that differentiated systems can function, albeit marginally, with only moderate or even low integration. One illustration of this situation appears in looking at some Wall Street stock brokerage houses. The degree of paper work, and the differentiation of the firm (in the organizational literature and business, 157

it is often called diversification, although this does not mean quite the same thing)* has left some firms actually bankrupt. It was remarked in this regard that several firms of this sort are broke, if they only knew it. The point is that the firm, or other system, can proceed along for quite awhile with marginal integration. It will certainly come to light in the case of a crisis, but may otherwise not come to light at all, or slowly. Many of the complaints about the urban system, the cities, and the municipal institutions are simply that the city, as a system, is very over-differentiated in respect to the amount of integration. Given that complexity is an important dimension of systems, we should like to pose two questions of importance here. Firstly, how can it be measured, or indicated, and, secondly, what effect does it have on its constituent units? This latter question deals with the interorganizational relations among units in the system, or between the system itself and units, such as persons and organizations, within it. The measure of system complexity is not easy because of the aspects of differentiation and integration contained within it, but with some qualifications the measure employed is system size. At the community level, it is the total population of the area. System size has been a fertile variable in almost every correlation analysis. It is one of the most common types of control variables used by social analysis, yet the meaning, in social terms, has remained obscure. Why, theoretically, should "things" be any different just because they are bigger, or smaller? One reason, of course, is the simple mathematics of size itself. Add one additional person to a group of 10, and a minimum of 10 new relationship potentials have been added, to say nothing of the new possibilities for coalition formation, in groups of various sizes. There are some limitations and problems which need elucidation in analyzing size. Early treatments of size, by Durkheim, did not assume that size and differentiation were the same.** Size was a precondition for organic solidarity though, as Durkheim referred to it, it was called social density. For Durkheim, though, increments in size did not always result in the development of organic solidarity. Indeed, he specifically mentions the case of the "horde," a large group in which organic solidarity had not occurred. He is unfortunately less than specific about the conditions under which size increments led to societal reorganization and transformation from mechanical solidariity into organic or led to the development of a horde, which, today, might be called mass society. One problem which needs to be considered, then, *Differentiation refers to the number of different parts and elements and the amount of variation within each of them. Diversification tends to refer chiefly to the number of different parts and elements, but especially to the parts. ** Emile Durkheim, The Division of Labor in Society (Glencoe: The Free Press, 1950). 138

is the extent to which size defines differentiation, or is a necessary but not sufficient condition for differentiation. A second problem deals with the other component of complexity, integration. It seems less clear that size, as such, measures integration, or integrative efforts, especially when, as we have noted, there can be some wide ranges of integrative effort at rather similar levels of differentiation. Perhaps further investigation will suggest another measure, such as per capita public administration, or per capita governmental expense, to measure the effort that the system in question is making at integration, rather than letting size stand alone. A third problem relates to the measurement of size and complexity. Size is not itself complexity, but an indicator for it. Two issues arise here. One is that some large systems may be simple, as in the case of the horde, or in the case of some large organizations with relatively simple technology. A large shirt laundry, for example, may be in fact less complex than a small computer consultation firm. Hence, there can be deviant cases. As important is the likelihood that complexity is in fact a step function, and does not occur with even increments but rather, may occur sporadically, and, at certain pivotal points, change dramatically. This curve is suggested in Figure 3.2. Such behavior is not at all uncommon in social data. Education, for example, behaves in somewhat this fashion. A person can attend seven years of grade school, three years of high school or college, and finish everything but his doctoral dissertation. But that extra year, which brings the degree in.each case, makes a significant difference in life chances. Hence, for education the steps occur at the degree periods, rather than smoothly. In any case, this possibility needs to be carefully considered in using size as a measure of complexity, and becomes an investigation within its own right. Let us move to the second main area of concern, other than measurement, the problem of the articulation of units, taking into consideration the level of complexity. The point becomes of particular concern when one is considering the relationship between an organization and its community. Organizations can be complex or simple. Communities can be complex or simple. If an organization is simple, and the environment is complicated, then the organization will probably go out of business, or suffer, and pass away over time. This is the lament of "small business" in American communities today. There are "so many rules and forms" that they cannot function. The converse also appears reasonable. If an organization is very complex, and the community is simple, then the organization will tend to dominate the community. Historically, this obtained in the "company town" situation, but appears in other guises as well. A major industry or business dominates major decisions, without actually taking over the vital functions. Some university communities 139

a) E'l 0 a) N a) 4-,) o CH 0 0.i.3 4-, V cn~~~~~~~~~~~U (A~~~~~~~~-., m o~~~~~~~~~~~~~t r C,-I r4PN'^^C \<u ^s, ~ a \I -r 0-'l, —H ~,-IO \~~~~~~~~~~~~~~~~~C \0) H 0 I.!x oIdmoD 0~ \40

are this way as well, with the university virtually dominating municipal life. Some of these considerations have been explored in the working paper, entitled "An Interorganizational Theory of Differentiation.* In sum, we are using the variable of size of a system, defined as the total number of members, to indicate the level of complexity characterizing that system. There are, however, problems with this indicator, a few of which we have discussed, but it seems reasonable to proceed providing that we are cognizant of the problems. Change and Stabilization One of the most important questions asked about the modern community today is whether it can be "governed" at all. Countless panels and professors discuss the urban crisis with attitudes ranging from skeptical to dispairing. A common theme which runs through many of these discussions is the degree to which it is possible to change the local community. The most negative argue that communities are simply too structured and cannot change. Others argue that change is occurring all the time. Almost none of the discussions attach any empirical referent to their arguments. We believe that the variable labeled "maturity/growth" taps some of this dimension. Basically, it separates communities into those which are older, which have older citizens, which have more foreign born citizens, less geographical mobility, more apartment living, from those communities in which many of the homes are one family, which have younger citizens, which are gaining population, and which are generally expanding as opposed to contracting. At first, thinking of communities in this way may seem inappropriate. Communities are social units, like persons and organizations, and do, in fact, age. It is reasonable to assume that older units have different problems from newer, growing units, but these need not be greater in number or more serious. Unquestionably, each type of unit has difficulties and trials of its own. But it could be important to recognize the different types of problems each unit has. For example, it is likely that more mature units are suffering from general system deterioration, in terms of housing, roads, and even persons.** See John E. Tropman "An Interorganizational Theory of Differentiation," unpublished paper, The University of Michigan, School of Social Work, November, 1969. See Appendix 5.A, Table 4. 141

On the other hand, newer, growing communities have younger people, with problems more attended to and centered around youth, children and the like. This is certainly not to suggest that both types of problems do not exist in each type of community, but rather, that on the community level, the types of problems in their modal presentation will be different, and priority and thrust will have to be different. Thus one must identify how that difference is structured in relation to the total system. We are particularly interested here in the special problem of how a mature or a growing community affects the organization, and especially the welfare organizations, within it. The empirical relationships between the maturity/growth dimensions and some selected social characteristics is presented in Table 3.2. The data presented in Table 3.2 suggest that generally speaking, as community maturity increases, more money is spent for private welfare (this is the amount raised, for the local Community Chest/United Fund), on a total, or per capita basis. A rather similar result occurs if we look at public welfare expenditures. It seems, therefore, that the growing community is not putting its money into social welfare, either public or private. This may place a serious strain on the ability of the community to offer services. On the other hand, the younger, more mobile population may have less need for such service. The more mature community may have, and may recognize that it does have, greater needs, and this fact may be reflected in greater per capita allocations from the available funds. One other hypothesis is that the level of need is the same, but that the ways in which this "need" is manifested in the two communities is sufficiently different. In the growing community informal and other nonorganizational networks pick up the deal with the problem, much like neighbors coming to aid a family in times of disaster. In point of fact, two types of clarification are needed. First, under some acceptable definition of need, we must establish relative need levels for communities. This will permit a serious assessment of the integrity and appropriateness of community effort. However, the establishment of this criterion must be joined with some machinery for assessing, and codifying, differential need modalities. We must be able to empirically assess the degree to which, within the same level of need, modal variations result in quite different community problem-solving styles. Summary and Conclusions We have identified and conceptualized for dimensions of community structure-stratification, mobility coagulation, complexity, and change vs. stabilization. Perhaps the best way to pull them together is to propose a series of hypotheses about their interrelationships: 142

C)J H r'0() 4-' -P r. tko o, 0,,-;,, a) 51 -L \0 N ON I 0\ 0 Kt\ 0 I 1 o- ir\ 0 W'3 0 0) C -HO ON 0 0 I 0 C I I I \ 0 t )- ) -P r-A Ea 3 1 4-P 0 -1 I r- E1H 3 CO 0 ) L4) - )0 0'00.H,U T3 (U) - P-1 0 0dEa a b00 Q) ri ta *ri (1u cz f co'...PC C * OS4 f-C ) 0) 0 0 ) 0 4a PH C'- 0 N- 4-P-0-) P *-1 — r- \D0 0 0N t^- O CJ _: K'\'^ \.O 40 ZP 0 03 tico 0 -P-' - r-H 0 C 0 r-H r-H CM K(\ P- c.,r-H.OC CE l- a)i i I I I I I COa' — III I 1 1 1 1 (U O -p ^ ^ Pj (U oC_) 40 C.0 H -H ) ), ~~l Cs 0) CO- a) P0 0=C G.0 -P 0) 0-)-' 14 H0)~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~r - Li 41C 4')H Q I 0J o 0.0) 0.. ^>3'3M e3 4->^.-^^>s~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~rd$4z - - 0U 0 ri 0 0 -. *H'H ) CM HI _:J-C- - H _-:,' r-H CM C — OJ I 0 0$ 14i H o o 14H- ~~~~~~~ ~~ ~~~~~~~~~~~~~ ~~~~~~~C') CO H CMj 0 0 0 0 H C 0 0) r) ) Cd I~~~~II 0 40) rHO 4C,) Cro'-O 00'0~-< O O O 00 0)^1 U (U O l 0 HON HO.~~~~~~~~~~4) 4-01 000)-4:: r-q Cr\ H a) z -1 Q~~~~~~~~~~~~~~1 ) Q)E C O O. -~~ -'^^ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~'~~~~~~~~~. 0)c.I I t i 00). H1 0) 0) H 4-) 0) CO C)o. +''0 -P: \ -H o 4 o-' 0 o Ho)CM1-O~ 0)4) 0 )X- — UC -s * 031 ON )3 0U C — 0 — H CO OH 10 Cd 00 0 H CM CM 00 H HO ) -04) CO Pci U to o 3C > rq C,CO0toH OH I 4 00)H'H 4-, a') O W CO +'H0) d C4Ot1 0)00 M ^C ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ir-1.~~~~~~~~C)) 0) LCNP- l00 ZE-'~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 0,,,-I...) )ONC- -,.-4~ 0 HIV'0. CoH.,.. ~1; Pq u t> 0 4; 0 4o 3 r-la4+.)0 0, 0 0-H 0 <C'-0 0) H 0)0) 0 0 0 00oo 4CM 4 0 o *H., H r 4.) o-0 CMC)H14 0)0) 0)4)14 H 1 -PC) 00E0 *O CO'^ PCo0) H0 HrOOCM Hj- o r1 O r0)aH-l 14H-U 0 0 - H 00 0ri 4 0 HON-, 0 OH. H L0H 4-H Co Co0 0 O 4-'Hr- oOo-P0) H0)I PiO^ ( 0P^- 003O n- 0U H- H N H(3'0 M 000 C( 00 C-HH C) M )( M 0 0 H CM 0 H H ) 0 O I-14 CM -ICC) 0) *HH *H0)OH |0\ -P 0 C -.p 0 Ea P P4 O~~~~~~~~~~~~~-'kO' 0SH 00 0) H 0 >~ > r-q 0}4A P 0)0) -0 i 0 d 4-' 00 U L\ -q r 001 H\D 0 C.) 4 0 -4 ) ~40 ~ C.4 N+ Co 000) Cd ON x~~~~~~~~~~~~~~14)0'0 H 4o) Hco \. CY\ co0, 0 -'8o'8C - P riU i- Pc a0a)+-> i r0 )4-' OH-H0)4 ) O 0)0) 0) I-~~~~~~~~~~~~~~~~~~-'0)14 Co HO~~jC~ K\ CM 0 C C)o W cdco d 00~' - M0 0O8oC l c0)0)l? - p 00 LPq a) 4-H -P-H00 H 00) OHHP 104-0) -H r- C - 00 0S)V. 0 0D ) 0' > 4- N) - \M (\ 0 or 0 30 H C *HQ)O -H HOC C 0))O 0i- H- 0M 0C fH H CMO O 00 CM H.0H'0 0 -PH^ CO?E4ld ) I 4II I I 0 4i' CdQ-H =!3l?-iLr,\O a?0 0) 4) - H0C PE HS< -P0 O HC0 S=! 03El 0 ) Co ONH 0 -.~ c~,,'-, o o ~ o o ~ co.- ~=~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~QcoX E- & > -P<U ^H~} 3 i-,1og! / 6-4 O - o o ~,.r r-,r-,.,u,-., 0C d) 0) — )0) 0)q C) Co O.) H 0)14 HC M 4 - CM140- ptP-P+0)o H-) 0 0 0)-H 0 0 0 0~' 00) H ~' H 0 i Hz E, ~ r-4 P.~Z0 1 - O )' - P U ) 0 )L 0 ) Ha ) Ea O N 4 4 4 a 0) o?- C~j a)0) ).. o) o 0 Co ~E-A' 0'+ -f C -1 < aS4' 4 -) -.0 - C) O -HQ -ON\ CO 0d CM0 0 CM CM 0M H CM CM aCM C C) 0 4 1+ 4-| -+0) 4)0-0 <CO -H HH I I I Ill l U+(-4C)-H - 0)1- 0) 0 1-P~ EaMC~~~~~~~ <+-i(0 C~~~O''0 0Cd EH -H 4-' 0 0C~j -4 -4 t ~ j 0 r\ \C O -HO Ul El 0)C)Q)0 ()0 ~~A EQ a) o cd.,A cd a~~~~-i- PI c i O) 0O 4-H 4a O U EQ CU'-^(U~~~~~~~~~l> U S M~~~~~r-lc6.rI ~r M f0 0 CM-H + — p( a HOCo 4pi.0 4- H cO -H. H-)1c 4-'0) 0)i ^ 0 T-l0 U —HH +'.. -P 1 -'CMH 14E0) H 144) C) 00 1CCC 0C o:. 4- o o- o ~ D+: pqLCT-lcd-fc0)P 14 0i 0 0)0) C 1 ) 0 ) 1440 0'0 ONiP^\0 0 0<~\ -HM H-l i-l c- CM ~ ir\ \D CO +^E i H 0 (: *Eor - H O 4-' 0)0 H 14 C) H-l1 1 00 T- 0)(|4-H > 04)- H c4-4 0) 0)0) 0 0 0) 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Q co)' S40- to (), > H HO 0 00) 0-1 ~~~~~ ~~ ~~ ~~~~~~~~ ~~~~ ~ ~~~~~0) H t0 C 0)+''0 1400 P' 0 HOHO 0 C)C) 0 -H-P 0) 0 0) Co -H H 140)00) 01 00) 0)0)0)0)4) c0H~~~~~~~~~~~~Q)FH U pq 0 P CM 0 -H P-H 00 o 0 0 - ~) 0 -H.0' o \ 0 o 0 O4'D' C 0 *O 0) HH MO 0 F H- 00)0 U0 >C 0- c )0 01 I 0040 ) -P 14 00 CO 0))0 14 ) IO) H0 C)c0 0)COCM~ ~ ~ -H! 6 } cO ^O-H 0 I )H ) ) ) M ) — 00C) -H H ) 0>' Ho CO p +).-pP4r1l -P ~ ~ ~ ~ ~? 0'0 CM 00 H'0 0 C 0 COPHH0CO 0 00) 0) C)U 0)E 0)c'+0c'Au O 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~2 T0 cd to r-r-l-lD.;.H CC o O. w *! - P.C)0 I H -H 0) 0-|) 0 co P.,C)r0) - r r-1 r-10'Q CU0 14 C) 1 0 0 +3:J 0O 0)- H — --- -*H —-) f} K - 0 —-------— O —--------------- 14 1 NH 0CM C) —Pzt 4'ON CO CM C O HH~~lT 4 ON 00 LC-l~ g 0 0 0-H 0H 0)0 0)0 0 0 H H C C& 0 < 0 0L 0C 0 P- -r 0 f M S. i -l - Ea T ) - l:i^ l r0:Ui- (U5 K Cl-q cl+ 0 C-P~ CO — — O- 0 —- F CCOCOCOPCO CO 4 U COC) 0> HU CM - Ir01 *H (U (U o) 04 oCd ^ u u C) C-^ *H 4.3 O-. - ^- - ^ - ^ - -\^^ H^l ^ N r0 0 C 1-1 U - 4-0 O cM o I- 0 \ LTN3\d *rQ1 Ord 00 CdO O r 1 r4 1 0 CM 0 0 0 0 03 ^' CQ — ^'-' rt'- S'^' "^ -^ ^-^'-' -^ —' — ^ 3 03 03 0 < PQU O r-CM K^ -

(1) The lower the position of the community in the stratification system, the greater the degree of mobility coagulation; (2) the greater the degree of rank crystallization in the stratification system in the community, the greater the degree of mobility coagulation; (5) the fewer the dimensions along which system units are stratified, the simpler the system will be; (4) the lower the position of the community in the stratification system, the more likely it will be that the community is "mature" rather than "growing;" (5) the greater the mobility coagulation within the community system, the more likely is the system to be complex; (6) the greater the degree of mobility coagulation in the system, the more likely it is that the community is mature, rather than growing; (7) The more complex the system, the more likely the community is mature rather than growing. These hypotheses can be tested out using some of the community (county).level data now coming to be at our disposal. Basically, we believe that some of the most critical problems occur in one configuration of communities. In mature communities, with a high degree of mobility coagulation, where residents occupy the lower positions on the stratification scale, hopelessness about solving problems appears to be rising, and requires special attention. 3.4. THEORETICAL APPROACHES TO COMMUNITY STRUCTURE Beginning at least with Durkheim's use of suicide as a measure of the social integration and cohesion of the group before the turn of the century, and developing more recently with Angell's attempt to uncover the degree of moral integration of American cities, sociologists have attempted to develop theoretical constructs to describe communities.* Sociological interest has, however, been neutral from the viewpoint of intervention. That is, most of the typologizing on geopolitical areas does not relate to the community's E. Durkheim, Suicide (Glencoe: The Free Press, 1957); Robert C. Angell, "The Moral Integration of American Cities," in P. Hatt and A. J. Reiss, Jr., Cities and Society (Glencoe: The Free Press, 1955). 144

ability to "do" something, or tell outside people what might be done. It has had little action or policy-related utility. Those sociologists who had some interests in problem-related and policy-related activities worked in the area not of community analysis, but of social problems and disorganization. Elliott and Merrill, in their introduction to the 1933 volume, Social Disorganization, say that "American sociologists have traditionally presumed an interest in the practical considerations of social welfare.* The relationship between the sociologists and the policy maker, despite this "interest" was at least once removed. In the past decade, however, there has been a definite shift in focus. Beginning in the early sixties with some work by Leonard Cottrell, an attempt was made to conceptualize the community in ways which would articulate with social intervention. Cottrell had earlier developed the notion of personal "competence," or loosely, the ability of a person to manage his affairs and relationships in socially suitable and personally satisfying ways.** This concept was transmuted to the community level, as community competence, and was used to refer to the ability of the community to deal with its affairs. One result of a breakdown of competence, would be social problems. The presence of competence would not necessarily indicate the absence of problems, but the ability, on the part of the unit in question, to effectively deal with those problems. Prima facia, of course, the concepts make excellent sense. Many people have an intuitive perception that communities differ in their ability to handle problems and prospects. Presthus, for example, in his comparative community study shows that the two communities under investigation were appreciably different in their ability to handle overtures from industrial concerns.*** As a practical matter, the Cottrell concept informed those persons responsible for administering the President's Committee on Juvenile Delinquency and Youth Development. The grants made by that body to communitieswerewith the intent not of controlling delinquency per se, directly, but of developing community problem-processing capacity such that the community itself could handle the problem. Hopefully, this increased problem processing capacity would enable the community to deal more adequately with the full range of problems facing it. This effort, and its Mabel A. Elliott and Francis E. Merrill, Social Disorganization, (New York: Harper and Brothers, 1933), p. ix. Nelson H. Foote and Leonard S. Cottrell, Jr., Identity and Interpersonal Competence (Chicago: University of Chicago Press, 1955). Robert Presthus, Men At The Top (New York: Oxford, 1964). 145

sequalae, the OEO Program and the Model Cities Program, are just now being studied.*,** The idea of helping local communities develop the competence to help themselves is an appealing idea, providing this approach does not become simply a cover for not providing needed resources. In any case, whatever programs for community help are developed, ability on the local level to receive funds, to plan expenditures, to develop problem priorities to provide some direction and integrity to the local system is essential. On a personal level, this attribute is often called community leadership, although this term personalizes the entire concept in a way that is not appropriate. However, the wish to equate community competence with certain personal competencies of specific leaders is quite understandable, because there is really no good measure for the concept. This fact, however, does not mean that one should dismiss the concept, but rather attempt to provide some operational indicators for it. In a previous study, we used the per capita amount raised by the Community Chest/United Fund campaign, but this figure is available only for named communities, rather than counties. While data are available on the state basis, and could be calculated and reallocated to counties, this procedure would simply not capture intercommunity differences which may be small, but otherwise significant. It seems also that there is a perhaps subtle, but important difference between what a community could do, and what it does. Some communities nray do more, others less, than is necessary. And, as anyone will say, there are some good reasons why the gift level does not, In any case, represent community effort. Therefore, while one would not wish to exclude entirely the possibility of adapting this available data to a county base, it seems that some other measure would be appropriate. It will be recalled that, in the discussion of complexity, we noted that the complexity variable actually represented two types of concepts-differentiation and integration. Further, we noted that the amount of integration, while related, was anything but a direct function of the differentiation of the unit. Further, we argued that in some degree, the total population of a unit was more likely to be an indicator of differentiation than of integration. It seems possible that there is a connection here between the conRoger M. Lind and John E. Tropman, Delinquency Planning and County Competence, Unclassified Report, Office of Juvenile Delinquency and Youth Development, Dept. of HEW, 1969. Daniel P. Moynihan, Maximum Feasible Misunderstanding, (New York: Free Press 1969), Peter Marris and Marten Rein, The Dilemmas of Social Reform (New York: Atherton, 1967). 146

cept of system of integration and the concept of community competence. As one considers the meaning of competence, or the several possible meanings that it could have, the notion evoked by integration seems to fit well, vizo, the ability of a unit to hang together, to act with some integrity and purpose. For these reasons, it seems that the amount of effort a system spends on integration is a measure of its competenceo For certain students of community, this measure will be less than satisfactory. Some prefer to use the concept as referring to "style and flourish." Hence, it is often thought that a "Lindsay administration" is "competent" while a "Daley machine" is not. Once again, the issue tends to become personalized around "His Honor." There is no question that the Mayor makes a difference~* But community structure differs, too. An interesting distinction between New York City and Chicago is that while the former has sustained a vigorous "reform" party for many years, the latter never-has. In any case, the thrust of the effort is to secure an adequate measure of how well the community in question is able to deal with the issues facing it. Relating competence to community efforts at integration, however, does not provide a measurable variable, useful for comparative analysis. Hypothetically, it seems that the number of persons employed in public administration, per 1000, would be a useful measure. These are the persons, though not exclusively, who have the job of running the system. To the extent that more people are involved in this task, it is likely that a more adequate job will be done in the local communityo These persons, in a system sense, are the community's management. The more complicated the system, the more management is necessary, although the community does not always secure ito Several problems are presented in using this measure, however, which may ultimately mean. that it should be replaced0 Firstly, it pays no attention to the quality of staff, a critical factor in any management analysiso Secondly, a system may need more management, but simply be unable, realistically, to add any moreo This might well be the situation in some of our large cities. A city administration not only has limited funds but, and this is important, it typically cannot compete for top managerial talent with other systems which can pay more, and make a more pleasant set of surroundingso Thirdly, we have no guidelines, as yet, to appropriate ratioso Crudely assuming that more is better may suffice for awhile, but patently ignores the problem of "top heavy" management, among otherso It also ignores what may be a more serious problem in some communities where the political system may be such as to permit the employment of many more persons than are necessary to perform the required tasks. Allan R. Talbot, The Mayor's Game (New York: Harper and Row, 1967). 147

These or analogous problems, though, one might have with any indicator. What is critical is to begin to relate the theoretical approaches to salient features of community structure to some which have proved empirically fruitful. The concept of community competence is first, a heuristic one, but with some operational indicator, can be useful, both to describe communities, and as an analytic tool in understanding their dynamics; and the effect upon the organizations which exist within their boundaries.* 3.5. THE COMMUNITY AND THE ORGANIZATION One of the important purposes of the project is to consider not only the community structure as such, but having developed some operational understandings of key indicators of that structure to look also at the relationship between community structure, on the one hand, and agency operations on the other. As a transactional human service organization, the welfare agency is a critical one, and of considerable interest. Portions of the work in exploring this problem have already been referred to (Appendix 35C)o Basically, we have hypothesized that as the community becomes increasingly'"lower class" the organizations serving the population will become distant and rigid, attempting to separate themselves from the population as much as possible. This hypothesis, however, is only part of the effort S and does not conceptualize the organization in such a way as to permit the empirical inspection of the relationship between organizational variables and aspects of community structureo Such a conceptualization is suggested in Figure 35.5 Basically, we a.re suggesting that an organization is an open system, with three main subsystems - the input subsystem, or resource garnering and processing subsystem, the throughput system, or input modifying and adjusting subsystem, and the output, or product subsystem, which provides exit from the organization for the changed product. Basically, this model simply proposes that an organization gathers input and brings it into the organization, processes or changes that input from its original character into some new character, and sends that product out of the organizationo Apart from general social. osmodic influences, the organization has explicit contact with the community at two points at the very least-the access boundary, at which resources and other elements are flowing into the organization and the egress boundary, at which the changed product is flowing out of the organizationo See John E. Tropman, "Social Stratification, Community Competence and Social Intervention" (Appendix 35.C)o 148

0 *H 0 0.-.r-{ *.r CO 5 0 0 C)O 0 CH: CH c3 i'd h 0 *H H -P a) EH (D - 0 %. 03 AT) * P h 4X; co I- o -4 O --' cP H,Q 0 1 Rm C-H w I ^^ +3 J X. 03*H ( QO *H O E a) E g ego o )0 o A o d H- -pi O M 0 P-20 02 __ _ C -H CH CH 0 4P *H oQ 3 -H P - r-I l P -H p.? - H 4 -) P -P CH 4r).C H 0 h o. 0 IrI 4| O -H E 1 t tCQ Ct.H - (kl _____________________ CO T-I ( o ( H 4r-0 o r'"rI I r-0 - r- r |?0Co)3 0 CQ. fcU) Q) j WR4 E-ACO k H Cd 0 - H C o CdH br - 3 0) 00 +0._)C.c)H+CP toOP 0C-H ) H0 0 C _Q -p r~~ H CO H4 C Q O CQ T- ^-( D r-!"- r-" Z P Nr-1~ ~ ~~~~;~ - oC r- N Kl a) H0 E-O to -Hc 0 )H ~ 3 C) -p a (dr o to 0 CQ IC C) H3 -- 0 0 0 0) C) E T H- MO O~ O O HHHHH ~ ~ 14

Community influences can be critical at either juncture. Community characteristics will heavily, but not entirely, determine the level, nature, quality, etc., of the input. Too much input causes overload, and system collapse. Too little (money, for example) causes an organizational shriveling and spasm. Input of too low a quality means that the organization must do additional processing, something which it may not be prepared to do, and which increases the per unit cost. The community also tends to control the output, through the structure of its market locations. For example, if there is no market for a particular product, this causes a backlog in the organization, resulting in a certain degree of swelling of the system, and possibly system collapse. An obvious example, here, is from the geriatric hospital field. Many geriatric hospitals were designed to be revolving doors, into which the aged patient would flow to receive needed medical attention, and out of which he would flow to a nursing home or some other infirmary type of arrangement for the necessary recuperation and care. The absence of nursing homes, however, turned the revolving door into a one-way street, with such hospitals bulging, and backlog of such proportions that many soon had to refuse new admissions. To specify the community influence empirically, measurable variables are required which tap these various aspects of the organization. The organization paradigm provides such a. conceptualization, and can be used in alloca.ting different organizational measures from the welfare administration questionnaire. Using the community measure discussed, we can attempt to specify relationships between the organization and environment. Despite all the measuring and operationalizing, however, there are some aspects about public welfare organizations which are more or less unique to them, and which need to be understood if the community influences can be accurately determined. There has yet to be written an authoritative monograph on the public welfare agency. Two points at least deserve mention here, first, public assistance agencies exist within a climate of public hostility which is rapidly increasing. They are subjected to checks, audits, monitoring and a. constant level of criticism. Some of these data. as they are perceived by the county welfare director, are reported in the section on the questionnaire. Secondly, they offer the unusual combination of help and money (a combination which is now under process of reorganization). This marriage of functions caused a certain organizational schizophrenia, from which it will probably not recover for some time. In any case, the sets of community expectations surrounding these two functions were not the same, and often tended to interfere seriously with one another.* Special efforts will have * nnn 7 -nn n A.i i i — ~ -r John E. Tropman, "Administrative Structure of Public Assistance and Organizational Change," unpublished paper, The University of Michigan, School of Social Work, 1969; John E. Tropman, "Comparative Administration and Public Policy: Three Programs Under the Social Security Act.," The University of Michigan, School of Social Work, 1969. 150

to be made to tap these special aspects of organizational climate for public assistance agencies. 3.6. A DEMOGRAPHY OF THE SELECTED STATES It might be useful as well as illustrative to consider, for the five states for which we have career and questionnaire data, the social and demographic structure, using, as a beginning point of analysis, those variables which we have identified as critical. It should be remembered, however, that these states are in no way "typical" of the population of states, and should not be thought to represent anything but themselves. Studies using some of these data for all states and the District of Columbia, are currently underway, and, when those analyses are available, we shall more accurately be able to place the states used here in some perspective. Further, the arbitrary nature by which these states were selected presents a, further reason for selecting a. nationa.l sample of counties. Selecting a "sample" of states from an N of 50 has the obvious problems of sampling size. The 350-county sample, described elsewhere, should help us to correct idiosyncratic influences due to the selection of states. However, the use of some states was essential, because in the administration of public welfare, the state is one of the critical units, particularly because the amountof money appropriated by the state legislature determines, within certain limits, the degree of federal participation. (This point will be discussed subsequently.) Basic data are displayed in Table 3.3. It is immediately clear that there are marked differences between the states. Total population, our measure of complexity, runs from 11 million down to 3 million. Three of the five states are smaller than New York City, and one is about the same size (as of 1960 figures) and one is larger. There are some interesting differences in the stratification measures for the states. One state is clearly above the rest, with a median stratification rank, taking into account all the variables, of 5. It has the second highest median income, the highest median education by 2 years, meaning that over half of the population graduated from high school. It has decidedly more people in white collar occupations and the lowest police expenditures per capita. Two states have a median stratification rank of 2, which means that they are relatively low, within the group. Two other states have middle ranks of 5 and the fact that these states are higher rather than lower within the stratification system of the United States is seen by comparing them to the national medians. All states have a higher median income than the National median ($5,660). Three states have a higher median educational level. Three 151

LI\ - r\ Lr\ OJ LU\ "\C (M _ r-. OJ' r-' - U ~J-' L L a1) rNr'\ Oj I"~',. — r-... O r- -. K',\ WN~I,~, r,(% r,~\ ~-..P., 5:11 ~~~~~~~~~~~~~~~~~~~~Lrl\ O () ('k C ~ ~ r-r' r- N- iKr\ I..r~ ~1..~\ r. — r'-A I'(" r-'l (!) N"\CJN~C -l HCMO r<~\ H r<"C\ 0C\\ K K^ O\ _t 1^ i\] N\j -P 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0 0'- 0- 0P -,- a) a 0) 0- O' —.1- - "-.' -t-t H J O O' L.r OC OJ.-.r- CM O OJ E!.., P Lr O ('C1 C.If r'-NI. I".. -. t O-Lr -) (ML'-1- ~ P ) co \N H 0\ t t' C\j _::I- Oj \10 C~i r-I CC) I b.0 0l) KO H 0 O) \0 r-1 0 1d P) ) C M..,1 - I o ^D*N^ q-t r-1004 -P C \CMCM:t L)JLI 0\C) IO (YCMC rC\A0 C OrC\ t C t A C~j, 4 Lr \< Q.O HL - Ifr rIJ LO CM-D0 0\f tCM C-I,, Z a) I ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I (Ul\ C~j.1:5 "J O) 2 ~c 1k - u' 4 — OJO~~h.r- Ckl>0 1 0\O O r OJO K ~ bCkl LP) ~ I C) H o-,.S 0 1 1 -4 C.)~', O Cz-l r-' -- CC I', \OJ- O O HV\- - I C- W I CM' *D< *IH... -p a.) if IIMI Cl)Co a) ) I- - Q) 0 C) I O OO H-O 0H C) 0 o i I 0 OC tP i t-) (kl ^,1 c o \ O t C INC jHCM CC) I 4OJ3 C2 Ucd a)~H.' H' 1-1 r-1 C3 0rQ U~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~r-4c^o Oo,O Oo~ r, ~ ~-P, Or —c c a?) I.- I'~ L HCk - Otr-CO II a)4 CO cOid- Ir r[_L \,D O -- H 0\ - 0P~~~~~~~~~~~~~~~~~I r-1 O _ rm S* H 1!-P cl rcQoON.-zl \D0I - -1 ^ t-^ i O1\-zf a)!-1 t-'I H' I POOJ 0 Lr\K \ I 0O I c i -PPI ~'\) FS3 rI1 0 c6 ^^OC^^O K^OJ K^~~~~li K^ JN^ 0000a o Cl) H?-l 0r\ * ~ ~ ** OJ....... (D C 0 0O-P CMOOCM O0\ 00 0 I ON a?) -— zf4O NCOJOJ I a) cliC) ~ ~ ~ ~ ~',Hzt? HI CM\MX.o(X Io,~ o ~~:l~a) E-4 QO~ ~4- P I C, (X - - C- Hd -- a) Q Cl). k.:',,k.1r, 00J ~r,' OJN O rc O O oO. i c CC) Cd (:~>0 r0 Q) — 51~ -3O 0:rt. 5i Po cd P~~~~~~~~~~~~~~~~~~~ 0 <4 bO~ ~~~~~~~x. F, - I —I ~r C/3~~~~~~~~~~~~~~~~- UU 0 ^ d rArl~ / S(aO p3,~~~~~~~~~~~~~~~0-P -PC)~ C o r qf 4 b0D 0l cY. a) H 0 0C.H.H C)Lt'r-0 a)-Pcli -JC) ) F- d b0 - d-rI;,E *H tt liH, C) OC C) rcI HOCI9 0 r-)cl -rI () clir U c Oa) 0 — a) r4roC H l-ai o +:H C C) F-4C a) a ) cli C)C)O (L>P < HCT3 C)H a) C) -P U2 00 >C a) C)00uhiXP~cli hi) b C)O c($ Cli C) pp0-34 > 4- ) 0 M)Fzcl(~0 Cd C)Ha 0 a).Hr 0 0 P-l!>'H ci H 5 0C)C C)C)-p0oa) C)C a^)C ) p (U')?) *b C) a)C cliliHC)C) oIcl hi Hci'4 c lTlig a h0P4C) C cl o ~a H* )OHa*- a 0 — )t'Hl ( cli *H OO0cl *H *HO0 H *H?Fl S ) )a 4) 00 )a) ) ) o> -) C)P4~~~? ('H 4 4'H ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~0 OC 03C)a ^~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~>.-4P c!i -P' 0: co Q0a a) *H 0CliCO a-P *H 4-^ 0 C)T).P C) CH~~a^|Pp4C)<^ -~P~-Ci x 4'H-P Cd a) C) *H a)OC)P'a *HH a)J H'H a)?-q i-4 0 -P 514 " C. d;:r-A 75 r-A 1 -,^ ttD r-I ( O Cd) 0 5 C)~+ HC)-I O PH lH hi)'aH a 0 iOM OC) OP-4 lHiC) P CO4 C) a)C3 4-P 0 0 0 -Pci Cd 9:5 -) O 0 oCQ 0E 152

states have a higher proportion white collar. All states have a higher proportion of sound housing, and three of the states spend less per capita on police than the national survey average, with one state spending at exactly the national mean ($10.30). The different states differ importantly in the percentage of Negro citizens they have, ranging from 9.2 to 0.7 percent. As can be seen from Table 3.3, the ranks of the'different states on the basis of percent Negro are similar to the ranking on maturity. This suggests that those communities which are more mature are more likely to have higher proportions of Negroes. The meaning of this for the communities and the Blacks remains to be explored. However, several possibilities can be considered now. We defined percent Negro as being a measure of mobility coagulation.* It appears that, to the extent that the Black population is living in the more mature, rather than the growing communities, there is additional support for this generalization. Consider, finally, the Maturity/Stability and change/growth modalities. One state is clearly mature and more or less stable, showing low growth rates, while one other state is clearly growing and developing. The growing and developing state is also one of the higher ranking states in the stratification system, while the state which is slowing down is one of the lower states in the stratification system. As an aside, it is important to note that growth is a combination of the gains minus the losses. Hence, since the mature state also tends to be the low status state, and the growing state the higher status state, and we know that poor people do in fact move around, the result must be that there is an influx of poor people into the mature area offset by an outflow of higher status persons. This possibility will have to be carefully checked. The other states rank in the middle, with some growth and some maturity, and do not become clearly identified as growing or stable. In sum, we find that, even among five states strategically selected for other phases of the research, there are some important differences. This, of course, is considering the state, as a whole, as the unit. For certain purposes, this is a good unit but we are at least as interested in looking at the county, where we expect to find intrastate variation. As we begin to consider the relationship between local social structure and local public welfare administration, then we will expect to find variation there as well. It will be interesting to see, for example, whether the intrastate variation is greater than or less than the interstate variation, using the county as a unit. The conclusion, though based upon these data presented thus far is that states differ markedly in their social makeup. This fact has implications for the welfare structure, for the whole structure of meeting social The 1962 and 1967 City-County Data Books differ here. The former reports "percent non-White" and the latter "percent Negro." We are using percent Negro, but will explore the relationship between the two. 153

problems within the state, and for the competence with which those problems are or can be met. 3.7. WELFARE STRUCTURE OF THE STATES Trying to understand how a social unit meets the social problems within its boundaries is a complicated endeavor. Often there are a bewildering number of formal systems, such as organizations, clubs, laws, etc., which can be invoked to provide help. These formal systems often exist at different levels within the unit. Further variation occurs depending upon the type of problem. Housing, for example, may be handled at a low level within the system, while another problem, likely securing adequate money, may be handled at another. (This, of course, causes little difficulty for the user if he has only one problem.) As if this complexity were insufficient, the public structure is often mirrored by a private one, sometimes as informal as a neighbor watching a child while the mother goes to the store (child minding) to talking over a personal problem with a friend (counseling) to lending or giving money in time of need (financial aid). The state system like all other systems, meets needs through a variety of formal and informal mechanisms. Our particular emphasis in this report is on the public welfare system, and the degree of effort a state makes through its own devices and that of its constituent counties to meet financial need. We need to first understand what in fact the state does in the public welfare area, before we can attempt to account for why it is different from other states, etc. However, it should be abundantly clear that attempting to describe the way a state meets social needs through looking at the welfare system is like trying to describe an elephant by feeling its trunk. The welfare system is one of the most public vehicles through which and by which a state meets social needs, and it has for this reason received a good deal of publicity and scrutiny. The welfare system, however, represents only a portion, and an unknown portion, of total problem-amelioration effort within the state. The level, scope, thrust of any particular program depends in part on what else is going on within the state. The Welfare Data Before we begin presenting different measures of welfare effort made in the public assistance program, we should discuss the type and nature of data available. There are some important problems and qualifications concerning the data which should be noted. 154

The first problem is finding the total number of welfare recipients. Per capita rates and mean grants for all states and counties are available for 1960. In addition, we secured the booklets assembled by the Welfare Administration and later by the Social and Rehabilitation Service containing multilith copies of the reporting forms sent by the states to the Department of Health, Education, and Welfare for 1964, 1966, and 1.968. These raw figures are, in our judgment, the most accurate possible source (not counting, of course, inaccuracies in reporting) and the one we shall use as a basis, -and from which we shall calculate other figures. These reports are not completely comparable, however. In.1.964 and 1966, the states were asked to report cases for OAA, AB, APTD, GA, and AFDC. For AFDC, they also reported the number of recipients and the number of children.* In 1968, the number of recipients was reported for all categories instead of the number of cases, except that states continued to report cases, children and recipients for the AFDC program, and they reported cases and recipients for General Assistance. The type of reporting unit is presented in Figure 3.4~ The different modes of reporting create some problems, both in comparison between years, in comparison amongprograms, and in comparison with other studies. Year Program 1960 1964 1966 1968 Old Age Assistance Recipients Cases Cases Recipients Aid to Permanently... Recipients Cases Cases Recipients & Totally Disabled Aid to Blind --- Cases Cases Recipients Medical Aid to Aged -- Cases Cases -- Aid to Families with (Families) Cases Cases Cases Dependent Children Recipients Recipients Recipients Recipients Children Children Children General Assistance - Cases Cases Cases Recipients Figure 35.4 Reporting formats for 1960, 1964, 1966, and 1.968 recipients of public assistance An anomaly here is that the recipient rate is reported for 1960. It appears, and we are attempting to find this out, that for two of the programs, OAA and AFDC, the rate is actually the case rate, called the recipient rate. Unless the Department had one reporting format in 1960, a different one in 1964 and 1966, and returned to the original one in 1968, this explanation must remaino In any case, comparisons between 1960, 1964, and 1968 can be made. 155

The first problem relates to the census. In the City-County Data Book, 1967, they present "Public Assistance Recipients, 1964" (variables 51). The note says that it is based upon the report for 1964 which we had. For awhile we could not duplicate their figure. However, after considerable work, it became clear how they had made their calculation. They had added the number of cases for OAA, AB, APTD, MAA, and GA to the number of recipients for AFDC and called that number "Total Recipients." In one sense, this procedure is not completely unreasonable. In most of the adult categories, a'ase" generally equals a "recipient," although there is some necessary understatement, of total persons being helped. In the AFDC program, however, a case decidedly does not equal a recipient, because each case contains an adult and children. Hence, it is imperative to consider, for this program, the number of recipients. In the General Assistance Program, however, it is less clear just what constitutes a case. There is some evidence on this point, however, because in 1968 states reported both cases and recipients for the General Assistance Program. The two figures for the five states which concern us here are presented in Table 3.4. It appears from Table 3.4 that there are, on the average from 1-1/2 to 3 people per case in General Assistance. This obviously reflects some individuals, and some families who are receiving this type of aid. The point is that there is some distortion in the census figure, because the number of cases in General Assistance cannot be considered as equal to the number of recipients. TABLE 3.4 NUMBER OF CASES AND NUMBER OF RECIPIENTS FOR GENERAL ASSISTANCE PROGRAMS, FIVE STATES, 1968 State Number of Number of Cases Recipients East Central 21,333 70,572 West Central 6,641 20,734 Eastern 33,238 42,860 Western 9,118 15,037 Central 5,230 14,192 Hence, we do not really have "cases" and we do not really have "recipients" in the census figure. This calculation was as close as one could come, in 1964, to securing the total number of recipients in a state or county. For this reason, and because we felt that the general distribution of the City-County Data Book, 1967 would make their figure the one available to other investigators, we decided to use their method of calculation for 1964 and 1966, recognizing that it does not really represent "total recipients." Since recipient figures are presented for all cases in 1968, we can easily calculate the total recipients 156

with some confidence. In looking at percentage increases, we obviously preferred a four-, rather than a two-year spano Hence, we wished to compare 1964 and 1968. For MAA, this was no problem, since there was no program in 1968. For OAA, APTD, and AB, we would have to compare cases in 1964 with recipients in 1968. Although as noted, the case and the recipient are quite close figures, we felt that the rate might have been misleading, hence we used only AFDC, where we had recipients for 1964 and 1968, and General Assistance, where we had cases for 1964 and 1968. While rates are of concern, money is also. We are interested in how much is paid to recipients in a state, how much the average per recipient is, how much on a standardized basis this represents. In the same state report mentioned previously, there is information on payments for each of the assistance categories, and it is possible to secure information on how much, within each category, recipients receive during one month. However, this dollar amount is only a portion of "welfare expenditures." There is also the amount for administration, including salaries, equipment, etc., items which are also welfare costs. There is also in the City-County Data Book, 1967 a variable (54) which reports the state and county expenditure for all welfare programs, including the federal ones, but also including local institutional help, and other miscellaneous items. Other types of welfare data are available but, for reasons discussed in the section on methodology, adding variables is somewhat tricky, particularly when, in two states, there have been changes in the numbers and identifications of counties between 1960 and 1968. We feel that these data provide a good base for extensive analysis, and, since the 1970 census will be available shortly, we can prepare questions to be answered with new data. 35.8. WELFARE DEMOGRAPHY The raw figures for rates and payments, for five states for each of three years is presented in Table 355. It must be recalled that the figures presented in Table 355 are for one month. The recipient population changes somewhat from month to month, but the total yearly figure can be obtained only by multiplying the monthly total by 12, although this approach is not completely accurate, since the monthly payments reflect the recipients during the monthO The dollar levels and persons are large numbers, yet comparative assessment is difficult without some notion of the persons and dollars in relationship to the population of the state. These data are presented in Table 3.6. Table 506 presents the rate of use of the various public assistance programs, per 1000 population, and the amount of money paid out to the recipients, 157

o ooo oo o cq oo Il r r~ o o~ o. ao, o 0 a 00 O 0 n CD O 000 0r 0 00 a' I 00 000 I" 000 uC Cm C l- >~ n c < In o mq 0 < - 00 0 mC 00 000 00 000 < C O 00C C 00 0, I) 00o 04 n In 0 0 0 000 I a0 00 CD - n oo In 00 0 M f c <t ~ID 1 In r- V) rI) JJ i^ cq o 0) 000 L0010 000Cir OO 0N00 O0Cl00 000 a0 In 0C0 10 00 10 000 r 0 ~o 0nc o ooo C oo oo r 0 a OO E o 0ID - 0 r I0n It -00 0D r ooc 00o ) I or 00 03 0' 0 C m n q 1 C) 4 ml < ao n oO O X 0(3 0?? "I I r c00 CD 0y o 0 0 00 00 0\0 nq 0 C4\0 10 0 m C0 O0 S O O 0 00 o 0 0.. 0 00. 0 0 o 0 0 158 r q > CL o 0 0 In 0 0 0 00 000 0 C', M o000 00 0 00 r0 0 000 InW n Lo 00 C0 00 000 00 0 t00 000 000 00 000 00 000 0 C 000 000 co 0 r 4, n 4 4 00 0 0 A

\ID LU-\ Lr\ i-l 0\ \D \D 01\ OC\J H N\ IU-\.r c-_l\ cO 0 H - O r- l u-\ 0 -P _:- CMjC \Co0 \ 0uL\ CO Lrf\ K^ Lr~r K\O' WK OJ OJ \,D OO 0 t- CJ CO \,DLr\ - r- 0 COJO N COJ 0! r- r- O r-C OC r- r — CNJ r-1 r-1 r-H cO3 G\'-l OJ LU\ Lc\ -\O 0 N'X 0 O'I CC) 0 0 Cn C CO -J' - t-. r'x r' x C OJ O' OJ r-H rCO "C ) 0 Cr — r —c r t-'C \>0J ON >-O C r-C\] S t 4^ OJ J c cOO LD K 0O c.00 0c O Lr\C) U)D 0 H COO0 c O\CC) O jON\ -o tE —Q c \ r 00 QP\ Ni r-\C-0\ Q0\s-0 r0 — OJU Lr-\] \c O' C'. 0 * h* c O oc O* *-O 4 ~ 0 C\ C cOQ - K~0,,-t- -- \C \\ 0 O Lx C \ V^ > O co oO cON Q CQ CMOJ \ 00~-00 CO [-iH ~; cL L\ rC\Zj- - \ Lr\KO L\ rKOC CM -ziCQ 0a C/ QOLr\ tL O\QO DOCO Q N C/) c. \]H C) O — m Lr CMJ C Q C * * e P +p u^0\ i iCt Ir-c\ I I S I ( I i1>- i COE- O]O Lr\ C \D-^ \OLr\ bV E;0q (L) J\D \O 1f \90 Cd- tL\ O ~ 2 CM LH' CXM rni rU2 Cs O 0~ ON OZ- HV 0C 0:=C H > CO - r &r 1 1 [ V-1 \70 Co c2 pq EHCOKO 0). -OHL- HCMOO J 0[\ V N Cr-M\D H o * * * o eo * * *oo ~ o't r- C ioC_ D ro-\ tN ONL co L \ C\ -' OP Q'O DK 0. O_:- 4 u- C a CM Naf:~ U'~..-~- ~-t 0..a~~CM: 0 DH 0, 00r 00 CMr~ - 0-I r-I -4 q ~-I r-I r - r-AO (ll _:K- I (l l r-'l r-i P0 V2 \ W H CO CM o O t —) 0C L\ Lf\ o:J- C - O DJ- E-4 O Cd H wO ucO 04O~ O t EAi t i C> * o* e eo * * o o ~* CO _ L CMCMH C MHH F-7 E-I u - J Od OJ CcO'xi- k.O C --- cOC' C \O-.L -Q 4 PZ, O O0.0 jO. 0\ L Lr\. N OKH O>:. 0,- P-p <L~ -p~r\-t V^CMCM Q\Q9\-O t-QDCM P pC) 4D 0\ O UC\K co - pC(a) C'J D A n Oj C O CO C] CH 0CQ O -H CU H H - I N CX~j CC) CO C O CO CO _:J O \ pq o D C\ \0 E00 o0 a, Ct \C0 \0 Lr C\ CQ \D 0 U0 4t3 0 030 Ep U2 e e * * * a * o e * *-1 e CC) C~ \O0 U K 0 \ C0 Lf- UN c o C Q)4 C) C) C \ \ \ \' \9 \ \ CO ( ) U n WW >n HO 159 V H-r

per 1000 population. This mode of presentation thus takes into account the fact that different states have different populations. A primary point of interest is the difference between programs. AB and APTD are small, both in rate of use and rate of expenditure. GA falls somewhere in the middle, larger than the previous two, but smaller than OAA and AFDC. These two latter programs are the largest. Interestingly enough, they occupy two very different positions within the political scene. The aged have recently benefited from Medicare and the drop in 1968 for this category partially reflects the availability of nursing home care. The AFDC population receives little except continuing resentment. We might also note that there is a considerable amount of variation, both within states and between states. This doubtless reflects both the availability of resources and the fact that there are need swings at various times of the year. Depending upon local conditions more or fewer people will have need for the public assistance program. Indeed, as we shall discuss, our sense is that there is considerably more elasticity in the demand than in the supply of cash, something which becomes critical for the understanding of the politics of welfare. Another important aspect of the welfare program which relates to the degree to today the program is increasing or decreasing. We have accurate data on two programs-AFDC and GA for the period between 1964 and 1968. These data are presented in Table 3.7. One fact becomes immediately clear. In the AFDC program, with one tiny exception, there have been increments of substantial proportions, regardless of how one looks at the data, by case, by total recipients, by the number of children, or by payments. Increments in total cases range from 2.4 to 43.1 percent. The increments in payments are even larger, ranging from 29.9 to 54.4 percent. These are not simply marginal adjustments, with a few people coming on and a few coming off, but rather represent very substantial increments, ones which could not avoid public scrutiny. Sometimes, talking in percentage terms, though it does provide a standardizing factor, obscures some of the dimensions of the problem. For this reason, we have included the absolute numbers as well. These percentages represent thousands of people, and millions of dollars. Consider the total number of AFDC recipients. One state lost 3,000 recipients, but each of the other states gained over 10,000 recipients, with the highest being over 30,000. Percentages aside, these absolute numbers represent amazing growth, and suggest one reason why the program has been under such scrutiny. 160

TABLE 3.7. CHANGE IN TWO PUBLIC WELFARE PROGRAMS, IN FIVE STATES BETWEEN 1964 AND 1968 AFDC GA State Total Cases Recipients Children Payments $ Cases Payments $ Percent Change East Central +16.5 +19.2 +22.4 +51.8 - 8.9 +37.6 West Central +27.9 +27.3 +26.1 +29.9 -29.4 -17-3 Eastern + 2.4 -.998 + 1.8 +27.0 -10.96 +10.8 Western +16.0 +17-5 +15.1 +43.5 +17.6 +38.4 Central +43.1 +44.5 +44.7 -54.4 -29.6 -37.7 Absolute Change East Central +6,510 +30,671 +26,040 +2,890,435 -2,072 +688,791 West Central +3,649 +13,100 + 9,770 + 753,952 -2,775 -133,687 Eastern +1,660 - 3,106 + 3,955 +2,526,904 -4,093 +245,908 Western +2,685 +11,356 + 7,061 +1,049,136 +1,366 +202,816 Central +4,945 +20,047 +15,169 +1,124,037 -2,200 -261,133 The level of persons is reflected in additional demands upon public funds. The smallest increase in payments within this program was three-quarters of a million dollars, with the largest being 2.6 million. This is for one program alone, and does not reflect the additional costs in personnel, equipment, and so forth which are obviously necessary if one is to process through an agency thousands of additional persons. The GA program has a slightly different, and rather interesting picture. Of the five states, two show decreases in both cases and allotments. One state shows an increase in both categories. Two states show decreases in cases, but increments in allotments. It may well be that the greater degree of local control over this program makes it more responsive to local pressures, either for more or less use of the program or whatever. The analysis of data on a county basis will perhaps provide more information on this point. The overall conclusion, however, is inescapable. Between 1964 and 1968 the use of welfare and its cost grew substantially. The only important decline (Table 3.5) was in OAA, where the implementation of Medicare probably reduced the rolls. When one begins to look into the degree of politicization surrounding welfare, one can see that the growing number of program users are approaching some vague "public limit" to the degree of acceptability of use. This did, and will continue to, generate friction. 161

The political ramifications can be seen in another way. Consider the average grant, per recipient, for the various programs. We have recipients'data for all programs in 1968, and that data is presented in Table 3.8. TABLE 3.5. AVERAGE GRANTS, FEBRUARY, 1968, FOR PUBLIC ASSISTANCE PROGRAMS IN FIVE STATES State OAA AB APTD MAA AFDC GA East Central $68.92 $ 86.83 $85.11 - $44.61 $35574 West Central 65.86 81.24 79.53 - 53561 30.77 Eastern 74.58 109.16 72.90 - 38.58 58.65 Western 68.22 86.50 78.41 - 45.48 48.61 Central 65.60 84.96 60.82 - 49.05 30-37 With some differences, the ranking tends to run as follows: AB recipients come first, followed by people on APTD. OAA recipients come in third, with AFDC fourth, and GA last, and lowest. A couple of points should be recalled to understand these figures fully. In the three adult programs, there is usually one recipient per case. However, in the AFDC program, and in GA, this is not so. Since each child is a recipient, one would have to multiply the amount of the per recipient average by the average family size to achieve the figure (approximate) which the average AFDC case (family) receives per week. There are approximately 4 persons per AFDC case (Table 3.10: 1 adult and 3 children), hence, for a rough total count, the per recipient figure should be multiplied by 4. Considering these data on a per recipient basis, there is a rough reflection of the political popularity of the program in the differences in amounts. The two highest amounts go to the programs for the blind and disabled persons who, within the cultural framework of American mores, have a clear "excuse" for being on the rolls. Very little public resentment accrues to this group. Next, on the average, comes OAA. Here again, the older citizen is one for whom there is some support, or at least an absence of overt hostility, in the system. This program is one which does not receive a lot of public attention, certainly nothing like the AFDC program. The two most unpopular programs, under this rule, are AFDC and GA. Certainly this finding corresponds to the public impression one has. Generally, when people talk about and against "relief" what they have in mind is the AFDC program. In fact, one of the main complaints about the program is that recipients receive so much money. It is for this reason that we mentioned the rule for calculating the approximate family allotment. This amount can top $200 per month, or $2,400 per year, for the larger families. This dollar 162

level looks large, until one recalls that it is also for the large family. However, even on this point the poor citizen who is not on "welfare" objects, because his own salary is in no way reflective of family size. He earns what he earns, and that's that. General assistance is unpopular because it is money which comes strictly out of local coffers, and money which could be used for other pressing local expenditures. Hence, for both the AFDC program and the GA program, there are good political reasons why they would be lower than the other programs. However, if we leave an examination of the intraprogram differences in mean monthly and look at the grants as a package, there is the clear sense that even the best of them is rather inadequate. Many states and counties provide, under a particular program, less than $1000 a year. If this is the only income the person has, he can hardly live. These low levels reflect another level of political reality surrounding public welfare in America-it is perhaps the most unpopular program at any governmental level.* Two final pieces of data are of interest. One reports on the total proportion of state populations who are on assistance, under any program, and the second relates to the family size under the AFDC program. Data about the first item is presented in Table 3.9. The data in Table 3.9 suggest that from 2 to 4 percent of the population is using welfare assistance. This proportion does not seem very high on a subjective perception basis, but does represent thousands of people. What becomes important, however, is to ask how well the state is doing in relationship to that group of people who might be "needy" or potential citizens in need of welfare assistance. These figures are difficult to estimate accurately, and, especially as we move to the county would be virtually impossible to obtain. A crude index does exist, however, and is available on a state and a county basis. This indicator is reported in the City-County Data Book, both 1962 and 1967 versions as percent of families who had incomes under $3,000 in 1959. In one sense, this indicator would probably understate the need, since the dollar level is 11 years old and now out of date. Yet even so, between 15.3 and 21.4 percent of the families in the states which we are considering fell into this poverty category. It is not possible to make a direct comparison between the percentage of persons receiving assistance and the proportion of poverty families thus defined, because the welfare utilization figure contains many single individuals, while the census defines a family as two or more persons. Since the figures for total families are available, however, and since it seems reasonable to consider, for the AFDC program especially, a "case" as equal to a * Some aspects of this program have been reviewed in John E. Tropman, "Administrative Structure, Public Opinion and Organizational Change," and "The Comparative Administration of Public Policy." Unpublished papers, The University of Michigan, School of Social Work. 163

TABLE 3.9. RECIPIENTS OF WELFARE AND POPULATION CHARACTERISTICS BY STATE AND YEAR (In Percentages) Percent Percent of Percent State Total State C s o < $3,000 State, Cases of Population T Per Year, Total on Welfare Fami 1959 Families East Central 64 3.15 - 15.7 66 3.11 68 4.08 2.5 West Central 64 3.05 --- 21.4 66 4.94 - 68 3-39 2.1 Eastern 64 3.89 -- 16.8 66 3.85 - 68 3.80 2.4 Western 64 4.71 15-3 66 4.57 68 4.46 2.7 Central 64 2.24 -- 17.4 66 2.29 --- 68 2.58 1.6 164

"family," we can compute the proportion of total families the AFDC caseload represents. Since AFDC families are not all the families on welfare, the proportion understates the ratio of helped families to total families. These data are presented in Table 3.9. The proportions range from 1.6 to 2.7 percent. Allowing for underestimation of the proportion of families being helped, we can say that there is, at minimum, a 10 percent poverty gap and, most likely this figure would be more accurate at 12-15 percent. This poverty discrepancy figure does not mean, necessarily, that every family which had under $3,000 income in 1959 was in need of welfare assistance, but it does suggest several things. First, many of those families need help, a point recognized in a belated and limited way under the Family Assistance plan proposed by President Nixon in 1969. Second, if we use the 10 percent figure as a minimal statement of the gap, it shows how far apart the organization is from the sense that it is meeting the needs, a fact which must be discouraging at the very least for those who administer the programs. Also, it indicates an elasticity in the potential applicant pool which is extremely difficult to manage. Third, the existence of a gap of this size simply reflects, as did the low average grants just discussed, the degree to which the United States has ignored the problem of the poor. The final piece of data on the states is presented in Table 3.10. These data on the number of adults per case and the.number of children per case suggest an interesting trend, and one which should be carefully investigated on a county basis, viz., that the number of adults is declining and the number of children is rising. This could mean larger families are becoming welfare clients, or, perhaps, more families with fewer adults, or both. In any case, the trend does not seem to be a localized one, because it is evidenced in every state. Overall, it appears that, although rates are rising and expenditures and costs are rising, grants are quite low. We have discussed only five states, but there is no reason to assume that other states will present a very different picture. For these states, at least, we have estimated that there is, minimally, a 10 percent difference between need, as represented by the proportion of families with incomes of less than $3,000 in 1959, and need meeting approaches, as represented by the proportion of cases receiving AFDC. Let us consider briefly the relationship between these variables and social structure, on a county basis. 165

TABLE 3.10. THE NUMBER OF ADULTS AND CHILDREN, PER CASE, AFDC PROGRAM, FIVE STATES, FOR JUNE 1964, JUNE 1966, AND FEBRUARY 1968 AFDC State No. of No. of Adults/Case Children/Case East Central 64 1.08 2.95 66 1.02 3.o8 68 1.03 3.09 West Central 64.80 2.86 66 1.02 2.86 68.82 2.82 Eastern 64 1.28 3.13 66 1.13 3.17 68 1.15 3.12 Western 64 1.08 2.79 66 1.06 2.84 68 1.15 2.76 Central 64.97 2.96 66.96 2.99 68.97 2.99 3.9. RATES AND UTILIZATION BY COUNTY At the state level there may be many factors which inhibit or otherwise mute relationships which may be clear on a smaller level of analysis. Since the county is the smallest level at which welfare decisions are made, it is important to look, even briefly, at the relationships among programs at that level. Currently, we have available data from 1960, on a standardized basis, for all counties in the United States. These rates have already been calculated by the Department of Health, Education, and Welfare, and, unlike the 166

raw figures we mentioned before, can be immediately used. Again, data for both rates of use and expenditure are available, but only for three federally aided programs-OAA, APTD, and AFDC. AB and GA figures are not available in this publication. Data on the intercorrelations of rates and grants are presented in Table 3.11. Insofar as the rates of use are concerned, there appears to be a strong correlation between the rate of use for one program and the rate of use for another. Because these data represent all counties in the United States, there is no question of sampling variation, etc. When the use of one program goes up, so does the use of another. These intercorrelations are not trivial, either. Figures of.56,.69, and.71 are very strong relationships for social data, and clearly suggest some relationships worth studying carefully. TABLE 3.11 THE INTERCORRELATION MATRIX OF PUBLIC ASSISTANCE RATES AND GRANTS, THREE FEDERALLY AIDED PROGRAMS, ALL COUNTIES, 1960 Rl R2 R3 Gl G2 G3 G4 Standardized Rates of Use per County R1 Old Age Assistance.56.69 -.23 -.55 -.42 -.32 R2 Aid to Families with Dependent Children.71 -.26 -.22 -.24 -.25 R3 Aid to Permanently and Totally Disabled -.32 -.29 -.34 -.30 Mean Grant Level per County Gl Old Age Assistance.67.75.72 G2 Aid to Families with Dependent Children, per Person.79.61 G3 Aid to Families with Dependent Children, per Family.68 G4 Aid to Permanently and Totally Disabled 167

Interestingly enough, the same type of correlation picture, at almost the same level of interrelationship, is observed for average grant levels. As one average grant in one program goes up, so do others, in the other programs. It seems clear that if a county provides higher support in one program, it does so in other programs as well. However, this does not mean either that the grant levels are adequate, or, that there are not rank differences, such as we discussed previously, between any two programs. Important as these facts are, the most stunning finding is revealed in the inter-cluster matrix, where we observe the relationships between the rates of use and the mean grants. Without exception, every correlation between the grants and the rates of use is negative, from -.23 to -.42. This can only suggest that, overall, as the rates of use go up, the allocations on an average grant basis go down. Precisely as need arises, the ability to meet need decreases. There can be at least two sets of reasons for this negative relationship. One could relate to the social conditions in certain counties. As the county becomes poorer as a local social system, demands for aid will, naturally, rise. Yet at this same time there will be decreased resources to deal with these increasing requests, by virtue of the very fact of poverty itself. This appears to be at least part of the reason. Relevant data are presented in Table 3.12. Table 3.12 presents the intercorrelation matrix of some selected measures of local social structure and the rates of use and mean grants for public assistance programs in all counties of the United States. These are the measures of social stratification which we previously mentioned, plus one measure of mobility coagulation (percent Non-White) and a measure of ethnicity, (percent Foreign Born). It becomes immediately clear that as the county is characterized by a more favorable position on the stratification scale, the rate of assistance usage goes down. On the other hand, the average payment goes up. This suggests that position in the stratification system is a key factor in understanding the negative relationship between rates and grants. Two measures of low position in the stratification system-percent of families with incomes under $3,000 in 1959 and percent of people 25 or older with five years of education or less-suggest the identical conclusion. These variables are negatively related to the mean grant, and positively related to the rate of use of assistance programs. The degree of mobility coagulation has a similar type of relationship, positive to rates and negative to the grants. 168

TABLE 5.12 THE INTERCORRELATION MATRIX OF SELECTED MEASURES OF SOCIAL STRATIFICATION AND RATES OF USE AND MEAN GRANT LEVEL, THREE FEDERALLY AIDED PROGRAMS, ALL COUNTIES, 1960 Rates of Use and Mean Grant Levels R1 R2 R3 G1 G2 G3 G4 Selected Stratification Measures Median Education -.49 -.42 -.52.51.42.49.48 Median Income -.59 -.47 -.11.48.26.53.48 Percent White Collar -3.1 -.19 -.36.33.43 -32 o32 Percent Units Sound -.58 -.51 -.60 50.42.50.48 Percent Under $3000.62.51.63 -.52 -.46 -.55 -.49 Percent Less than Five Years Education.60.44.54 -.56 -.50 -.56 -.49 Percent $10,000+ -045 -.36 -.48.41.34.42 o42 Percent High School+ -o53 -.49 -.54.57.46.54.51 Selected Mobility and Ethnicity Measures Percent Non-White (mobility coagulation).44.20.43 -.47 -.45 -.49 -.40 Percent Foreign Stock -.42 -533 -.46.57.53 -59.51 While it should be emphasized again that "correlation is not causation" there seems to be at least a clear indication that stratification provides at least part of the explanation. Needier communities can afford lesse It is almost tautological. Yet the state plays an important role, and counties do not have to depend upon their own resources in these matters, and only a small amount, if any, county resources would be involved in the programs under analysis here. Hence, one must ask if, in addition to the social factors themselves, these contextual variables somehow interact with organizational variables, at the state and county level, to produce the observed negative relationship. We think that there is such an interaction, and that the problem occurs between the relative elasticity of the demand for aid, and the relative inelasticity of the supply of money. For this reason, states, and thence 169

counties, tend to solve the problem of increasing requests for aid by cutting the grants. This hypothesis, which is confirmed by welfare practitioners, reflects political and organizational factors, as well as social structural ones. Basically, the welfare grant is made up of money from three sources-the federal government, the state, and, for some programs and in some states, the county. While the funds are mixed in the total grant which a recipient receives, there are substantially different calculi which obtain for them. The county and state funds are appropriated, as part of the respective budgets of the jurisdiction, and have dollar limits. In either case, if the administration wants more money, it must go to the legislative body and request more. The federal money has a limit, but it is a proportional one, not a dollar one. Hence, on a proportional basis, the federal government will contribute to all expenditures for the designated programs. In effect, this policy means that the state appropriation becomes the most salient parameter in understanding what the recipients in the state will receive. Since the federal contribution is a known proportion, what varies is the amount the state will allocate. Now at this point, we must look to the welfare budgetary apparatus to understand how much is requested. Given some sense of what is "necessary" to live in the state, and some estimate of the flow of applicants, a request is made, which, basically, will be what the department received last year, plus or minus a small amount.* After some wrangling, the budget is accepted, and this amount is the amount that the department has to work with. The reason we believe that this amount is inelastic stems, firstly, from a trouble quotient. That is, to get more, either at the state or county level, the relevant administration must invoke the appropriations process. At the state level, this may mean, and has meant, special sessions of the legislature to consider requests for special appropriations. The State of Illinois, for example, often had to meet in special session over just this item-additional money for public welfare. This becomes troublesome, and an administrator is somewhat less than likely to do it unless he has absolutely no other recourse. At the minimum it brings a new round of denunciations of welfare recipients, calls for investigations, state senators living for one day on a welfare budget and asserting that if they can do it anyone can do it, etc. And, of course, apart from this level of "flack," once in session, there is absolutely no guarantee that the legislature will in fact grant the request. In one state, during the early part of this decade, the request was for a time not granted, state funds were nonexistent, checks did not go out, and breadlines were formed in a major American city for over a week. In short, the program, the administration and the clients must pay for an extraordinary convocation of appropriations machinery. * See Aaron Wildavsky, The Politics of the Budgetary Process (New York: Little Brown, 1964). 170

But why, one asks, could not the appropriations have been made adequate in the first place, thus saving the necessity for undergoing such a painful process? The reason is political, and has at least two components. Firstly, the American people have shown disdain or apathy toward the poor and an unwillingness to support them. The representatives of the people in the appropriations process honor this attitude by keeping appropriations lowO One can be sure that these requests are considered with great care. Knowing the attitude of the legislators, administrators submit a budget they think they can get, and this usually reflects only incremental change. Putting welfare aside for a moment, the work completed so far on the budgetary process indicates that only in the most unusual instances does one secure large increments. Hence, in the welfare case, simply beginning at a low level is enough to keep the program at a low fiscal level, since appropriations officers are almost universally unwilling, short of war, to grant what appear to be large increments. This rule seems to be true for almost all programs, and would certainly be even more the case for a, program with the political liabilities carried by public welfare. From the budget side, then, these reasons tend to keep appropriations low. Even so, it might not be a problem if from the applicant side, there were not such elasticity. It will be recalled that we developed, by comparing the proportion of families with incomes under $3,000 with the proportion of families receiving welfare assistance, a gap of about 10 percent. This proportion is a conservative estimate of "need" which, at any moment, could present itself at the offices of the county agencyo It is quite difficult to estimate the social conditions which might move some people out of the nonrecipient category into the applicant categoryo But whatever they are, the more the recipients, the faster the state department of welfare is going to run through its appropriation, unless it cuts grantso Basically, as applicant pressure builds up, and with only dim prospects of additional allocations, the state department can turn people away at the applicant phase, and keep the recipients it has at a higher levelo Alternatively, it can admit many people to recipient status, and shave the allotment. This process can occur even as the state and county are actually increasing the amounts they pay for public welfare program. It is easy to see that, as the number of recipients begins to rise, and the allocation remains fixed, there is a point at which grants will begin to drop. If the rise continues beyond that point, the state has to put in more money to prevent even further reductions. This, to us, explains the bitter truculence which often erupts when state fiscal officers and clients get together, the former pointing to ever increasing efforts the state is making, and the latter pointing to inadequate grants. This problem is serious enough in an average state, but taking a state or locality which is itself poor, in which revenues are dropping and demand rising, the problem can become disastrous, particularly when the program is not a politically popular oneo This, then, is a hypothesize scenario of the welfare crisis. The more the demand, the more inadequate the fundso 171

PART IV RESEARCH DATA SYSTEMS: AN EXPLORATION 173

4.1. INTRODUCTION In recent years, and at an increasing rate, there.has been discussion about developing information systems for formal organizations, the army, for control of flights into outer space, and the like. Vast amounts of time, energy, and talent have gone into the development of such systems. The results, however, we think it is fair to say, have been mixed. On the one hand, we can find examples of systems which generally wQrk very well, such as the computerized airlines reservation system. On the other hand, computerized magazine subscription and billing procedures are the source of a rising cacophony of complaint. Very recently human service organizations, including public welfare agencies have become interested in the utilization of automated information processing systems.* The research enterprise, however, has lagged behind the industrial and aerospace complexes in developing data systems for its own research. Perhaps this is because we tend to regard data processing as simply preliminary to the important work of theorizing, or perhaps it is because the senior people who make many of the research design decisions are still ruled by counter-sorter technology, and thus think of one card per case, with 70 index measures, plus ten columns for identification. Perhaps other factors are important. In any case, social scientists have lagged in developing the technology and electronic data processing systems (EDPS) to deal effectively with the data being collected and for the type and style of analyses we wish to undertake. It has become painfully clear to the investigators on this project that we have ignored these items to our peril. This project, no less than any other, experienced the usual run of research problems —securing data, checking it, coding it, keeping track of it, and soforth. Unlike many other projects, however, each part was dependent upon the manipulation of very large sets of data by computer. Well prior to data analysis phases, problems of data management became apparent. For example, the child welfare data for one state, for two years, arrived on electronic tape as the equivalent of over 100,000 IBM cards. Clearly, alternatives such as hand sorts and manual tabulations were completely impossible. The scope, extensivity, and general intractibility of problems in this sort of data processing proved to be an underanticipated source of difficulty. Our discussions with other researchers indicate that these problems are far from unique, but are somehow generally unheralded. See Robert Elder, "Applying PPBS to Public Welfare," Public Welfare, Vol. XXVII (February, 1969); John J. Harris, "Systems Designs for Welfare Programs," Public Welfare 24 (April, 1966), pp. 112-117. 175

We are firmly convinced that social researchers need to attend to the development of their instruments. Many psychological experiments could not be carried out without specially designed equipment. Psychologists have designed instrument shops where such equipment is constructed, and have employed special, particular people to build it. Graduate students are valued if they have had experience operating certain instruments. These points are anticipating a parallel. It is now clear that social researchers need to similarly attend to the development of EDPS. This must become a phase in its own right, and be fully anticipated, in time and cost terms, by researchers. Work and thinking in this area must become codified, to prevent the endless repetition of mistakes and errors. Problems in large-scale data management are particularly serious, and in this regard, handling of social science data has some unique characteristics not typical in the management and processing of other types of data. It therefore seems both appropriate and necessary that we report some of the computer-related problems which have developed in this project. They are not uncommon problems, although their particular manifestation may be somewhat unique. Hopefully, this portion of the report will serve to sensitize other investigators to this arena of issues, and help them avoid some pitfalls. 4.2. INFORMATION PROBLEMATICS The Welfare Research Project on Client Careers drew its data from three main sources. First, there were the records kept on juvenile wards of the state in each state. Secondly, a questionnaire was sent to all county welfare administrators in five states, as well as to welfare administrators in a national sample of 360 counties. Thirdly, social and demographic data was secured on all counties of the United States. These data comprised two data sets, the City-County Data Book, 1967, which contained 140 variables, and a tape from the Office of Economic Opportunity (OEO) which contained 1045 variables for each county in the country. Basically, we were faced with the following problems, each of which has rather extensive EDPS components: (1) securing the data; (2) understanding the data; (3) managing and processing the data; and (4) analyzing the data. These facets are not, of course, completely separate, but they do have at least a modicum of discreteness. Let us consider the first three of these. 4.3. SECURING THE DATA Securing the career and county social indicator data would appear to be an easy task, but in fact it became extremely time-consuming. One reason was 176

that state welfare departments experienced many of the same EDP problems that we and others have encountered. In some states data were only available in files or on index cards, and these were essentially useless for us because of the prohibitive costs of retrieval of the information which we wished to analyze. Parenthetically, it could be noted that these files were also essentially useless for the agency because one state would have thousands of cards and/or files and had no procedure for quick and systematic information abstraction. In other states local autonomy was preeminent, and counties did not report data about individual clients to the state. Only summary reports were submitted by the county to the state. In those states where it was believed that the data were available, the processing of our request revealed many problematics which they or we had not known about in advance. It became quickly apparent that the majority of state welfare agencies today face serious problems in information processing, and many were grappling with some of the same problems which confronted us in data collection, processing, management, and utilization. A small number of states were observed to have fairly well developed information systems, and these departments could quickly and easily provide us with the information requested. These departments also gave evidence of extensive utilization of their data for several purposes. It became clear in this phase of the research that any operational data systems, regardless of the generality of its original goals, tended to become more and more unifunctional over time. State welfare department data systems tend to accumulate that information which is needed to prepare reports required by the various funding and accounting agencies. They encounter much difficulty in servicing nonroutine requests. As we reported in Part I, and suggested above, marked differences were apparent among the five states in the information systems developed to secure and process data, about child welfare wards. These differences appeared to reflect agency goal priorities, although some general findings emerged. In comparing information about child welfare wards who were dependent and neglected children with those wards who were delinquents, it was noted that more comprehensive and well-prepared data were usually available about the correctional populations. The latter departments appeared to be far more interested in analyzing characteristics of offender populations and outcomes of various types of intervention efforts than were the child welfare departments. At least three reasons might account for some of these differences: (1) Technologies for serving delinquent youth are better developed than are those for serving dependent and neglected children. Information required for the utilization of the technologies can be made more explicit. (2) Correctional departments seemed to employ staff who were trained in research and data processing to a greater extent than the child welfare departments serving dependent children. 177

(3) Correctional populations deal with problems of social control in the society, and therefore, more extensive information must be routinely obtained and processed if these populations are to be controlled effectively. Also, social control generally is better funded than social helping. It is manifestly clear that the next decade will see the development of routine auditing of social agency performance. That such auditing of professional services is feasible has been amply demonstrated by such organizations as the Commission on Professional Hospital Activities which audit medical care in hospital on a routine basis.* Securing county social indicator data had its own problems. The CityCounty Data Book, 1967 is now available on electronic tape, but that tape contained only two variables which related to public welfare: total recipients, 1964, and total welfare expenditures. We needed to obtain more extensive measures of county welfare structures and services. After considerable search, a data set prepared by the Office of Economic Opportunity was located and became available. This tape had over 1000 measures of various types of social demographic data for each county. Although it did not have extensive information about welfare agency operations, it did provide us with useful social indicator data relevant to welfare services. Additional welfare data were obtained and coded separately. It is thus apparent that there are many problems in securing data sets for a study of this type. Several key points are important to keep in mind for future research of this type: (1) The initial data sets which we required were large, and therefore had to be recorded on electronic tape. (2) The use of tape posed a problem because some welfare agencies were not accustomed to working with tape. (3) Where tapes were available, they were often incomplete and the method of preparation was such that transformation was difficult for our types of analysis. (4) We were required to deal with data which we did not need in order to have some that was necessary; at the same time we had to be versatile in learning procedures for remedying data insufficiencies. * See Virgil Slee, "The Medical Audit," Ann Arbor: Commission on Professional Hospital Activities, 1967. 178

These problems may appear relatively mundane, but when they involve millions of bits of information, stored on high-speed instruments, the problem achieves new dimensions. It must also be recognized that the same degree of resistance and difficulty that one would anticipate in obtaining a high rate of survey questionnaire returns could also be expected in obtaining data sets. 4.4. UNDERSTANDING THE DATA Karl W. Deutch* has commented that: "We also need an agreement, if possible, on combatting the formation of local dialects among computer centers..." We could not agree more. In particular, there is a coat of many colors when it comes to writing information on a tape. Apparently simple, this procedure can be carried out at various speeds, in various languages, and in varying formats, by various machines, on various kinds of tape (magnetic or punched). This is to say nothing of the individual styles of particular information specialists. We faced a number of problems in this area which should be mentioned, because, in many cases, it took considerable time, after having secured the tape, and after having been informed of how it was written, to understand it. Proper Formating. -Considerable time was spent in attempting to secure the state and county information in ways which could be read by our computer. However, despite much discussion with the particular computing centers involved, small local habits about which we were not informed, or which we misunderstood, often prevented us from understanding what was on the tape for some time. Inadequate Data Description.-A related problem had to do with the descriptions which accompained the data. Very often they were the ones written for the people within the computing system which produced the original data, and for this reason often left unexplained certain conventions and formats which would be well known there but were unknown to us. Data Error and Omission. —It is important, of course, that the data on the tape be what it is supposed to be. Often, there were problems of data Karl W. Deutch, "On Methodological Problems in Quantitative Research," in Mattei Dogan and Stein Rokkan, eds., Quantitative Ecological Analysis in the Social Sciences (Cambridge: The M.I.T. Press, 1969), p. 35. 179

omission and error, even when the most careful attention had been paid to developing complete transmission schedules. The problem of omission was often reflected in blanks or spaces, where we expected to find data. The problem of omission occurred in high rates of nonresponse and of "notascertained" in certain categories. These problems were partially exacerbated by the fact that the states themselves have certain inadequacies in their own data processing systems, and they have, themselves, experienced some of the same problems. 4.5. DATA PROCESSING AND MANAGEMENT Having secured and understood the data, the next hurdle was preparation and management of data for easy retrieval and analysis. In much research, this phase is merged with data collection. In this case, however, at least two phases of the work-that explicitly on client careers and that on demographic characteristics-were "secondary analysis." Because of that an additional step was required, namely preparation of the data for our particular study, storage and retrieval, and developing programs for accomplishing these ends. Data Storage/Retrieval.-As we have mentioned, the data arrived in very large amounts. At current computer costs (averaging $300.00 per hour) it was absolutely necessary to develop a considerable number of storage devices, holding selected, and limited, amounts of data. Often this involved selecting from the general tape subsets of that tape. At other times it has involved, and will involve, merging data from two sets of tapes, or three. Data Development.-In addition to simply moving the data from place to place within the hardware complex, and developing adequate software to do this, a considerable amount of secondary and tertiary processing also had to be completed. This increased both the amount of work and the amount of data. Deutch comments: "Most of our information about the world is contained, not in raw data, but in the relations among them. Ratios and rates of change are of as much, and often more, interest than the raw data are.... The growth of total data in our data libraries therefore, will be proportional not merely to the influx of raw data but more likely to some power of those data. It might be as high as the square of the data coming in. I do not know how high this will go, but we must be prepared to budget for it, and we must already count on this very large growth of secondary data."* Ibid., p. 36. 180

We can only endorse, strongly, Deutch's anticipation as a firm reality in our own case. We can't say it better! Programming.-For both data storage/retrieval and development, programs to operate the hardware are required. These are not only time-consuming to write, but require a competency which is quite hard to secure, today. This University, as well as many other computing centers, has available a number of canned programs, but it became increasingly clear to us that these programs would have limited utility for us because they had been designed for different types of problems, for different data formats, and often had variable and case limits which were unacceptable to us. For these reasons, programming competency became even more crucial, because it was not just a matter of invoking a prepared program, but of designing our own. This required more time and greater skill than had been initially anticipated. Problems are now being solved quite efficiently and successfully, but this learning has been invaluable to us when we consider the utilization of some of the knowledge we have gained for the development of information processing and decisional systems in welfare agencies. In his thoughtful review of the state of the act of information processing, Sherman Blumenthal recently commented that very few industrial organizations have as yet developed the capacity to use EDP systems for other than financial or routine administrative tasks.* He also points out that information systems are needed which reach into and tap all parts of the agency, and in turn, provide data for use by all of those parts in their decision-making. We have repeatedly noted that welfare agencies are changing; but it must be emphasized further that the rate of change is accelerating within the agency and in the environment with which the agency must contend. See Sherman Blumenthal, Management Information Systems. Englewood Cliffs, N. J.: Prentice Hall, 1969, pp. 3-24. 181

APPENDIX 1 183

A COMPUTER APPROACH TO EVALUATING PROGRAM ALTERNATIVES FOR CHILD WELFARE CLIENTS Despite the emphasis on nonparametric statistics and their computerization since Shephard's breakthrough in the area of nonmetric sealing, parametric statistics are not only desirable but also necessary in many situationso Several recently developed computer routines are capable of coping with nonmetric scales (by writing dummy variables for each category of each variable) as long as these scales underlie independent variables. The scale of the dependent variable(s), however, is usually assumed to be of at least interval strength. Thus, the question arises: what can be done if one's raw data are of ordinal nature at best? There are a number of techniques that have been developed for spelling out the metric implications in basically nonmetric measurements, The Thurstone and Likert methods of interval scale construction or the semantic differential procedure attempt to avoid the pitfalls of a direct application of the categoryscale model to subject scales.* Even though they seem successfully to overcome the shortcomings of direct interval-scale construction, they still do not appear readily amenable to interpreting the dimensions used by the raters in assessing the relative valence of the original scale items, And if the researcher should succeed in identifying the raters' frame of reference, he still is at loss in determining the goodness of fit of this frame of reference with respect to the actual ratings, These residual ills are partly remedied by Coombs' unfolding scale method whereby I-scales are transformed into J-scaleso** In a study, part of which is being reported on here, we were interested in developing procedures for further research where we would focus on scaling as a method (ie., the development of some, preferably unidimensional, scales) and not as a criterion (ioeo, the exploration of certain observed behaviors). Yet the Coombsian unfolding technique is more effective as a scaling criterion than as a scaling method because of its vulnerability (= degree of reproducibility) even though this could be reduced by introducing probabilistic versions as Lazarsfeld did with the Guttman scalogram analysis. Moreover, because the Coombs' method relies considerably on intuition-at least at its present stage of developmentit does not lend itself to computer analysis, And this, of course, is a liability which makes the processing of even a moderate-sized sample of raters or ratings quite unwieldy, *See Hubert M. Blalock, Jr. and Ann B. Blalock, Methodology in Social Research (New York: McGraw-Hill Book Company, 1968), pp, 60-108. **Clyde Coombs, The Theory of Data (New York: John Wiley and Sons, 1964), ppo 80-159. 185

Both problems mentioned above-one's inability to determine the underlying dimension(s) and the lack of computerized processing-recently have been overcome by a joint effort of Louis Guttman and James Lingoes which resulted in a series of computer routines for scalogram analysis.* In the Guttman sense, a set of respondents and a set of items are said to be scalable when all respondents and a set of items are said to be scalable when all respondents in the set make reference to the same criterion property, even though in varying degrees, in deciding how to rate each item in the set. Depending on a rater's choice of, and position on, the scale property he will differ in his evaluation of the universe of attributes. In the simplified case of acceptance versus rejection, he will accept the item if he perceives it as embodying less of the scale property than that which defines his decision criterion. Otherwise, he will reject it. If the range of possible responses to the submitted items is extended he will simply allocate a specific score to each item without being forced to either accept or reject it. Thus, his precise position on the underlying dimension(s) will not be known immediately, yet the technique can be employed more dynamically if the rater is not limited by a strict dichotomy, somewhat comparable to the dynamics of an open versus closed type of question. This latter procedure appears preferable if one is not primarily interested in the relative position of an individual rater with respect to the underlying criterion but rather in the overall structure of the whole rating process. The Guttman-Lingoes series of computer routines offers different programs for studying the structure of data depending on the nature of one's data with respect to scaling properties, comparability, and reciprocity.** For the problem at hand (see below) the Smallest Space Analysis (SSA-1A) seems best suited. This program is designed to reduce a given number of input variables (items) into as few as possible or desired output dimensions and to compute the space coordinates (comparable to "loadings" in the factor analysis) of each item with respect to the resulting dimensions. It is beyond the ken and purpose of this paper to discuss the detailed mathematics involved in the SSA as it has *See John C. Lingoes, "An IBM-7090 Program for Guttman-Lingoes Smallest Space Analysis," Behavioral Science, 1965, 10, 183-84; L. Guttman, "A General Non-Metric Technique for Finding the Smallest Coordinate Space for a Configuration of Points," Psychometrika, 1968, 33, 469-506; J. C. Lingoes, EE.C.I. Rosham, L. Guttman, "An Empirical Study of Two Multidimensional Scaling Algorithms,"Multivar. Behav. Res., 1969, 4 (in press). **James C. Lingoes, "The Rationale of the Guttman-Lingoes Nonmetric Series: A Letter to Dr. Philip Runkel," Multivar. Behav. Res. (October 1968), pp. 495507. 186

been discussed extensively elsewhere.* Yet to the reader who is entirely unfamiliar with SSA, the following analogy may prove helpful. The n items as evaluated by the respondent can be thought of as n points, each connected to the remaining n - 1 points by wires of different length. The lengths of the wire are determined by the degree of similarity among the responses to any two items. The task of SSA consists in arranging these n points in such a way that all the wires will bettaut. If all points can be spread on a line so that little or no slack remains in the net of wires, a one-dimensional solution results. If this can be accomplished by spreading the points on a table, a twodimensional solution is the result. If a cube is required to do the job, a three-dimensional solution is found. No matter what the dimensionality of the chosen solution, the amount of slack (= variance not explained) is measured by the coefficient of alienation. If the goodness of fit of an m-vectorial solution is acceptable (a commonly accepted though arbitrary criterion requires the coefficient of alienation to be <.20), the vertical projections of the individual points (original items) into the resulting vectors can be used as the relative values of an interval scale for the original (nominal or ordinal scale) items of comparison. Visual inspection of the graphic portrayal of the data structure enables one to label the vectors representing the criterion property or frame of reference of the respondents in their attempt to rate the items. To illustrate this rather abstract discussion of constructing an interval scale, we now present a specific problem that we attempted to tackle according to the above outlined procedure. In conjunction with a large scale study of child welfare clients, we needed to develop some basis for comparing and evaluating programs which serve these children. For further analytical purposes, it was deemed necessary to generate such bases on an interval scale. Thus, we prepared several brief vignettes of hypothetical situations. These vignettes were then given to a sample of 58 advanced graduate students and 33 faculty members in the School of Social Work at The University of Michigan with the request to rank the possible programs (these represent the available alternatives in the State of Michigan) according to the perceived quality of service provided to clients in these programs. "Perceived quality of merit and service" refers to social desirability as defined by the raters all of whom had firsthand knowledge of these programs and reflected professional standards in treating child welfare clients. Although extenuating circumstances may well affect the decision in a particular case, the raters were asked to simply rule out such contaminating factors and to base their decision on the general principles and practices of their profession. They were likewise instructed to use the full range of the scale for the total number of alternatives associated with each vignette as reproduced below. *Louis Guttman, "A General Nonmetric Technique for Finding the Smallest Coordinate Space for a Configuration of Points," Psychometrika, 33 (1968), pp. 469-506; James C. Lingoes, "New Computer Developments in Pattern Analysis and Nonmetric Techniques," Uses of Computers in Psychological Research (Paris: Gauthier-Villars, 1966), pp. 1-22; see also: Wolfgang L. Grichting, When Authority Fails (Washington, D.C.: Cara, in print). 187

VIGNETTE 1 A child of seven is referred to the county welfare department by a neighbor because of the mother's illness which will require extended hospitalization. The parents were divorced, and there is no contact with the father. Placement possibilities include the following. Please rank each as to the quality of service expected for this case. Least Most Desirable Desirable 1 2 3 4 5 6 7 8 9 10 1. Temporary foster home / / / / / / / / 2. Children's Institute / / / / / / / / / / 3. Private residential institution for children / / / / / / / / 4. Relatives / / / / / / / 5. Local "group" foster home / / / / / / / / / / 6. Local public detention home / / / / / / / / / 7. Adoptive home / / / / / / / / / / VIGNETTE 2 A boy of 15 is adjudicated as a delinquent by the juvenile court because of three offenses for shoplifting. He c.ould be assigned to any of the following programs. Rate each as to the quality of service expected for this case. Least Most Desirable Desirable 1 2 3 4 5 6 7 8 9 10 1. Group home for adolescent boys / / / / / / / / / / 2. County probation of juvenile court / / / / / / / / / / 35 Boys' Training SchoolInstitutional Program / / / / / / / / / / 4. Boys' Training School-Camp Program / / / / / / / / / / 5. County probation serviceDepartment of Social Services / / / / / / / / 6. Foster home / / / / / / / / / / 7. County "youth" home for delinquents / / / / / / / / / / 8. Private institution for delinquent boys / / / / / / / / / / 188

VIGNETTE 3 A child of 3 is referred to the county welfare department by a neighbor because of serious physical and psychological neglect by his natural parents, The agency is granted wardship and could assign the child to the following programs. Please rank each for this case. Least Most Desirable Desirable 1 2 3 4 5 6 7 8 9 10 1. Foster home / / / / / / / / / / 2. Adoptive home / / / / / / / / / / 3. Children's Institute / / / / / / / / / 4. Private residential home for children children / / / / / / / 5. Own home with supervision by protective services unit / VIGNETTE 4 A Black handicapped child of 6 weeks born to an unmarried 14 year old mother is referred to the agency for placement by the hospital staff. The handicap is a mild form of orthopedic deformity which is surgically correctable in a few years. Rate each of the following possible placements for this child, Least Most Desirable Desirable 1 2 3 4 5 6 7 8 9 10 la Own home with supervision / / / / / / / / / 2. Foster home / / / / / / / / / / 35 Adoptive home / / / / / / / / / 4o Rehabilitation Center / / / / / / / / / / 5. Children's Institute/ / / / / / / / 6. Relatives / / / / / / / / 189

VIGNETTE 5 A child of eight is referred by a school social worker because of serious emotional illness and bizarre behavior at school. Service to the parents has not been possible because of their lack of concern. Please rank each of the following. Least Most Desirable Desirable 1 2 3 4 5 6 7 8 9 10 1. Public residential treatment center for mentally ill children / / / / / / / / / / 2. Children's Institute / / / / / / / / / / 35 Local child guidance clinic / / / / / / / / / 4. Local family service agency / / / 5. Private residential treatment / / / / / / / / / / 6. State mental hospitalChildren's unit / / / / / / / / 7. Foster home / / / / / / / / / / Upon return of the student and faculty questionnaire, respectively, the mean and standard deviation for each item in each respondent group were compute These statistics indicated that the two groups of raters were similar with respect to all items as the corresponding distributions of ratings were highly homoscedastic. Thus, it was decided to combine the two sets of scores for each item and to treat them as one homogeneous set of responses which could be analyzed as if it had been obtained from just one group of informants. The average absolute difference of the two groups of raters amounts to.6945 over all 33 items.* Next we checked to determine whether or not the different programs were scored differently depending on the particular child welfare case at hand. As expected the ratings for the same institution vary widely as different clients are to be treated. For instance, adoption is rated the least viable solution for the child whose mother's illness required extended hospitalization (vignette 1) whereas it comes to be seen as a preferred solution in the situation where *None of these differences are statistically significant as they would approximately 1 our of 2 times occur by pure chance. 190

serious physical and psychological neglect by the parents is the cause for referral (vignette 3)0 This indicates that the criterion for program evaluation quite clearly is a function of the case at hand and that the five hypothetical situations must be studied individually. As obvious as this may seem, differential evaluation is not applied to all program alternatives to the same degree. Thus, the Children's Institute appears to be a rather comparable solution for all cases for which it was given as a program alternative. Nevertheless, the overall complexity of the rating process makes individual analysis for all five case situations mandatory. Even though "perceived quality of service" was given as the criterion of evaluation, it does not follow that this criterion denotes the same entity for all respondents. If it did, a one-dimensional solution would have to result for all five case situations, respectively. However to the extent that "perceived quality of service" connotes either several dimensions in general or implies a different order of perspectives for different raters (and hence different ranks of the alternatives) a two- or more-dimensional solution of ranking criteria is required, Depending on a respondent's position on a custodytreatment continum, perceived quality of service will mean quite different things. Preliminary analyses of the data have been completed using the GuttmanLingoes Smallest Space Analysis0 Results from this analysis are presented in the following pages for vignette 2o It can be observed that a two-dimension solution has some definite advantageso Clear preference is expressed by the raters for retaining this type of client in his own community and providing services to him there rather than commitment to a state training facility. Likewise, the raters prefer those types of setting where intervention is only in terms of selected role behaviors rather than a more holistic attempt at control of all major roles of the cliento Additional and new analysis procedures are currently underway so that this instrument can be refined. We also plan then to test its utility with practitioners in a sample of field settingso 191

o o - w'0 00 UN tf\ rC <M I,o0 0 0 0 0 0 0 o. ON9- 7 oj 0 W w al l w W w w O W 0' 4 N 0' 4 -u L 0n'0 IL VWI 0 N m 0 7 A M a, G 0' 0 UI E'f - I Pf u in (a' OI a N Cp 0 4 N IA go 0' 0 N >0 * i-.o IA'- 7. - 7cu 7.. 7.-c V V 0 0 9 S S S S S 0 0 > LU U~~~~~lij LU UJ UJ~~M 0 j LUi 4 N N ~~~~~~~~4 N N+ N- C- N U' o o - a, w iA uj o - rM O N ( ^ m n cn n 0 0 0 0I I Iv a a a a 0 0 0 0 0' u C W LU o UJ, LW W uj'4' ~ ~ ~ ~ ~ a 0'ir' - - IA u *o I 0 i IA 0 4 0n I A I r re i'0 (/I I I I ) a, a, n a, 4 Nl CY aN % N ) 0' N N - 0 0 < - WN' a, tm tlr! c-o w Le\ O -t ]O CII " W I-~~~~ 4~ N.- 4 A 0 5t sC cc C, It C) mC o o 0 aa 0' 0 ~fC o LU ~ ~ a a V) o co0 < LU LU U J IU l LU U LU U~ W r 7- co, N t7- 0 IiC V) < cc o st m 1-f co t Oc rt < o (^ ooN oo oa. O cc, r- c 0C 0 0L c N f 0' r0 LA co,. UJ o I N N N N N N N N N z 0 0 o- 0 0 0 a1.0 r- - LA - a, N 7. c0 L oC a, 04 0 a, UN o 0 1 -4 N N % P zoo.... Z Z ^ ~ Wa;C CO 0 0 0 0 0 0 0 0a ~~ooo o o c o o IWLJ (/) l, a QI I ILU ~~ ~ ~~~~~UJ J L UJ LL LJL UJ 111: 0400' r -4 No o 4) LA 1 7- a, 11,f rctn(2j rC o ~~' <c'^ )Io ocN- c' ~ - c zoo z. ~... o-0L -0 r (*<~~~~~~~9 UJ I 0 CO * (^ ~~ I ujO a. </> CO <IQ LCOU U U L~U ~ L U 41~rn \ a: (/ h- co s h ^o u'c aIo I I I I Ia Icr si fc i ^ co < ) oc U~a C1 LJ * * CI </) ~ J 0 0I I r <I~ in 0s < * I ol oo r* 0 192

z LU LU uj L, 0 0 uul o 0 0 - 1 I I It I I - I II I U \L N m m o' 0' 0 4I~~~~~ NI I I I I 0 N1 I0 C- C0 r-0 0 0""*< 0-" 0 00 0 0 rc C-s 0 - 0 - Ncr o a M. r Z U U U U U U U 0 3 ~ 3 ^ 3 3 3 - 3 C C C C C C oC O0~o' \j ~ - s 0** **

(- <^ I- -z o ~-. 0L z I LI) I 4, >; I zr z l I C) I I I - z O o3 I I l 4,) I I I I I I I ie 2: I I - h- UI 00 0,I -tm% IU I I - 4,)U z I N0 JI I I I I I I I Il 2 I 0 > Icc* 5 U.:, 0 a z LU I II LL ZL/ I S N Z L Li Z. 4U, O V 0. I I C IV I I I) 1 - <~ >: UJ 2' O h- 0 ui CjI Ir I Z z I ~C C) I Z - Q I F 2- I (.. ~ LU I L Ui ~~~ Q I C)r I I 00I O l % I- oI I Nr-o f II-' 4,) V)iI I QI4 L Z zr I *. *... * L O w I 0NI I i r I -- (~ / I I I I I NI 4 UN LU I i l- l t I c o LUx. I ooocL0ol ro - C,,j LL oc,2: ** LU. UJ II I ( \JI ul U) ^ u-4 L\C0 O4 00 V) 4,J Lt~t'- r U 4 u I 2: - i no I I I Irlr oo oo cL/ o I 1 o " I I I OLu 0n I I oI c l ~c r\ t I-~ ~ ~~~a o~ oIoINos o - uj ZIo ch cr^ uo > 0t I LU ^ ( I | 0 UJ ) ( X LI I 2: - fr W I 00 I-LI) ~-J 2- p-1 cc~ d II u Ic I U I > C < < ar: ~- U 1 ly CIS uj1~ >- (N n rCIoo t~~oo ~^- rl 5: ac 1 1 - 3 194

z* o )LJ 0 l2: * C I LII V) I I z N N l Q t I J - I _ I 1 O r w II I C) LU C I C) c- < L UI11 t, U 0 0 0 0 W U)I I %C C) U II C) * ll I O- tr ll 0 JJ t/) Z> Z I N 0; I 2: C) I O r'ItOO V0! * * * * * * * LU 7 C)I I I tro o - 0. I LC\NC)00..VP - 0 0 0 C o,C I( C\ I - > I %0 rlr- Io oj (n U LU U I: C)* * IL Vz V) I 7 Q ) I I^ O.%lro FM \4lcr ( 4 4 I I lu P I t t U,) U 2: r-I NC)OO OOI " - 00.4 I * * * * * * * * - > Lr I o o o u4 CLU c/ ) r o - z I(7S- r^ uuoo 07 0 ( r r ( |I| CD. U UJ 0I I I U Q 2: - WI N - I.- - 1 - -J U ) of > I

o^^ culo oo o4N al<olo~o ruir ^< J^' u alf t o<\ j uoo fM~ab ~i~ 3 (p ol~ccu rtr^ ^^uoiNI~ oo1~ f^^f lo oooo cu o OD ODoalott311 I\VIr 1 i ul ur cr~l(IV UC) n~~~ul\aa O~taa1bOII C* l ( I I * t l ( I l ( I l I l i l | I | ^ 0 0 0* * * 0 0 ~ * * ^^4..~ * ~ ~5 u * 0* ~~~~~~* *0i * ~~~~~~~~~* *< * * 01 * 3 ~~ 0 0) * *~ * ^ * * I * * * * t o) LIMT 0 0 0t* * - *0 * * i*v S* * *- 0 0 z O w <M** *N * -^~< 0* *' 0 * 0 V " C0 LUJ ~j LNLJ^ w 2* * * I O"N* *N 0S 0 * * T * 0 * -* C *0 o* -* i^* ~ *^N ^* C4Ml 0*l I I I I I I I I I I I i * i I ** + I ^~,. I (H s I n I I I'T 9 * ^* *9E * * <:3>* La* * I C "4'& * 3 9o^ </ 0* SJ* ^*0 I * k I I I ~ ( I- *? r L- 0 ^ J ~0 Q. I ~* ~ ~~~ * ~ ~ * ~ ~~~ ~ ~' I ~ I ( *~ ~~*~*~*~~~'<~~**~~~*~*~~ ~ I 0 *^ffO40OMC' ~()O^00MO''<)0( *^O^O''~~'O OPO ^ OMO ^~MO 00 00 OI I I II~l ll I.~ *1 3 ~ I- 0* * *..J "4* * 9- N~~~~~~~~~9 0 l 0'0N@04.0'0N0'0N0'0N@04.0I N 4. N'004.@0N'00N'00N O04.@0NcO 0 @Nin~N e J" " "4 9 N NNI I I @ 0 0' 0'0 I Inc'- 9999999999999999999999- cT'31 1 I I I~l 00 ~ 0 n I I I I I I,~ wJ w w n I I I I I I I I k I I I > 0 i I I I i I I I I I I I I I I I 196I

s) z U. UUJ 0 Co~D ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~S 0 I I I I 1 I I I I I I r-< o o a a av CCre( O 0 O reC Cr~ I I I I (i (c o 0 C 0 0 0 0 0 197

APPENDIX 2 199

2.A. SURVEY QUESTIONNAIRE ENTITLED "TO PROVIDE HOPE"-FOR ADMINISTRATORS The School of Social Work at The University of Michigan would like to gain a more adequate picture of the policies and practices of the different County Welfare Agencies. The present questionnaire was designed to collect the relevant information. We are all too well aware of your limited time and how much we must impinge on your professional work as we ask you to complete this questionnaire. Yet we do not know of any other way to get the information needed to study, compare, and hopefully improve the services presently rendered. Thus we ask you to give us some of your time that will enable us to learn from your insights and experiences, Please rest assured that all information will be treated strictly confidential and that your questionnaire answers will be seen by no one except the research staff, Thus do not sign your name on this questionnaire. Almost all questions can be answered by making a simple check mark in the box beside the answer you choose. Please ignore the extra numbers under the answers and in the margins to the right of the double line. These are there to help in tabulating the answers by IBM machines. We should like you to feel that you are expressing your true feelings, opinions, and views as you answer the questionnaire. Please write your comments in the margin or at the end of the booklet when you feel a question is unclear or does not allow you to express exactly how you feel. If you should be interested in receiving a report of the findings of this study, please sign the enclosed postcard and mail it separately from this questionnaire, As soon as you have completed this questionnaire, please put it into the preaddressed envelope, seal it, and return it to us at your earliest possible convenienceo We hope you will enjoy filling out this questionnaire and would like to thank you for your cooperation. 201

CODE IBM 0 01 Do not sign your name Do not sign your name 939 02-04 01 05-06 THE UNIVERSITY OF MICHIGAN SCHOOL OF SOCIAL WORK 07-09 ANN ARBOR, MICHIGAN 10-11 INTERSTATE STUDY OF COUNTY WELFARE ORGANIZATIONS 12-14 (DIRECTOR'S QUESTIONNAIRE) To begin, we shoutd tike to ask a ew queation6 about yout county depaotnent in geneAa. We aAe acwre that pedhap4 some of the quetions arte difdicuZt to anmuwe o04 do not futLy apply to youw. 4ituation. 16 thiLs houed be the cae, ptea.e let u4 know by writing in the margin. 1. How long has your county department been in existence in its present form? e__ars Y a 15-16 2. How many client applications (all programs from all sources) were made to your agency during 1967? Applications 17-21 3. a. How many of these applications were submitted for the first time (excluding applications for food stamps)? First-time applications 22-26 b. How many applications for food stamps only were made to your agency during 1967? Food stamps 27-31 c. Most organizations are unable to help all applicants. How many of the 1967 applications (as mentioned in question 3a) were you unable to help? Not helped 32-35 202

CODE [ IBM d. We realize that you were not able to help everybody to the same extent. Please estimate the percentage of those that you were able to help completely, somewhat, or only a little. % Completely 36-37 % Somewhat 38-39 % Little 40-41 4. How many of your agency are in each of the following categories? Executive and administrative 42-43 ___ Supervision 44-46 Research 47-48 Total number of welfare workers 49-51 _ Full-time equivalent welfare 52-54 workers (except indigenous workers) Total number indigenous workers 55-57 Full-time equivalent 58-60 indigenous workers Secretarial and clerical 61-63 Other professionals 64-65 Full-time equivalent lawyers 66 Full-time equivalent 67-68 psychologists Full-time equivalent 69-70 psychiatrists Full-time equivalent clergy 71 Full-time equivalent nurses 72-73 Other (specify: )74-75 Total staff 76-79 203

CODE IBM D 01 We now move into an area wheAe we would tike to know a tettte bZt 939 02-04 about the admintArative ^tutctaue od youw agency. 02 05-06 07-09 10-11 12-14 5. What is the case load for a full-time equivalent welfare worker for the following programs? AFDC 15-17 Aged 18-20 Aid to the Blind 21-23 APTD 24-26 General Assistance 27-29 Medicare & Medicaid 30-32 Child Welfare 33-35 6. How many full-time professional staff members joined and how many left your agency during 1967? Joined 36-38 _ Left 139-41 7. How important are the following criteria for promotion in your agency? Very Unimportant Important Seniority I I -, I p J 42 1 2 3 4 5 6 Securing M.S.W. I I I I I 43 1 2 3 4 5 6 Competitive examinations I I i,I,, I _ a 44 1 2 3 4 5 6 Political considerations I,,,,,,I 1 45 1 2 3 4 5 6 Demonstrated competence I,,i i, 46 1 2 3 4 5 6 204

CODE IBM Very Unimportant Important Colleague judgment,,,, I 47 1 2 3 4 5 6 1. Other (specify: ),,, I I 48 1 2 3 4 5 6 2. Other (specify: 1 2 I I I 1 49 1 2 3 4 5 6 8. We are interested in knowing how welfare workers define their job. We would like to get at this by asking you to rank several other jobs, which are listed below, according to their similarity to welfare work. Please write the figure "1" next to the job which most closely resembles welfare work in your opinion; then write "2" next to the one which is next most similar; and so on down to number "7," the one least similar to welfare work. Please do not use the same number more than once, and please assign a number to each of the jobs. Job Rank Clergyman 0__ _ Detective ___ 51 Psychiatrist _ _ 52 Lawyer _ 53 Bookkeeper _ 54 Teacher __ _ 55 Automobile insurance claims adjuster _ 56 9. a. There are about nine major b. Since it is not always possible 0 01 areas in which a welfare to put first things first, agency is called upon to please tell us what percentage 939 02-04 help. Please indicate what (again to the next 5 or 10%) of 03 05-06 percentage (to the next 5 your total effort (time and 07-09 or 10%) of your total effort personnel) should ideally be 10-11 (time and personnel) is allocated to each of the 12-14 actually allocated to each following areas of concern: of the following areas of concern: Actual Ideal 1. Financial aid % __ 15-18 2. Family-centered problems % __ 19-22 205

CODE IBM Actual Ideal 3. Child welfare services _ % X __ % 23-26 4. Mental health and addiction _ % %__ 27-30 5. Migrant workers _ % _ % 131-34 6. Aged citizens' care _ % _ % _ 35-38 7. Medical care % % 39-42 8. Legal services _ % % L 43-46 9. Permanent and/or total disability _ % _ % 47-50 10. Other (specify: ) % % _ 151-54 11. Other (specify: ) % _ % ____ 55-58 100% 100% 10. There are only five days to a week and thus one cannot always give as 0 01 much attention to each of these nine realms of concern as one might want. Please tell us which of the following nine areas you feel is 939 02-04 most adequately (1), second most adequately (2),... least adequately 04 05-06 (9) served by your agency. Do this with regard to when you took your 07-09 office, with regard to now, and with regard to how you would like it to 10-11be. 12-14 Past Now Ideal 1. Financial aid ___ 15-17 2. Family-centered problems _ 18-20 3. Child welfare services _ __ __ 21-23 4. Mental health and addiction _ _ 24-26 5. Migrant workers _ _ 27-29 6. Aged citizens' care _ __ 30-32 7. Medical care 1_ _33-35 8. Legal services _ _ 36-38 9. Permanent and/or total disability ___ 39-41 10. Other (specify: ) ___ 42-44 206

CODE IBM 11. Comparing your own agency with other county welfare agencies, would you say you are better off than most of them, pretty much like all others, or are you poorer than most of them with regard to Better Same Poorer a. Adequacy of funds ___ __ 45 1 2 3 b. Adequacy of personnel 46 1 2 3 c. Adequacy of service facilities 47 1 2 3 12. One of the ways used in research to help understand people's opinions and attitudes is to ask them to complete sentences by saying the first thing that comes into their minds. Thus, please finish the following sentences, putting down whatever comes to your mind as you read the opening phrase. a. I can't understand why so many social workers 48-49 50-51 b. It's a shame that many clients 52-53 54-55 c. I can't understand why our county agency _ 56-57 58-59 d. It's a shame that our financial resources 60-61 207~____~~________~ _62-63 207

CODE IBM e. I wonder why the government 64-65 66-67 f. It's strange how people 68-69. _____________________ _70-71 13. Every agency has some rules and regulations for employees. We have 0 01 listed some of the workers' behaviors which violate such rules. Would you mark the order of disciplinary measures you would actually take in 939 02-04 the case of deviance. In this case, "1" is the first thing, "2" is the 05 05-06 second thing you would do, etc. 07-09 10-11 12-14 Sanctions Applied Workers' Behavior Discharge Suspend Reprimand Talk Do Him Him Him to Him Nothing a. Is absent a lot 15-19 b. Is often late ___' ~_______________________ _ __ _20-24 c. Doesn't follow the rules 25-29 d. Unprofessional conduct...._______3o~~_________________________~ _30-34 e. Does other business on the job 35-39 f. Is disloyal to agency...._______________________ ____________________ _40-45 208

CODE IBM 14. On the other hand, we would like to know how your agency reacts to exemplary behavior of the workers. Please indicate the order of rewards you are most likely to give in the following situations. (Follow the same procedure as above.) Rewards Given Situation Delegate Salary Special Give Do Promotion Increase Authority Praise Nothing a. Does more than is expected.........__________ 46-50 b. Is always prompt with work ~______________________ ___ _~___________~_____ 51-55 c. Takes initiative......_____________________'....... 56-60 d. Shows creativity on the job..........._...._._..._________ _61-64 e. Cooperates with other staff..... ___________________ __________________________ _____65-69 f. Is conscientious.................._______________________________ 70-74 15. The following series of statements concern supervision in your 0 01 department. We would very much like to have your opinion as to how true or false each of the statements is as far as your department is 939 02-04 concerned. (Circle corresponding number.) 06 05-06 07-09 On the whole, I believe the statement is: 10-11 12-14 Definitely Probably Probably Definitely True True False False a. There can be little action taken around here until a supervisor approves a decision 1 2 3 4 15 209

CODE IBM On the whole, I believe the statement is: Definitely Probably Probably Definitely True True False False b. A person who wants to make his own decisions would quickly be discouraged here 1 2 3 4 16 c. Even small matters have to be referred to someone higher up for final answer 1 2 3 4 17 d. I have to ask the director before I do almost anything 1 2 3 4 18 16. Now, we would like to have your views about other aspects of your work and your department. a. Which is more important in getting along in this department: how much one knows or whom one knows? (Circle corresponding number.) lHow much one knows 1 Whom one knows 2 19 b. How much influence do the State and Federal offices have on the way this department is run? A Lot Some Little Very Little State Office 20 1 2 3 4 Federal Office 21 1 2 3 4 210

CODE IBM 17. Welfare departments around the country differ somewhat in their objectives regarding the AFDC program. The following statements express some of these differences. For each of them, please indicate in column A whether it agrees with the policy in your department. Then in column B indicate your own agreement or disagreement. (Circle the corresponding answers.) A B Department Personal Objective Preference a. To provide financial assistance to the Yes No Yes No maximum number of eligible families 1 5 1 5 22-23 b. To emphasize rehabilitation through provision of services rather than Yes No Yes No financial support 1 5 1 5 24-25 c. To continue assistance as long as it remains helpful rather than terminate Yes No Yes No it as soon as possible 1 5 1 5 26-27 d. If there is a shortage of money, to provide small grants to assist a Yes No Yes No relatively large number of families 1 5 1 5 28-29 e. To maintain grants at high levels even if it means assisting a smaller number Yes No Yes No of families 1 5 1 5 30-31 f. To concentrate on legally defined obligations toward the client, rather than trying to go beyond existing Yes No Yes No provisions to help the clients 1 5 1 5 32-33 18. Welfare workers not infrequently experience considerable role conflict in trying to reconcile various responsibilities. Please tell us to whom you feel the workers owe most loyalty (1), second most loyalty (2), third most loyalty (3), and least loyalty (4) among the following alternatives. Amount of Loyalty Profession _ 34 Client 35 Se f __ 36 Department __ 37 211

CODE IBM 19. In some public welfare departments there are situations or conditions which make it hard for caseworkers to provide adequate services to their AFDC clients. Some of these situations are listed below. Have any of these made it difficult for you to provide services to your AFDC clients? Please check all situations either true or false. True False a. Case toads are too large 38 1 5 b. Too much paper work 39 1 5 c. Inadequate training ___ 40 1 5 d. Conflicting directives 41 1 5 e. Too many routine decisions require approval at higher levels 42 ~~1 5 f. Pressure from persons and groups outside the department 43 1 5 g. Too much responsibility for continuing eligibility determination 44 5 h. Inadequate supervision _ 45 5 i. Shortage of caseworkers _ 46 1 5 j. Client hostility _____ 47 5 k. Public hostility _ 48 1 5 49-50 I you have come thi4 6ar in an6weAing the questionnaire, you are prtobaby wondering when it i6 going to end. Actualty, we are coming ctose and ask youa kind cooperation for anotheAr de pages. We know we aAe asking a tot of you. But you ptobably agree that the study vwit produce the moat significant body of info-mation that we have eveA had about weleaAre agencies. And you ate entteed to Seet a certain pride in having been pant of it. 212

CODE IBM 01 Now we woud Like to preent a 6eAieA of 6tatement6 to you 6ort you 1 939 02-04 conzideAdtion. We 4houtd tike to know both youwi own pemonat opiiZon and youx e6timaQte of youw fuiend6' attitude on the^e item6. Ptease u4e the 07 05-06 fotowing categoaie6: 07-09 SA = STRONGLY AGREE 10-11 A = AGREE 12-14 MA = MILDLY AGREE MD = MILDLY DISAGREE D = DISAGREE SD = STRONGLY DISAGREE 20. It's a shame that our organization can't do more to prevent misery rather than just alleviate it. You yourself SA A ~MA O MD WD SD 15 1 2 3 4 5 6 Your fr- snds WQ SA O A O MA O MD O D W SD 16 1 2 3 4 5 6 21. Except when there is a depression, anyone in our country can get a job if he really tries. You yourself W SA W A a MA MD [ D [ SD 17 1 2 3 4 5 6 Your friends W SA O A - MA W- MD WO D W SD 18 1 2 3 4 5 6 22. It's all right for welfare workers to cancel meetings with me when they are pressed for time. You yourself E SA W A W MA A- MD W- D W SD 19 1 2 3 4 5 6 Your friends O SA O A O MA W MD O D O SD 20 1 2 3 4 5 6 213

CODE IBM 23. Sometimes my welfare workers overidentify with their clients. You yourself DSA W A D MA D MD W D SD 21 1 2 3 4 5 6 Your friends O SA A [A O A MD O D ~ SD 22 1 2 3 4 5 6 24. Having to struggle for what you get in life is the best way to develop character. You yourself W SA OA D MA D MD O D D SD 23 1 2 3 4 5 6 Your friends 2 SA D A D MA D MD D D D SD 24 1 2 3 4 5 6 25. The main purpose of our agency should be to provide the poor and needy with the necessary money. You yourself O SA OA D MA D MD OD C SD 25 1 2 3 4 5 6 Your friends SA A D MA D MD D D O SD 26 1 2 3 4 5 6 26. I don't like the idea of letting clients tell us how to run this place. You yourself W SA W A D MA D MD O D D SD 27 1 2 3 4 5 6 Your friends D SA D A D MA D MD O D n SD 28 2 3 4 5 6 27. Most people respect welfare workers. You yourself D SA D A D MA D MD O D D SD 29 1 2 3 4 5 6 Your friends D SA D A D MA W MD O D D SD 30 1 2 3 4 5 6 28. It's a public stigma to accept help from our agency. You yourself D SA A D MA D MD O D O SD __ 31 1 2 3 4 5 6 Your friends SA W A ~ MA D MoD D D D SD 32 1 2 3 4 5 6 214.

CODE IBM 29. Welfare workers should meet deadlines for review and budgets even if it means a delay in providing assistance. You yourself G SA A G MA G MD G D G SD 33 1 2 3 4 5 6 Your friends SA A JMA MD GOD SD 34 1 2 3 4 5 6 30. Most clients think we are doing a pretty good job. You yourself 2 SA W A [ MA [ MD Ga D 1 SD 35 1 2 3 4 5 6 Your friends G SA EC A MA G MD GE] D G SD 36 1 2 3 4 5 6 31. Welfare recipients should not be entitled to vote. You yourself G SA O A W MA G MD W D G SD 37 1 2 3 4 5 6 our friends G SA G A O MA Ga MD -1 D [] SD 38 1 2 3 4 5 6 32. By and large, black welfare workers should take care of black clients. You yourself GU SA O A GMA G MD GO D [SD sD 39 1 2 3 4 5 6 Your friends G SA ] A G MA G MD G — D G SD 40 1 2 3 4 5 6 33. The criteria of our state of eligibility for help are poor. You yourself G SA O A [ MA G MD ] D SD 41 1 2 3 4 5 6 Your friends G SA G A - MA [ MD G] D [ ] SD _ 42 1 2 3 4 5 6 34. Welfare recipients generally are unwilling to accept responsibilities. You yourself G SA A G MA G MD G D G SD _ 43 1 2 3 4 5 6 Your friends G SA [ A I MA [ MD [I D I] SD 44 1 2 3 4 5 6 215

CODE IBM 35. Welfare workers are not adequately paid for the work they do. You yourself L SA A IMA MD Z D O SD 45 1 2 3 4 5 6 Your friends O SA O A MA [] MD I D O SD _ 46 1 2 3 4 5 6 36. Any able-bodied individual who refuses to take a job should not receive assistance. You yourself C SA O A 3 MA LM D O D O SD __ 47 1 2 3 4 5 6 Your friends [ SA L A C MA I MD I D LI SD 4 48 1 2 3 4 5 6 37. People can actually do very little to change their lives. You yourself U SA I A I MA L MD L D O SD _ 49 1 2 3 4 5 6 Your friends L SA EA MA A MD IO D O SD 50 1 2 3 4 5 6 38. Welfare workers should belong to unions. You yourself I SA W A W MA MD W D SD 51 2 3 4 5 6 Your friends I SA 1 A W MA W MD W D O SD _ 52 1 2 3 4 5 6 39. Welfare recipients, by and large, live an easy life. You yourself L SA A O MA L MD F D ISD $ 53 1 2 3 4 5 6 Your friends aL SA C] A I MA LI MD LI D O SD 54 1 2 3 4 5 6 40. I don't think our agency can learn much from the clients it serves. You yourself L SA W A D MA j MD L D O SD L 55 1 2 3 4 5 6 Your friends L SA L A I MA I MD O D O SD $ 56 2 3 5 6 216

CODE IBM 41. Very few clients will ask for help unless they really need it. You yourself 2 SA A IMA [ ]MD E D E SD 57 1 2 3 4 5 6 Your friends [ SA [ A j MA [I MD [] D [ SD 58 1 2 3 4 5 6 42. Welfare workers, by and large, fail to respect the client's private life. You yourself - SA A A MA O MD I D SD 1 59 1 2 3 4 5 6 Your friends L SA O A MA [ ]MD L[ D D SD 60 1 2 3 4 5 6 43. I expect my welfare workers to see their clients after work if necessary. You yourself SI SA A I MA I MD D L OSD 61 1 2 3 4 5 6 Your friends r7 SA 2 A [ MA L MD []D E SD 62 2 3 4 5 6 44. If a person,ried hard enough he could find work and support himself. You yourself L SA O A O MA IMD OI D ~ SD 63 1 2 3 4 5 6 Your friends ] SA [OA [ MA [ MD a D I SD 64 1 2 3 4 5 6 45. A welfare worker should not talk to his clients about his own family. You yourself L SA OA D MA L MD L D D SD 65 1 2 3 4 5 6 Your friends I SA [] A LI MA ]I MD Z D I SD 66 1 2 3 4 5 6 46. Some welfare workers are too lenient with their clients. You yourself L SA A I MA IMD L D L SD 67 1 2 3 4 5 6 Your friends I SA LI A LI MA ]I MD ZI D [I -SD 68 1 2 3 4 5 6 217

CODE IBM 47. When a married couple with children is having serious problems getting along together, their first consideration should be to keep the family together at all costs. You yourself SA A D MA MD D - D SD 69 1 2 3 4 5 6 Your friends Q SA Q - A - MA F MD - D _-_ SD 70 1 2 3 4 5 6 48. Welfare recipients are basically inadequate people. You yourself W SA W A D MA W- MD W D D SD 71 1 2 3 4 5 6 Your friends W SA W A W MA W MD W D W SD 72 1 2 3 4 5 6 Congatutration! You made it alt the way up to herte, and we ate confident that you wilt continue 4ince you have aLteady completed the majot paAt of the questionnaire. StilZ there aLe a few mote quebtionh that ak 6or youwt 6u attention. 49. As you know, there is a great deal of concern with AFDC mothers these o 01 days, and people have all kinds of ideas about what they are like. Listed below are some words which have been used to describe AFDC 939 02-04 mothers. We would like you to do two things: 08 05-06 First, please give us your opinion as to whether or not people in general would apply these words to AFDC mothers. For each word, circle 07-09 the appropriate answer under "A" below. 10-11 Next, please tell us what proportion of AFDC mothers in your community you believe each of these words applies to. For each word, circle the 12-14 appropriate number under "B" below. A B People in general believe AFDC mothers I believe this applies to: are: Most Some Few Hardly Any 1. Unfortunate Yes No 1 2 3 4 15-16 1 5 2. Law-abiding Yes No 1 2 3 4 17-18 1 5 218

CODE IBM People in general believe AFDC mothers I believe this applies to: are: Most Some Few Hardly Any 3. Promiscuous Yes No 1 2 3 4 19-20 1 5 4. Scheming Yes No 1 2 3 4 21-22 1 5 5. Decent Yes No 1 2 3 4 23-24 1 5 6. Responsible Yes No 1 2 3 4 _ 25-26 1 5 7. Immature Yes No 1 2 3 4 27-28 1 5 8. Religious Yes No 1 2 3 4 29-30 1 5 9. Lazy Yes No 1 2 3 4 31-32 1 5 10. Maternal Yes No 1 2 3 4 33-34 1 5 11. Exploiting taxpayers Yes No 1 2 3 4 35-36 1 5 12. Immoral Yes No 1 2 3 4 37-38 1 5 13. Deserving Yes No 1 2 3 4 39-40 1 5 14. Stupid Yes No 1 2 3 4 41-42 1 5 15. Dishonest Yes No 1 2 3 4 43-44 1 5 16. Conscientious Yes No 1 2 3 4 45-46 1 5 17. Greedy Yes No 1 2 3 4 47-48 1 5 18. Family-minded Yes No 1 2 3 4 49-50 1 5 219

CODE IBM People in general believe AFDC mothers I believe this applies to: are: Most Some Few Hardly Any 19. Hardworking Yes No 1 2 3 4 51-52 1 5 20. Grateful Yes No 1 2 3 4 53-54 1 5 The 6oUlowing item { (to the Zeft of the page) have been chosen to O 01 deQ4Cibe your agency. On the horizontal tine two ZabetL indicate the extent to which they may be tAue oa yout agency. Treat each o0 them a a 939 02-04 continuous scate from the extPLe:e o6 one end to that od the otheA. PZeaQe do three thungs as doaZouw: 09 05-06 1. Kindly put an n (= now) on each line at the point which, in your 107-09 experience, describes your agency best at the present time. 10-11 2. Kindly put a w (= want) on each line at the point where you want to have your agency fall with regard to that item. 12-14 3. Kindly put a p (= past) on each line at the point where you feel your agency was when you took office. Here i. an example. Suppose you now feet usually veAy close to youL 6ocuia workeru; put an n at the 6aoA edt. But when you took odflce you delt you weAe uveAy faA apaAt fAom them; put a p at the 6la tight. And you deet youw prtesent retationzhip is just ideat; put a w on top od the n. Example: Usually Far Extent to which you feel very close apart close to your welfare W workers n, n I,.,, a, p J (Should you need to difereentiate betoeen diferent agency group6 feet f6ee to do so any place you want to.) 50. Extent to which the It always It seldom agency tries to use the tries to tries to ideas, views, and do so do so opinions of your welfare workers 15-17 8 7 6 5 4 3 2 1 220

CODE I BM 51. Extent to which the It seldom tries It always tries agency tries to use your to do so to do so clients' ideas, views, and opinions i,, 18-20 1 2 3 4 5 6 7 8 52. Extent to which you Feel not at Feel completely feel the welfare workers all free to free to do so feel free to discuss do so agency policies with you,, 21-23 1 2 3 4 5 6 7 8 53. Extent to which you make Very Very yourself available to much little individual welfare workers 24-26 8 7 6 5 4 3 2 1 54. Extent to which the Very Very welfare workers are much little sympathetic to their clients 27-29 8 7 6 5 4 3 2 1 55. Amount of responsibility Very Very your welfare workers little much show toward agency money 30-32 1 2 3 4 5 6 7 8 56. Extent to which you permit welfare workers to bend Very Quite rules and regulations in little a bit order to meet special case situations,,,, 33-35 1 2 3 4 5 6 7 8 57. Extent to which you feel Very Very the welfare workers much little should become advocates of clients' rights, i,,,, 36-38 8 7 6 5 4 3 2 1 221

CODE IBM 58. Extent tb which you feel Very Very the welfare workers much little accept agency policies made by you, i _ 39-41 8 7 6 5 4 3 2 1 Very Very 59. Extent to which welfare little much workers make their own decisions,, I | I 42-44 1 2 3 4 5 6 7 8 Not at Very 60. Extent to which your all much welfare workers cooperate with one another I I I I, 45-47 1 2 3 4 5 6 7 8 61. Extent to which you feel Very Very your welfare workers keep little much up with new developments of social welfare practice,,,,, 48-50 1 2 3 4 5 6 7 8 62. Extent to which you have Full No confidence trust and confidence in confidence at all the ability of your welfare workers,,,, i i 51-53 8 7 6 5 4 3 2 1 63. Extent to which you think Full No confidence your welfare workers have confidence at all trust and confidence in your ability, I I 54-56 8 7 6 5 4 3 2 1 64. Extent to which you feel Not at Quite government policies all a bit interfere with the work you think needs to be done, i, 57-59 8 7 6 5 4 3 2 1 65. Extent to which you feel Quite Very politicians try to a bit little interfere with the work of your agency, I, 60-62 1 2 3 4 5 6 7 8 222

CODE IBM 0 01 Exce2tentl You've got it made. Just a ew moae que.tion6 about youA 939 02-04 peVuonat background. ALthough they ask about you, keep in mind that the 10 05-06inounmation Z6 to be u6ed in a 4tat4etica way oney, much tike the U.S. 07-09 censu6s. 1O10-11 12-14 66. What was your age on your last birthday? Years 15-16 67. Which of the following degrees do you hold? (Check as many as apply.) High School 17 B.A. or B.S. 18 M.A. or M.S. 19 Certificate in Social Work 20 _, M.S. W. 21 L.L.B. 22 _ M.P.A. 23 Ph.D. 24 _ D.S. W. W. 25 68. Social Work Training A. In school. Classroom Hours Supervised Field Work Undergraduate hours 26-29 Graduate hours 0 ____ 30-33 Other graduate training _ 34-37 3. On the job. Describe the training you have had on the job for this or other social work positions held. List the number of any special training courses, conferences, institutes, etc., which you have attended. No. of conferences 38 No. of extension courses 39 No. of special institutes 40 _ No. of leadership training seminars 41 Other (specify:. _ 42 Other (speci fy:) 43 223

CODE IBM 69. What is your official grade here at the present time (if appropriate)? Grade 44-45 70. How long have you held your present position? Years Months 46-49 71. Number of years and/or months you have worked in Public Assistance. Years Months 50-53 72. What was your first job? First job 54-55 73. Would you tell us what job your father had when you were about 16 years old. Father's job 56-57 74. Please indicate your annual salary. Less than $10,000 $14,000 - $15,999 1 4 $10o,00 - $11,999 $16,000 - $17,999 2 5 $12,000 - $13,999 $18,000 - $19,999 3 6 $20,000 or more 58 7 75. What is your race? American Indian Chinese 1 3 ____ B lack Japanese 2 4 Caucasian 59 224 224

CODE I BM 76. What is your religious preference? Protestant 1 Catholic 2 _ Jew 3 None 4 __ Other (specify: ) 60 5 77. What is your sex? Male _ Female ____61.What is your marital status? Single (never married) GO TO QUESTION 80 Married (first time) 2 Remarried 3 Separated 4 Divorced 5 ___ Widowed _ 62 6 79. How many children do you have? _ Children 63-64 225

CODE IBM 80. Would you like to see your children become welfare workers? Yes No 65 1 1 5 1 80a. Why? 80b. Why not? 81. Generally speaking, do you think of yourself as a Republican or a Democrat? Republican Democrat Independent Other _ 1 1 1, I Would you call yourself a Do you think of yourself strong Republican/Democrat as being closer to the or not a strong Republican/ Republican or Democratic Democrat? party? Strong Republican _ Republican 1 3 Not strong Republican _ Democrat 2 5 _____ Not strong Democrat ___ Neither 6 4 Strong Democrat 7 ____________________________ _________________________ _67 82. Would you mind telling us for whom you voted in the recent presidential election? Did you favor Humphrey, Nixon or Wallace? ____ Humphrey _ Nixon _ Wallace _ None of above 68 1 2 3 4 Now that you have completed this que^tionnaire, just.stick it into the envelope, seal it, and dLop it into the nearesmt maiebox. Thank you so much ot youA coopewtion. 226

This page has been left empty for any comments you care to make. V-A-2 227

2.B. HOW RESPONSES WERE SOUGHT FOR COUNTIES WITH IRREGULAR ORGANIZATIONAL DESIGNS When a Director was responsible for one or more counties in a region included in the sample, a special attachment and supplements were included with the questionnaire being sent to that particular Director. Likewise, a unique form letter was sent to him. The special attachment was stapled to the first page of the questionnaire explaining the information needed. For example, when one Director is responsible for more than one county-including only one in our sample the attachment is as such: Do not sign your name Do not sign your name THE UNIVERSITY OF MICHIGAN SCHOOL OF SOCIAL WORK ANN ARBOR, MICHIGAN INTERSTATE STUDY OF COUNTY WELFARE ORGANIZATIONS (DIRECTOR'S QUESTIONNAIRE) To begin, we should like to ask a few questions about your county department. The county randomly selected within your jurisdiction is. Do you have separate data for this county? /7 Yes. If so, please give /a No. If the county data is these statistics to questions not separate, but rather, 1 through 6 in the questionnaire collected as a whole in your where applicable. department, then please state the totals for both of the counties in the questionnaire. We are aware that perhaps some of the questions are difficult to answer or do not fully apply to your situation. If this should be the case please let us know by writing in the margin. 228

No supplements were included because we were requesting information on only one county. However, in the same situation when we were sampling more than one county in a Director's region or district the attachment used was this: Do not sign your name Do not sign your name THE UNIVERSITY OF MICHIGAN SCHOOL OF SOCIAL WORK ANN ARBOR, MICHIGAN INTERSTATE STUDY OF COUNTY WELFARE ORGANIZATIONS (DIRECTOR'S QUESTIONNAIRE) To begin, we should like to ask a few questions about your regional office. The counties randomly selected within your region are and Do you have separate data for these counties? / / Yes. If so, please give - these statistic pleae give No. If your county data is not these statistics on the attached supplements for each separate, but rather, collected as county where applicable. Then a whole in your region, then please continue with question 7 on state the totals for your entire page 3 in the questionnaire region in the questionnaire. We are aware that perhaps some of the questions are difficult to answer or do not fully apply to your situation. If this should be the case please let us know by writing in the margin. Here two counties in the Director's region were sampled. 229

When three counties were sampled the attachment used was: Do not sign your name Do not sign your name THE UNIVERSITY OF MICHIGAN SCHOOL OF SOCIAL WORK ANN ARBOR, MICHIGAN INTERSTATE STUDY OF COUNTY WELFARE ORGANIZATIONS (DIRECTOR'S QUESTIONNAIRE) To begin, we should like to ask a few questions about your field office. The counties randomly selected within your jurisdiction are ______ and. Do you have separate data for these counties? /7 Yes. If so, please give / No. If your county data is not these statistics on the separate, but rather, collected as attached supplemerts for each a whole in your department, then county where applicable. Then please state the totals for your continue with question 7 on page entire area. 3 in the questionnaire. We are aware that perhaps some of the questions are difficult to answer or do not fully apply to your situation. If this should be the case please let us know by writing in the margin. Where more than three counties were sampled a corresponding number of blanks were included to indicate which counties these were. The other alteration in the attachments was the name used in that state for the areas of jurisdiction. These names were: region, district, field office, and jurisdiction. Examples of the possible attachments used follows. 230

Do not sign your name Do not sign your name THE UNIVERSITY OF MICHIGAN SCHOOL OF SOCIAL WORK ANN ARBOR, MICHIGAN INTERSTATE STUDY OF COUNTY WELFARE ORGANIZATIONS (DIRECTOR'S QUESTIONNAIRE) To begin, we should like to ask a few questions about your field office. The county randomly selected within your jurisdiction is _____. Do you have separate data for this county? /7 Yes. If so, please give / No. If the county data is these statistics to questions not separate, but rather, 1 through 6 in the questionnaire collected as a whole in your where applicable. department, then please state the totals for your entire area. We are aware that perhaps some of the questions are difficult to answer or do not fully apply to your situation. If this should be the case please let us know by writing in the margin. Do not sign your name Do not sign your name THE UNIVERSITY OF MICHIGAN SCHOOL OF SOCIAL WORK ANN ARBOR, MICHIGAN INTERSTATE STUDY OF COUNTY WELFARE ORGANIZATIONS (DIRECTOR'S QUESTIONNAIRE) To begin, we should like to ask a few questions about your county department. The county randomly selected within your jurisdiction is ___. Do you have separate data for this county? /7 Yes. If so, please give /7 No. If the county data is these statistics to questions not separate, but rather, 1 through 6 in the questionnaire collected as a whole in your where applicable. department, then please state the totals for both of the counties in the questionnaire. We are aware that perhaps some of the questions are difficult to answer or do not fully apply to your situation. If this should be the case please let us know by writing in the margin. 251

Do not sign your name Do not sign your name THE UNIVERSITY OF MICHIGAN SCHOOL OF SOCIAL WORK ANN ARBOR, MICHIGAN INTERSTATE STUDY OF COUNTY WELFARE ORGANIZATIONS (DIRECTOR'S QUESTIONNAIRE) To begin, we should like to ask a few questions about your district office. The counties randomly selected within your district are and Do you have separate data for these counties? /7 Yes. If so, please give /7 No. If your county data is not these statistics on the attached separate, but rather, collected as supplements for each county a whole in your diserict, then where applicable. Then continue please state the totals for your with question 7 on page 3 in the entire district in the questionnaire. questionnaire. We are aware that perhaps some of the questions are difficult to answer or do not fully apply to your situation. If this should be the case please let us know by writing in the margin. Do no your Do not sign your name t THE UNIVERSITY OF MICHIGAN SCHOOL OF SOCIAL WORK ANN ARBOR, MICHIGAN INTERSTATE STUDY OF COUNTY WELFARE ORGANIZATIONS (DIRECTOR'S QUESTIONNAIRE) To begin, we should like to ask a few questions about your regional office. The counties randomly selected within your region are ____and ___ Do you have separate data for these counties? / Yes. If so, please give /7 No. If your county data is not these statistics on the separate, but rather, collected as attached supplements for each a whole in your region, then please county where applicable. Then state the totals for your entire continue with question 7 non in the questionnaire. page 3 in the questionnaire. We are aware that perhaps some of the questions are difficult to answer or do not fully apply to your situation. If this should be the case please let us know by writing in the margin. 232

Do not sign your name Do not sign your name THE UNIVERSITY OF MICHIGAN SCHOOL OF SOCIAL WORK ANN ARBOR, MICHIGAN INTERSTATE STUDY OF COUNTY WELFARE ORGANIZATIONS (DIRECTOR'S QUESTIONNAIRE) To begin, we should like to ask a few questions about your district office. The county randomly selected within your district is. Do you have separate data for this county? / Yes. If so, please give L/ No. If the county data these statistics to questions is not separate, but rather, 1 through 6 in the questionnaire collected as a whole in your where applicable. district, then please state the totals for your entire district in the questionnaire. We are aware that perhaps some of the questions are difficult to answer or do not fully apply to your situation. If this should be the case please let us know by writing in the margin. Do not sign your name Do not sign your name THE UNIVERSITY OF MICHIGAN SCHOOL OF SOCIAL WORK ANN ARBOR, MICHIGAN INTERSTATE STUDY OF COUNTY WELFARE ORGANIZATIONS (DIRECTOR'S QUESTIONNAIRE) To begin, we should like to ask a few questions about your regional office. The county randomly selected within your region is _____. Do you have separate data for this county? /-7 Yes. If so, please give /a No. If the county data these statistics to questions is not separate, but rather, 1 through 6 in the question- collected as a whole in your naire where applicable, region, then please state the totals for your entire region in the questionnaire. We are aware that perhaps some of the questions are difficult to answer or do not fully apply to your situation. If this should be the case please let us know by writing in the margin. 255

Do not sign your name Do not sign your name THE UNIVERSITY OF MICHIGAN SCHOOL OF SOCIAL WORK ANN ARBOR, MICHIGAN INTERSTATE STUDY OF COUNTY WELFARE ORGANIZATIONS (DIRECTOR'S QUESTIONNAIRE) To begin, we should like to ask a few questions about your field office. The counties randomly selected within your jurisdiction are, and ____. Do you have separate data for these counties? /7 Yes. If so, please give /7 No. If your county data is not these statistics on the separate, but rather, collected as attached supplemerts for each a whole in your department, then county where applicable. Then please state the totals for your continue with question 7 on page entire area. 3 in the questionnaire. We are aware that perhaps some of the questions are difficult to answer or do not fully apply to your situation. If this should be the case please let us know by writing in the margin. When one or more counties in a district or region were sampled a supplement was included in the questionnaire for each county sampled. The supplement consists of the first six questions asked in the questionnaire. 254

2.C. RESPONDENT AND REQUEST RATES A. RESPONSE RATES-ADMINISTRATORS The following four tables report the response rate of the questionnaires as of January 28, 1970. Table 1 is for the four original states only. Table 2 is for only those questionnaires in the four original states which were included in the national sample. Table 3 is for the national sample excluding those from the four original states and those for New England. Table 4 is for the national sample including the four original states and excluding New England. Eleven more responses are needed in the national sample to reach a 607 return rate. This is encouraging in that the second follow-up letters were sent out only recently. The questionnaires have just been sent out to the New England states, so we expect to exceed the 60% return rate easily. Table I ORIGINAL FOUR STATES (Total Q's) STATE SENT OUT RESPONDED RETURN RATE Central 72 36 50o0% Western 31 23 74.2% West Central 87 69 79.37 Eastern * 134 83 61.9% TOTALS 324 211 65.1% * The response rates for public assistance and child welfare agencies are very similar. These figures combine the two types of agencies. 235

Table 2 ORIGINAL FOUR STATES (Q's in National Sample) STATE SENT OUT RESPONDED RETURN RATE Central 4 1 25.0% Western 13 11 84.67% West Central 7 5 71.4% Eastern * 36 24 6607% TOTALS 60 41 68037% Table 3 NATIONAL SAMPLE (Excluding Four Original States) (Excluding New England) SENT OUT RESPONDED RETURN RATE 307 168 54077% Table 4 NATIONAL SAMPLE (Including Four Original States) (Excluding New England) SENT OUT RESPONDED RETURN RATE 367 209 56.9 * The response rates for public assistance and child welfare agencies are very similar. These figures combine the two types of agencies. 236

B. REPORT REQUESTS FROM RESPONDENTS The enclosed four tables indicate the request rates for the findings of the project as of January 7, 1970. These rates are an indication that respondents feel they have for the information the project may provide. Of the 651%7 who responded in the Four Original States, 74.0% requested a copy of the findings. In the National Sample 73.87 of the 53.4% who have thus far responded, requested a copy of the findings. These high rates of request suggest to us that public welfare directors are aware of the need for information concerning the improvement of welfare services. In view of the present uncertainty re local welfare structure, and the criticism being experienced from many groups, it is probable that local administrators are seeking information about how colleagues have dealt with problems, about their orientations, and about their operational patterns. Such information could be useful to them in working with local boards, committees, staff, and pressure groups as well as with the general public. One further observation on the enclosed figures is the similarity between the response rates for the Four Original States and the National Sample - a 027% difference. In asking ourselves why these two rates are similar a few initial thoughts have been raised: 1o The similarity is due to the good methodology in sampling. 2. The four Original States are all northern states plus one of the reasons for their selection was that they had relatively well developed information systems. High response and general interest rates were expected to be high. Therefore, the request rate for the National Sample - which includes all the states irregardless of the factors mentioned above - was expected to be lower than that of the Four Original States. 3. The National Sample which does include the southern states and others which are generally thought of as having more problems in the area of public welfare may have been the equalizing factor in the rateso That is, these county directors may feel the greater need and therefore request the findings more frequently, 4. The request rate would be expected to be similar for all those that responded, because the fact that they responded makes them similar. 5, Report "collecting" is an executive function so it would be expected that welfare directors would request reports at a high and even rate across the states. 237

Table I Four Original States (Total Sample) STATE RESPONDED FINDINGS REQUESTED REQUEST RATE (%) Eastern1, 2,3 (CW) 42 30 71.4 (PA) 40 25 62.5 sub totals 83* 55 68.8 Central3 36 29 81.0 West Central4'5 69 52 75.4 Western 23 20 87.0 TOTALS 211 156 74.0 * One Q could not be differentiated between CW or PA. 1. One R asked for two copies. 2. Five R's repeated requests for findings. 3. Two Q's, not included in table, did not respond lut askad for findings. 4. Two R's repeated requests for findings. 5. One Q, not included in table, did not respond but asked for findings. 238

Table II Four OrigLnal States (Nationally Sampled Only) STATE RESPONDED FINDINGS REQUESTED REQUEST RATE (%) Eastern (CW) 13 6 46.2 (PA) 11 8 73.0 sub totals 24 14 68.3 Central 1 1 100.0 West Central 5 2 40.0 Western 11 11 100.0 TOTALS 41 28 68.3 Table III National Sample (Excluding Four Original States) (Excluding New England) RESPONDED FINDINGS REQUESTED REQUEST RATE (%) 164 122 74.1 Table IV National Sample (Including Four Original States) (Excluding New England) RESPONDED FINDINGS REQUESTED REQUEST RATE (%) 205 149 72.7 239

2oD. PROCEDURES FOR HANDLING QUESTIONNAIRE DATA IN RELATION TO COUNTY ECOLOGICAL DATA Two types of analysis can be applied to the Provide Hope Questionnaire: (1) organizational and (2) ecological. The first can be approached by treating the organization as the unit of analysis with each questionnaire received representing one and only one welfare organization. The ecological analysis, which involves relating the responses in the questionnaire to demographicecological data on counties, requires adjustment in the treatment of the questionnaire data for two reasons. The first is that since some welfare organizations represent more than one county and, the second is that some counties contain more than one welfare organization. A. PROCEDURE FOR HANDLING QUESTIONNAIRES THAT REPRESENT MORE THAN ONE COUNTY In all cases, attitudinal and background data were duplicated, triplicated, etc., as the case may be, for each additional county represented by the questionnaireo Each data card was uniquely identified by the county I.Do number. However, in order to distinguish between the original card and the duplicates, the following code was employed in Deck 01, col, 01: 1 = original 2 = duplicate 3 = triplicate etCo The purpose of adding this information was to abstract the duplicates for the purpose of organizational analysiso In some cases, such organiizational data as number of applications or size of staff, represented the entire welfare district composed of more than one county~ In these cases, these data were duplicated as the attitudi-nal and background data wereo In other cases, the respondent may have actually distifnguished the number of applications and size of staff for the counrty or counties in the study as distinguished from the total for the entire district, In these cases, while the attitudinal and background data were duplicated, the organizational data were noto The county coded as "original"s was to be the largest of the co.runties, in terms of caseload and size of staffO Where the organizational data was for the district and not the county, an adjustment was necessary to make the data comparable. The number of counties represented by the organizational data was as follows: 01 = only one county represented in the totals (true for one-county agencies and multi-county agencies in which data for each county is available) 240

02 = two counties represented in organizational data 03 = three counties represented in organizational data etc. For the purposes of ecological analysis, adjustment in organizational data were adjusted by dividing the data on client applications, size of staff, etc., by the values. In the national sample, where only one county was selected from a district, it was coded "1" in Deck 01, col. 1; the total number of counties in the district in which it was located (if organizational data represented the district) was coded in Deck 10, cols. 69-70. If Organizational data represented the county sampled only, it was coded "01" in Deck 10, cols. 69-70. In the national sample, where more than one county was selected from a district, original and duplicates were coded in Deck 01, col. 1. If the data on clients, staff, etc., were for each county separately, they were coded "01" in all cases in Deck 10, cols. 69-70. If the data were for the entire district, the number of counties in the district were coded in Deck 10, cols. 69-70. The same procedures are followed for the study of the four states. Since every county is "sampled," duplicates were added and other adjustments made accordingly. B. PROCEDURE FOR HANDLING QUESTIONNAIRES THAT REPRESENT LESS THAN ONE COUNTY There are relatively few problems of this type. Some counties are divided into smaller districts that are sufficiently autonomous to rate a separate interview. This is true where a metropolitan area like King County (Seattle), Washington, was divided into four parts (as of mid-1967), and where some New York State counties were divided into two welfare departments, one for the central city and one for the remainder of the county. Col. 12 in the I.D. number was used to differentiate between the departmental units. Thus, county "017" was coded as "117," "217," "317," and "417" for the four departments representing the county. "017" was coded for an additional card designed to represent the county as a whole in the ecological analysis. One response was selected to represent the county; the most complete questionnaire. However, should one respondent seem to reply for the county in the organizational data, he was assumed to represent the county. While the attitudinal data and background data were merely reproduced for the representative card, the organizational data were duplicated only if it was clear that the respondent's answers were for the whole county. If each respondent in the county responded for his section only, the individual re241

sponses were added for the representative card to obtain organizational data for the entire county. The composite or representative card was coded "0" in Deck 01, col. 1. Since the data for the entire county were entered as a coding decision for the organizational variables, code "01, was to be used in Deck 10, cols. 69-70. For the purpose of ecological analysis, the composite representative card will be employed in the analysis rather than the original responses. This means that "00" will be coded in Deck 10, cols. 69-70 for the original respondents enabling these responses to be filtered out of the analysis. For the organizational analysis, all the original responses for the departments within the county will be coded as "1" in Deck 01, col. 1 in order to include them all, while the composite representative card will be coded as "0" in that column, in order to exclude it from the organizational analysis. C. SUMMARY For the organizational analysis, include all data coded "1" in Deck 01, col. 1, excluding all "O's," "2's," "3's," etc. For the ecological analysis, exclude all data coded "00" in Deck 10, cols. 69-70 and adjust organizational measures by dividing by value in Deck 10, cols. 69-70. 242

2.E. GENERAL CODING INSTRUCTIONS Code all numbers directly unless otherwise specified. In codes where the field width is larger than the actual response given, the leading O's (zero's) must be coded, thus if reply to Q2 says 715, the correct code is 00715 and not 715, "O's" are used as missing data codes wherever possible. For interval scales where "O" is a valid code, "9's" are used for missing data codes, e.go, for number of children, an individual may have no children, thus "99" is missing data code. Interval scales where "0" is illogical or impossible, use "O's" for missing data codes, e.g., for age, no respondent can be "0" years old, thus "00" is missing data code. When a question is left blank, there are two possibilities. The first is that the respondent is unwilling or unable to answer the question and will be coded as missing data. The second is that the respondent intends a zero response and should be coded accordingly. The latter can be detected when a blank response is part of a larger set of responses which are not left blank, e.g., the number of welfare workers is filled in, but the number of indigenous workers is left blank. In cases where there is a response of 98 +, and there are only two columns available for coding, code 98 and not the red number. For example, 108 is given and there are only two columns, code 98. 243

DECK COLUMN QUESTION AND CODE 01 01 ORIGINAL & DUPLICATE COUNTY REPRESENTATION Code all questionnaires indicating whether they represent only one county or more than one county, i.e., each questionnaire which represents only one county will be coded (1) for the only original card. If a questionnaire represents more than one county, it is distinquished by using the following code: 0 = Card for representative Dept. in Co. with more t;hfan I Dapt. 1 = original (as above)' 2 = duplicate 3 = triplicate; etc. 01 07-09 QUESTIONNAIRE I.D. This I.D. will be repeated for each of the ten decks (1-10), representing the number of the questionnaire itself. This identifying number is assigned upon the receipt of the questionnaire in the office. 01 10-11 STATE I.D. This I.D. will be repeated for each of the ten decks (1-10), representing the number identifying a particular state by the County and City Data Book 1962, i.e., 35 = Ohio; 48 = Washington. 01 12-14 COUNTY I.D. This I.D. will be repeated for each of the ten decks (1-10), representing the number identifying a particular county within a particular state by the County and City Data Book, 1962, i.e., "002" - Allen County in Ohio; "026" -= Pend Oreille County in Washington. EXCEPTIONS: then there is more than one Questionnaire sent to a 12 particular county. For example,'all Child!'elfare questionnaires will be identified with a 5, 6, or 7 in place of the 0, i.e., "502 = Allen County Chil!4eilfare-Department in Ohio. In the case of Texas where there are 200+ counties, child welfare counties are identified as: Co. Bd. of Child Asst. Wielf. "7001 = 501 101 = 601 201 = 701 12 In the case where a questionnaire represents less than one county (where it is divided into smaller districts) as in the case of King County (Seattle) Washington, being divided into 4 parts, a code is used to differentiate between these departmental units, i.e., county "017" is coded as "117", 244

DECK COLUMN QUESTION & CODE 12 "217", "317", & "417" for the four departments representing tFe county. "017' is coded for an additional card designed to represent the county as a whole in the ecological analysis. 12-14 When a questionnaire cannot be identified, code the first one as such in each state "901l" next, "902"; etc. For additional explanation, see Matthew Silberman's paper titled "Report on Procedures for Handling Questionnaire Data in Relation to County Level Ecological Data." 01 15-16 1. LENGTH OF AGENCY EXISTENCE 00 = No answer 1C 17-21 Q2. 1967 CLIENT APPLICATIONS 00000 = No answers + unknowns 01 22-26 Q3a. FIRST TI;lE APPLICATION 00000= no answer, or "He do not keep this information," "This information is not available locally" 01 27-31 Q3b. FOOD STAMP APPLICATION 00000= None or indicate no program, i.e., "We do not participate in the food stamps program" etcetera. 99999 = No answers (blank) 01 32-35 Q3c. 1967 APPLICATIONS UNABLE TO HELP 9999 = No answers (blank) 01 36-41 Q3d. PERCENTAGE OF DEGREES OF HELP 99 = No answers (blank) EXCEPT when 100% is reached without filling in each category, then code blanks 0. 98 = Code if response is 98+, e.g., 75% Completely (code 75) 25% Somewhat (code 25) % Little (code 00) 01 42-79 04. AGENCY PERSONNEL O's s When have answered other sections in far left column ("A," refer to following page), and the "Total" seems to correspond to the figures given & no other information is available in back of Q. 245

DICK COLUMN QUEST:ON AND CODE 52-54 999 = When "Total number of welfare workers" and/or "Tot.l and number indigenous workers" in "A" group are given & 58-60 the corresponding indented categories in section "B" are blank. 66-73 0 = No answers 1 = Partime or fulltime. 64-65 Put total of indented categories ("C") in "Other Professionals." Code "Other" in terms of existinq categories when this seems reasonable, e.g., bookkeeper & case aides - secretarial & clerical homemakers + indigenous workers (55-57 only) Likewise, coding professionals (or fractions thereof) from "Other" category if same titles as given in indention ("C"). 4. How many of your agency are in each of the following categories? __ecutive and administrative Supervision J, / /Research / // TTotal number of welfare workers ~/ / ~/ ~/ _ uFull-time equivalent welfare ~/ -~// / workers (except indigenous,//2i~ // ~workerPs)'//// _ Total number indigenous rkers // Full-time equivalent. A //^y~~~~~ ^^indigenous workers A _>~ __ _____ /Secretarial and clerical Other professionals Full-time equivalent lawyers Full-time equivalent psychologists Full-time equivalent psychiatrists C Full-time equivalent clergy/ Full-time equivalent nurses Other (specify: __________) Total staff 246

DECK COLUMN QUESTION AND CODE 02 15-35 999 = No answers, when 2 or more categories are left blank and it appears that the Director did not want to answer the question completely. 000 = No answers, when 1 category is left blank and it appears that the Director does not administer that particular category of assistance OR 000 = In counties where a certain program is not administered through this particular aqency, i.e., in states where child welfare & other public assistance programs are administered through different offices, in same county. In cases where it is shown that the workers carry a mixed load, see example below, take the total number given Pa divide by number of all programs except Aid to Blind. Code this number for each category, while coding Aid to Blind 998 (signifying to the computer it is a mixed load) 5. What is the case load for a full-time equivalent welfare worker for the following programs? AFDC 1 i Aged 18-20 O4LLUiAL / ____ Aid to the S&ind 21-23'l A______ APTVD | 24-26 Gene A Ad 6'tance 27-29 26 MQ.ediccue 6 Medicaid 30-32 40 Child Welfare33 (When dividing and there is a remainder of number of cases, as in this particular questionnaire, use this oriority in giving the higher numbers: 1) AFDC 4) General Assistance 2) Child Welfare 5) Aged 3) Medicare 6) APTD 02 36-41 Q6. STAFF TURNOVER 999 = No answer if both categories left blank. 000 = if only one category left blank 02 42-46 Q7. PROMOTION CRITERIA 0 = when blank 247

DECK COLUMN QUESTION A,!D CODE 02 50-56 Q8. S. U. ROLE PERCEPTION Complete ranking if only one rank Is missing, e.g., if only the "2" is missing, insert a "2" to complete the ranking. Otherwise code 0 for missing ranks. 03 15-58 Q9. ACTUAL-IDEAL DISCREPANCY 15-18 Col. 15 & 16 are used for "actual" col. 17 & 18 are used for "ideal". If answer reads 15% 5% the proper code is 1505. This applies to the entire go., i.e., to the entire deck 03. ETC. 99 no answer, Except code O's when responses add up to 100% or more in each column. (Code 9's only when obvious R did not answer for complete question) 98 98+ response 04 15-44 Q10. RESOURCE ALLOCATION 15-17 Q10. (1) FINANCIAL AID Col. 15 is used for "past" Col. 16 is used for "now" Col. 17 is used for "ideal" ETC. O's = no answer Ordinal ranks used for ties, e.g., ordinal ranks are used in the case of ties (when respondent ranked, but not as requested), e.g., if he ranked 1, 2, 3 in first three columns and used 4's in the remainder. You would add 4 and 9 (as that is the maximum rankinq number) getting 13; then divide by 2, getting an answer of 6.5. As 7 is the next larger number, code all the 4 responses 7. 04 45-47 Qll. COMPARISOMi OF AGENCY O's = no answers. 04 48-71 Q12. One of the ways used in research to help understand people's opinions and attitudes is to ask them to complete sentences by saying the first thing that comes into their minds. Thus, please finish the following sentences, putting down whatever comes to your mind as you read the opening phrase. CODES FOLLOW. Code the first thing mentioned in responses with double statements 248

12a. I CAN'T UNDERSTAND WHY SOCIAL WORKERS.., Primary Concern of the Director 1. The Profession: A. Workers manifest interest in the profession: (n.e.c.) (11) I. Despite lack of public support: e.g., "go into our field, which receives so little support from the public," "have so much dedication —considering community apathy," "underpaid," "feel they must work for lower salaries." (12) II. No reason given: e.g.,'continues with their vocation." B. Workers manifest unprofessional conduct and/or interests: (21) I. By leaving the field or ob: e.g., "leave,"' "quit working after several years," "do not return to the field after their families are raised." (22) II. Due to a lack of proper education and training: e.g., "lack a well rounded education with which to modify their inbred social concepts III. Due to indifference to the work (as a profession): (23) a. Manifest as a concern for money: e.g., "think of salaries first." (24) b. General: e.g., "indifferent to establishing social work as a profession," "do not stand up for their own rights (as professionals)," "fail to see the challenge of the field," "improve their image as a professional person." "seemingly lack motivation for their jobs." (27) c. Unconcerned about wages &/or afraid to ask for more: e.g., "are reluctant to ask for a living wage," "are so passive with respect to their own wages." IV. Dual standard of professional training and practice: (25) a. General: e.g.,'waste their educational training," "don't practice what they preach," "don't fulfill their capabilities.' (26) b. Mention dislike for public welfare (including not wanting to work in public welfare): e.g., "dislike public welfare," "look down on public welfare service." (28) V. No reason given (shows general lack of professio:~nlism or mentioning specific unprofessional conduct or interest): e.g., "are more interested in personal gain than service to the most needy," "try to emulate psychiatrists," "are more interested in obtaining professional status than concentrating on more effective ways to deal with clients." 2. The Workers: A. Hang-ups (fears, apprehensions, difficulties): (31) I. a. Ideas-oriented: e.g., "have a fear of new concepts." b. Problem-oriented: (32) i. hang-up over problem is personal: e.g., "become so agitated over small problems," "hesitant to make decisions and act independently." (33) ii. hang-up over problem is in relationship with client: e.g., "are afraid to frankly and honestly express themselves, and to carry this over in working with families needing help." "have difficulty seeing the problems of their client in depth." 249

I. Absence of hang-ups —-great ego-strength: (34) a. Ideas-oriented: e.g., "are so ideal in the help they wish to give." b. Problem-oriented: (35) i. Personally self-assured: e.g., "are such positive thinkers in the face of the many problems encountered and work loads." (36) ii. Excellent relationships with clients: B. Attitudes (feelings, orientations, wants): I. Concerns for "action orientation," urban work, social problems: (41) a. Supervisor for —worker against: e.g., "have been content to be hangers-on and not social action leaders," "are so reluctant to become involved —and are unwilling to express themselves on issues except around the coffee table," "don't speak out on social issues." (42) b. Supervisor against —worker for: e.g., "want to work in urban areas," II. a. Workers are bureaucrats: (43) i. Seek easy bureaucratic solutions (without "understandin clients) e.g., feel that giving financial assistanc.e solves all problems," "fail to extend themselves beyond a minimum level of effort," "devote their time exclusively to determine eligibility." (44) ii. Rule-bound, narrow-minded —-bureaucratic personality: e.g., "are so punitive and restrictive," "are so inflexible and dogmatic." (45) b. Workers are not bureaucrats (but they should be): e.g., "feel that adequate financial services is not one of the most important services you can give a client." (46)111. Trained & untrained workers do not get along: e.g., "with MS Degrees keep aloof from other workers without MS Degrees," "who are trained look down their noses at untrained workers." (47) IV. Other attitudes: i.e., "are troubled by the use of money," (51)C. General vindictive comments about workers: "are so inadequate," "talk so much," "ask such dumb questions," "are unable to see the trees through the forest." 3. The Work: (61)A. Problems workers have in relationships with clients (not due to hangups): e.g., "refuse to really listen to clients," "do not become more involved with client and caseload," "cannot communicate with clients." (71)B. Problems with non-client related aspects of the work: e.g., "get bogged down in paper work,' "are so prone to depend on supervisors to make decisions for them," "have difficulty meeting deadlines," "complain of too heavy workloads." 250

12b. IT'S A SHAME THAT MANY CLIENTS... Primary Concern of the Director 1. Client Deficiency: A. In relation to welfare agency: (11) I. Lack of knowledge (or information) of what services are available: e.g., "do not know how to use social services." "don't understand the services offered by the agency," "have a misconception of what services are available," "do not understand extent of services available and can't accept help offered," "do not understand our function, other than financial assistance," "don't know their rights." (12) II. Lack of confidence in agency (manifest in resistance to being helped and/or apprehension): e.g., "don't trust us' "are uncooperative," "do not realize we are there to help them." (13) III. Delay in seeking help or failure to seek help (because of stigma attached to welfare manifest as shame or pride for example): e.g., "are ashamed to ask for help with their problems," "art- too full of pride to seek our services," "do not feel free to come to the office before they are destitute." (14) IV. Failure to use services available or misuse services available (no reason given): e.g., "do not utilize all of the services available," "are unable to take advantage of the help that is available," "don't care for help." B. Deficiency is personal: (21) I. Lack of education, opportunity, source of deprivr'ion in the past (general, not specific to dealing with anything Lit perhaps life in general): e.g., "do not have a better education," "have not had a better chance." (22) II. Lack will or ability (to be better people, live better): e.g., "lack initiative," "are left unmotivated," "have large families," "are not more self motivated," "haven't the capacity to better themselves." (23) III. Lack of ability to cope with problems: e.g., "can't recognize their own problem," "have such difficulty in coping with their problems," "can't face their problems realistically," "cut their own throats," "are unable to give of themselves," "make the same mistake so many times," "are defensive." (24) IV. Have a poor self-concept: e.g., "down-grade their own self-worth." (25) V. Lack intelligence: e.g., "are mentally retarded," "are of such limited intelligence." 251

(26) VI. Are deprived (doesn't mention lack of will to change circumstances): e.g., "are poor," "neglected by relatives and friends," "unable to achieve a fuller life," (27) VII. Inability to communicate: e.g., "are unable to be more articulate in stating their needs." (28)VIII. Not involved in organizations to help them obtain their rights: e.g., "do not organize into effective groups," "aren't involved in parent-teacher organizations, advisory committees and other community groups." 2. Welfare Deficiency: A. Agencies provide inadequate assistance to clients: (31) I. General: e.g., "aren't helped more meaningfully." (32) II. Financial: e.g., "do not receive an adequate grant." (33) III. Non-financial: e.g., "do not receive necessary compassion, understanding and service," "are so seldom provided with realistic options." (34) IV. Lack of assistance is due to bureaucracy. impersonal relations, work demands (where specific services not mentioned): e.g., "cannot be properly served because of staff and time limitations." B. Organization of welfare: (41) I. Bureaucratic structure of welfare (assistance organization so organized as to require client to get aid from more than one agency and/or see more than one agency,whether or not agency able to provide assistance; refers to inter-organizational bureaucratization) e.g., "get bounced from one agency to another,"'"ust apply for, and be eligible for more than one public assistance program." (42) II. Selection of welfare personnel inadequate: e.g., "can't have better selection of board members." (51) C. Stigma attached to receiving welfare (welfare clients not considered like other people or not understood by others): e.g., "have to go through the degrading aspects of applying for help," "are made to feel as though they are'second-class citizens'," "have to be treated the way they are by society," "are thought of as inferior." 252

12c. I CAN'T UNDERSTAND WHY OUR COUNTY AGENCY... Primary Concern of the Director 1. Limitations Exist on Scope of Operations of Agency: A. Due to externally imposed restrictions: (11) I. General lack of autonomy: e.g., "shouldn't be given broad guidelines, objectives, and allowed to develop its own program," "cannot operate with more independent decision-making authority," "isn't permitted greater flexibility by the State office," "is losing more and more control." (12) II. Rules, directions and need to do paperwork (to satisfy controlling agencies implied): e.g., "is plagued by so many state office bulletins, memos, directives," "is so strapped with rules and regulations," "can't have less paper work,' "does as good a job as we do, what with state & federal bureaucrats throwing more and more paper work at us," B. Operations of agency should be expanded: (21) I. Do better job; serve clients, (general improvement desired, nothing specific mentioned): e.g., "can't do better," "can't be more imaginative and flexible," "is not able to sense the needs of our clients," "doesn't stand up and use its expertise," "doesn't have more answers-to have to resort to trial and error," "can't get more involved in more projects," "could not be a full service agency, delivering as one unit the fragmented services now delivered by several agencies." (22) II. Specific improvements of services mentioned; (excluding staff additions): e.g., "cannot do more in the area of mental health," "cannot get better housing," "does not do more...-senior citizen housing - medical care facility," "can't be a family agency rather than a financial agency." 2. Deficiencies Exist in Operation of Agency for Various Reasons: A. Lack of resources: (31) I. Lack of funds for effective operation: e.g., "doesn't have more funds allocated to us rather than new poverty programs as we know where the problems are," "seems to be the'poor sister' in the county set-up," "can't have some money available for tangible needs such as buying sheets, blankets, etc.," "cannot qualify for federal funds for demonstration." (32) II. Lack of funds for salaries: e.g., "is so damn tight with salaries, cannot meet comparable salaries for work in same fields." (33) III. Lack of resources (no elaberation): e.g., "is so cheap." 255

B. Staff problems: (41) I. Need more or better staff (general; no specific type of staff mentioned): e.g., "isn't given sufficient staff to cover public relation and community organization," "cannot be staffed more adequately," "as a rural area, is not attractive to graduate social workers..." "is so understaffed," "has such a high caseload (42) II. Need more staff (mentions specific type of staff needed): e.g., "isn't alloted a research position," "can't have a medical consultant," "cannot have our own legal department." C. Poor physical plant: (51) I. Not enough space (include general statements): e.g., "has such an inadequate physical plant," "is subjected to such poor working conditions," "can't get better space." (52) II. Poor use of space: e.g., "cannot have staff housed in one building with outreach offices in more remote sections of the county." (53) III. Not enough equipment: e.g., "has to wait forever before facilities...are provided." D. Organizational problems: (61) I. Unnecessary responsibilities delegated to agency: e.g., "is responsible for the county farm." (62) II. Inability to be flexible: e.g., "cannot change to meet managerial needs," "doesn't always stay abreast of trends & changing concepts in social welfare." (63) III. Instability of operations: e.g., "is always operating on a crisis basis," "has such an abundance of work changes," "has to depend on the whims of the county for proper housing," "must often move in divergent and often contradictory directions." (64) IV. Conflict within (include problems of communication): e.g., "has such a history of advisory committee —executive quarrels," "is run by unqualified farmers when the trained professional is the only one who really knows how to do this job right." (01)E. General organization not effective (no reason): e.g., "cannot be more effective in the communities as social innovators," "doesn't become a state agency,' "'needs to be merged w.ith the larger state agency." 3. Relationskips \i.t' Group Outside Agency Difficult A. Community relationships (includint maws media assume eneral coent refer -o cort unity) ~ (71) I. The public does not understand the agency' e.g., can't get to people of the community-wit! better understanding of the real reasons for welfare," "isn't better understood in the community," "is so confused with the agencies both local and federal." 254

(72) II. The public does not "accept" the agency (not liked, unsupported): e.g.,' is not as accepted as it should be locally," "doesn't have more support from the public," "does not receive recognition for the work it does," "is disliked," "remains a public image of something distasteful," "must constantly defend our function because of the attitude of the public." (73) III. Te public epects too much: e.g., "is expected to cure all ills," "gets all the blame for every govenmental ill in the county." (74) IV. Agency has public relation problems: e.g., "is not more aware of its public relations responsibility," "must have all details of salaries, etc. published locally when no other state or county office or agency does," "does not become more recognized in the community as a resource for information regarding welfare problems, "has so few inquiries about how we spend the tax payer's money." B. Relationship with clients: (81) I. Clients do not "accept" the agency: e.g., "is not better accepted by the client population." (82) II. Agency cannot "accept" clients (as clients, include negative statements about some or all clients): e.g., "finds it so difficult to accept its role as servant." "doesn't do something about the chislers." C. Relationship with other agencies and organizations: (91) I. Problems with the legal system: e.g., "is not fully understood by the commissioners and court." (92) II. Problems with other agencies: e.g., "is considered the place for other agencies to send their problems to." 255

12d. IT'S A SHAME THAT CUR FINANCIAL RESOURCES... The Primary Concern of the Director: 1. Agency Resources Viewed as Problematic: A. Limited financial resources: (11) I. General (mentions short supply of available funds without elaboration): e.g., "are not larger," "aren't unlimited," "are not more," "are not adequate." (12) II. Specific source of limitation mentioned: e.g., "are very limited by the county commissioners," "are limited and controlled by commission." III. Consequences on service of limited resources mentioned (doesn't necessarily mention "limit" on resources; simply mentions consequence): (15) a. General (no particular type of service mentioned): e.g., enable us to barely scratch the surface," " will not permit us to do all that needs to be done to meet human needs," "cannot be used to meet all of the client's needs." (16) b. Financial (mentions concern for better standard of living or need to give clients more financial aid): e.g., "are not more adequate to meet a health & decency standard," "do not permit for the granting of allowances in keeping with a decent minimal cost of living," "do not fully provide for family needs on a higher subsistence level." (13) c. Non-financial service: e.g., "are not good enough to provide good child welfare services," "limits in all services that we would want to provide," "aren't greater to allow for needed program expansion at this time," "do not permit upgrading of food and transportation allowance." (14) IV. Consequences on personnel and facilities of limited resources: e.g., "limit us in the hiring of adequate staff," "are not more adequate so we could add persons needed to work in certain problem areas," "are limited when it comes to salaries," "don't allow us to get new quarters." B. Inadequate distribution of resources: (21) I. General (no mention of specific misuse of funds, a distribution problem) e.g., "aren't utilized more effectively," "are not more flexible," "are used more selectively in aiding clients," "are limited to certain needs, I think we should allow for other needs," "are not geared to the needs of the people," "are limited by law so you cannot use the fund to best advantage." (22) II. Mentions "specific" misuse of financial resources: e.g., "are spent for unnecessary administrative cost," "continue to be used for money grants when there are so many other needs.' 256

2. Source of resources viewed as problematic (and all prior categories don't apply; no mention of limitation): A. Original source of revenue: (31) I. Taxpayer and public viewed as problematic (taxes as source may be mentioned): e.g., "have to be drawn from the property tax base," "must be considered from the point of view of how much this is costing the taxpayer instead of what the need in the community is, "are granted more the idea of appeasing the taxpayer rather than meeting the needs of the client." (32) II. Politicians make decisions (with taxpayer or selves in mind): e.g., "depend on politics," "are tied in the political actions." B. Immediate source of revenue: (41) I. Federal government control (include complaints about not enough local or state funds) seen from point of view of the John Birch Society: e.g., "must come from the federal government," "have so many federal-state strings attached," (42) II. Local & state government control (include complaints about not enough Federal funds) seen from point of view of Federalists: e.g., "cannot be less locally denied," "are all controlled by the commissioners," "have always been questioned by county commissioners," "are not shifted more to state & federal." C. Specific source of difficulty not mentioned (mention problem of external controls): (51) I. Bureaucratic demands: e.g., "must have so much red tape and strings attached," "must be expended in such bureaucratic detail," "are questioned & scrutinized continually," "are tied with categories." (52) II. Inconsistent environment: e.g., "vary from year to year and program to program," "are not more uniform and flexible." 3. Societal Resources Viewed as Problematic: A. Inadequate distribution of societal resources: (61) I. Resources should be used to prevent problems from arising in advance: e.g., "are not better spent in prevention of social ills rather than stop gap, often punitive measures later." (62) II. Misuse of resources (distribution should be different without mention of functions such as prevention of problems): e.g., "are so misused in our society," "are spent on war, cosmetics, cars, sports stadium, etc." (71)B. Limitation on total resources: e.g., "can't eliminate poverty." (81)4. Resources are adequate; no Problem: e.g., "no problem whatsoever," "on the contrary our financial resources are adequate." 257

Matt Silberman 9-11-69 12e. I WONDER WHY THE GOVERNMENT... Primary Concern of the Director: 1. The government imposes too many restrictions: A. Bureaucratic Problems: (11) I. Pperwork is too much: e.g., "requires such excessive amounts of paperwork." (12)11. Too many rules and regulations: e.g., "imposes so many rules and regulations," "does not cut down complicating red tape and paperwork," "requires so much documentation before we can assist a person," "has to make such complicated procedures for such a simple task." (21) B. Less federal control desirable (the state & local government should have more control): e.g., "doesn't give more money to the states to carry on their own programs," "doesn't let the states operate more freely," "doesn't insist that states pay more of the Welfare Bill." "does not allow more local control —under supervision," "is working towards control," "is allowing itself to become a'parent' to all its citizens." (31) C. Less local control desirable: e.g., "wants so much control locally." 2. There is not enough communication between the government and others in establishing policy: (41) A. Lack of communication with professionals competent in social welfare: e.g., "doesn't consult with'first line' caseworkers before they set mandates & policies that affect local D. of S." "doesn't council more with grass roots public assistance administrators." "at higher levels is so uninformed of actual practice at lower levels." B. Lack of communication with the pubic & community in establishing policy(42) 1. Government should listen to the public: e.g., is so slow in reacting to public opinions, continues to retard community.organization in public welfare programs." (43) B2. The public should listen to the government: e.g.,'cannot interpret its programs so that the public has a better understanding." 3. Wrong policies: A. Resources are allocated improperly (includes money spent in the wrong places, and independent of liberal or conservative policies): (51) I. The government is ignorant of the needs of the deprived: e.g., "is blind to the many special needs of the poor, aside from money," "keeps trying to prove that enough money can solve peoples problems," "feels that negative income tax will restore the self-esteem to the poor." 258

(52) II. The deprived are not a primary concern not enough concern for the deprived in general or a particular category of deprived persons): e.g., "doesn't realize priorities and do more for the poor of this nation," "cannot allow more to client than they do in moon ventures and other services for people," "is so liberal with farm subsidies and not welfare funds?" "provides so much for everyone but the normal child who happens to be neglected." (53) III. Resources should be allocated for prevention of the causes of poverty rather than the treating of symptoms: e.g., "doesn't think in very long terms with regard to prevention of problems rather than wanting to cut costs and think in terms of two, three or four years," "doesn't place more emphasis on prevention of the cause of Welfare." B. Policy is too liberal: (61) I. The government should spend less: e.g., "has become so liberal and how soon a Republican administration will make the necessary changes." (62) II. The taxpayer is not of primary concern: e.g., "is not more concerned about the rights of the taxpayers in the bill of welfare." (63)111. The government is involved in too many programs (include too many categories of assistace: e.g., "must be involved in all programs including farming." "is considered the best source of help for clients." "doesn't return its functions to the citizenry." (64) IV. There should be less concern with national uniformity in welfare programs: e.g., "attempts to make all welfare programs uniform throughout the country." "feels the closer to Federal regulation one gets the better the service." C. Policy is not liberal enough: (71) I. The government should establish a standard or minimum grant.(include negative income tax and the ability of people on welfare keeping part of the money they earn seen as desirable): e.g., "does not set a minimum grant." "does not establish a guaranteed annual fncome plan." "is so slow in allowing clients to retain a reasonably substantial portion of any earned income.' (72) II. The government should do more for the poor (includes spending ore and no mention of misplaced priorities or misspendings): e.g., "doesn't help more with allocations." "appropriates less money for child welfare than the legal limit." "has trouble appropriating adequate funds for welfare." (73) III. The government is too punitive toward the deprived: e.g., "feel people with problems should be punished." "has to be so liberal in punishing militant groups that cause damage." 259

(81)D. What the government does is wrong (no specific policy mentioned): e.g., "does many of the things it does." "goes on as they have." "is so illogical." "can't see the problems of today." "isn't more creative." 4. There is something wrong with the organization and implementation of policy: A. Too many agencies: (91) I. Different agencies established to solve same problems (specific mention of agency): e.g., "chose to conduct its'War on Poverty' via new agencies instead of handling new funds through existing structures and agencies." (92) II. Different agencies established to solve different problems which should be united: e.g., "doesn't unite child welfare and public assistance. B. Doesn't do things right in order to enable the realization of objectives: (93) I. Not enough involvement in what happens in agency nor providing it with the capacity to do its job: e.g., "is so slow in activating approved projects." "doesn't assume more leadership in planning and implementing a more effective service delivery system." "never made our agency a necessity in the county until recently," " at the state level doesn't provide us with more help in in-service training for new caseworkers." "doesn't give us better support for the many problems we try to solve." (94) II. The criteria for determining who gets help are incorrect, inadequate, or too comlex (too many categories of assistance): e.g., "can't set one criteria, that of need —and allow counties to administer one program." "tries to define poverty in terms of a dollar and cents value when in reality many people below the so called poverty line have no complaints." "retains a multiplicity of public assistance categories." "duplicates and makes complicated so many programs." "continues to initiate new programs rather than strengthening existing ones." (95)111. Doesn't know how to go about things or poor social control mechanisms: e.g., "doesn't start where the problem is." "can't seem to police itself," "cannot better control misuse of funds intended for the poor." (96) IV. Lack of uniformity in policy implementation: e.g., "can't maintain a direction in welfare programs." "keeps changing policy statements or procedure detail." (97) V. "General" (system inadequate to get job done or government is poorly organized): e.g., "hasn't developed a better system for the delivery of services and meeting material needs." "clings to an outmoded welfare system." "is so loaded with incompetent employees." (98) V. There is something wrong with the "manner" of implementation of olicy (includes attitudinal way of doing things): e.g., has to wheel and deal." "does not become more offensive regarding welfare program rather than always being on the defensive." 260

12f. IT'S STRANGE HOW PEOPLE... Primary Concern of the Director: 1. Clients are strange: (11) A. Clients feel that society owes them: e.g., "feel they have so much owing them." "feel the government owes them a living." (21) B. Clients are on welfare for so long: e.g., "become dependent on the social service offices after being on the rolls for some time." "continue to have the same problems generation after generation." (31) C. Clients resent help: e.g., "react to proferred help by social agencies." "resent the help you give them." (41) D. Clients expect so much from the agency: e.g., "feel the agency can solve all problems in one visit." (51) E. Clients are dumb (don't understand): e.g., "get themselves into such situations." "fail to understand the simplest directions." 2. The public is strange: A. Attitudes towards clients are strange: (61) I. People are punitive toward the needy: e.g., "feel people with problems should be punished." "indict the poor and fail to understand the many factors that made them that way." "think of punishing parents rather than helping children." (62) II. People believe welfare recipients are unworthy or inferior people: e.g., "have the opinion that all welfare recipients are lazy or immoral." "look down on public assistance clients." (63) III. People dislike welfare recipients and/or don't understand them (no mention of reason such as thought of as unworthy-: e.g., "feel about public assistance recipients." "hate public welfare recipients." "negatively react to their neighbors receipt of public assistance." "think either poor people or deprived persons are either all bad or all good." "who have the necessities of life criticize those who are less fortunate and feel that people on welfare'really have made it'." "often react to welfare recipients." B. Attitudes toward welfare (and welfare agencies) are strange: (71) I. The public rejects welfare and the needs of clients: e.g., denounce the welfare office." "resent'welfare'." "care more about their animals than human beings." "react to welfare —a give away program." "still object to public assistance." "react to the giving of grants to the needy." "ignore need." "resent a welfare agency." 261

(72) II. The public criticizes welfare: e.g., "are so quick to point out our faults and so slow to recognize a job well done." "are so free to criticize when they know so little about welfare." (73) III. The public doesn't understand welfare (or lack of knowledge: e.g., "don't understand welfare programs." "fear public welfare programs."'n"isunderstand current welfare programs and the goals of current welfare programs." "know so little about the program they support." "continue to refer to welfare as a way of life," "interpret public welfare." C. Attitudes in general are strange: (81) I. The public is narrow-minded: e.g., "stereotype situations. "quickly see prejudices in others but fail to recognize the same in themselves." "criticize most the things they are least familiar with." (82) II. The public has dual standards (willing to spend money in some areas but not on welfare): e.g., "feel about helping people when it costs money." "want services for those they know —but not for all those others." "are so often paradoxical in their attitudes and behavior." "attack welfare until their family needs help." "expect service but do not support payment at the same time." (83) III. People are concerned with themselves and not others: e.g., "are more and more preoccupied with money, materialism and the race for it." "are so self-concerned and yet have to be social to survive." "are willing to accept the status quo and unwilling to see the other fellow's vie.Tpoint." "do not familiarize themselves with social ill.until they become personally involved with a problem." "can't see the poor." (84) IV. People are strange: i. General miscellaneous negative statements about _pople: e.g., "react." "seem to think that people are strange." "are chronically unsatisfied," "get so involved." (85) ii. Miscellaneous negative statements with other specific attitudes: e.g., "feel about any government project." "feel the solution to the problem of poverty is so complicated and so in need of constant study." (91) D. People are good (include all positive statements about people i. general): e.g., "are becoming more accepting of welfare," "will come forth if we would only let them know what the problems are." 262

DECK COLUMN QUESTIONI AND CODE 05 15-44 Q13. PUNISHIENT STRUCTURE 15-19 Q13a. ABSENT A LOT Col. 15 is used for "discharge him" Col. 16 is used for "suspend him" ETC. 0 = no answer. 05 45-74 Q14. REWARD STRUCTURE ETC. Same as Q13. 06 15-18 Q15. DEPARTMENT SUPERVISION O = no answer 06 19-21 Q16. ADDITIONAL ASPECTS OF DEPARTMENT O = no answer 05 22-33 017. DEPARTMENT VS. DIRECTOR 22-23 Q17a. PROVIDE FINANCIAL ASSISTANCE TO MAXI?1UM ELIGIBLE Col. 22 is used for Department objective Col. 23 is used for Personal Preference ETC. 0 = no answer 05 34-37 Q18. WORKER'S LOYALTY O = no answer 06 38-50 Q19. COMPLAINT INDEX 0 = no answer 49-50 Total number of "l's" (true responses), e.g., 04. 07 15-72 Q20-48. LIKERT ITEMS 0 = If wording of statement has been changed in order to make it "more logical" or "reasonable," or no answer. 263

DECK COLUMN QUESTION1 AND CODE b0 15-54 Q49. AFDC MOTHERS 15-16 Q49 (1). AFDC MOTHERS ARE UNFORTUNATE Col. 15 is used for "people believe" (A) Col. 16 is used for "I believe" (B) ETC. 0 = no answer 09 15-62 Q50-65. DIRECTOR - S.W. 15-17 Q50. ASK OPINION OF WELFARE WORKERS Col. 15 is used for p (when you took office) Col. 16 is used for n (present time) Col. 17 is used for w (ideal) ETC. 0 = If wording of statement has been changed or no answer. ul 15-16 Q66. AGE 0 = no answer 10 17-25 067. EDUCATIONAL ATTAINMENT OF R 1 = Has respective degree 5 = Does not have respective degree 0 = When R leaves entire question blank 10 26-37 Q68a. FORMAL S. W. TRAINING 26-29 Q68a. UNDERGRADUATE HOURS Col. 26-27 are used for classroom hours Col. 28-29 are used for supervised F. W. Code absolute #, if + gg = gg If stated had an MSW, but did not record number of hours in graduate hours, code as U. of M.'s School of S.W.4016. ETC. 99 = nonresponse EXCEPT code 0 if respondent appears to have answeredcomipletely, code 99 for illogical responses, e.g., 120 undergraduate hours 98 = 98+ response 264

DECK! COLUMN QUESTION AMD CODE iO 38 Q6Sb. NUMBER OF CONFERENCES 0-4 = 0 55-64 = 6 5-14 = 1 65-74 = 7 15-24 = 2 75+ &"too many to say" = 8 25-34 = 3 No response = 9 35-44 = 4 Answered others, but not particular one = 0 45-54 = 5 10 39-43 Q68b. OTHER TYPES OF ACTIVITIES 1 = 1-3 extension courses 2 = 4-6 extension courses 3 = 7-9 extension courses 4 = 10-12 extension courses 5 = 13-15 extension courses 6 = 16-18 extension courses 7 = 19-21 extension courses 8 = 22+ & "too many to say" 9 = no answer at all for all parts 0 = answered others, but not particular one. 10 44-45 069. OFFICIAL GRADE See state manual 00 = no code available 10 46-49 Q70. LENGTH OF PRESENT POSITION 00 = no answer 10 50-53 Q71. LENGTH IN PUBLIC WELFARE 00 = no answer 10 54-55 Q72. FIRST JOB (DUNCAN) See socio-economic index (Duncan Score) 0 = no answer 10 56-57 Q73. FATHER'S JOB (DUNCAN) Code same as cols. 54-55 10 58 Q74. ANNUAL SALARY 0 = no answer 10 59 Q75. RACE 0 = no answer 265

DECK COLUMN QUESTION AND CODE 10 60 Q76. R'S REL. PREFERENCE Code major Prof. denom. (e.g., Lutheran) as 1 even if respondent checked "other" 0 = no response 10 61 Q77. SEX 0 = no answer 10 62 Q78. MARITIAL STATUS 0 = no answer 10 63-64 Q79. NUMBER OF CHILDREN 99 = no answer to this question nor other personal questions 00 = if responding to other personal data & left blank, it is assumed there are no children 10 65 Q30. Would you like to see your children become welfare workers? Yes. -- thy? No. -- Why not? 0 = no answer 10 66 Q80a/b. WHY -- HHY NOT W.W. These are positive replies; negative codes are the corresponding opposites: 1) Working conditions, e.g., "security,""variety of work," "work readily available" 2) Personal satisfaction, e.g., "I enjoy work and hope they may also," "wonderful challenge and full exiting future." 3) Altruism, e.g., "To help the underprivileged." "My children ought to care about others." 4) Desirable life associated with job, e.g., "A good life," "It doesn't pay enough." 5) Child's own preference, e.g., "If it interests them," "provided they like it." 6) Professional service & satisfactione.g., "There are more rewarding professions," "it is a satisfying productive occupation." 7) No preference, e.g., "I have no feelings in the matter." 8) Societal need and reaction, e.g., "We need them," "hopefully society can be developed to a point where welfare workers won't be needed." O) no answer 266

DECK COLUMN QUESTION AND CODE 10 67 Q81. R'S POL. PREFERENCE If respondent made check marks in both boxes, code only LEFT box. O no answer 10 68 Q82 R'S PRESIDENTIAL CHOICE O no answer 10 69-70 NUMBER COUNTIES REPRESENTED BY WORKLOAD AND SIZE VARIABLES 00 = R's welfare department for part of a county 01 = R's welfare department for one county or more than one county, but data is for only one county 02 = R's welfare department is for two counties & data is for both etc. 267

APPENDIX 3 269

5.A. Critical Dimensions of Community Structure A Reexamination of the Hadden-Borgatta Findings JOHN E. TROPMAN THERE HAVE BEEN many attempts over the years to develop a set of critical variables for analyzing community structure.Most often such efforts were deductive in orientation, and began with a set of assumptions about what variables should be the most important. Naturally, geographical considerations received much attention, and communities were often differentiated in terms of their location. Another set of assumptions saw communities in terms of economic structure and function, often differentiating communities in terms of the kind of industry that predominated, or the complex of industries that characterized the community. Finally, some students have been primarily interested in the quality of community life, e.g., the "moral integration" of American communities. An alternative, inductive approach was used by Hadden and Borgatta in their recent volume, American Cities: Their Social Characteristics AUTHOR'S NOTE: This analysis was part of the author's doctoral dissertation. Portions of this research were performed under a National Institute of Mental Health Traineeship and a National Science Foundation Grant (GS-1407). I should like to thank Professor Henry J. Meyer of The University of Michigan for a critical reading of an earlier version of this paper. 271

URBAN AFFAIRS QUARTERLY (1965). They took a number of measures of community structure and empirically investigated which variables were most useful to describe and differentiate communities. This approach has been less popular than the deductive approach, in part because until recently there has been a lack of computer facilities which could feasibly process the large batches of data required for such analysis. Perhaps more important, such an approach was thought to be empiricist and atheoretic, and hence of dubious utility. Availability of adequate computational facilities does not still the charge of empiricism. One review of the Hadden-Borgatta volume (Winsborough, 1966: 432) commented that "scholars busy pursuing their favorite set of theoretical ideas are unlikely to find time to devise ways to use an atheoretical scheme." However, the fact that Hadden and Borgatta used no specific theory in developing their list of critical community variables does not mean that one cannot proceed to theorize on the basis of their results, and attempt to relate these results to other work on community structure. In order to facilitiate such use, the authors provide a considerable amount of data in the book itself. Jonassen (1967: 432) felt that this would in fact be its main contribution. The main contribution of this volume is the basic reference material which may be useful for further analysis and synthetic evaluation. Particularly useful are the many tables of intercorrelations of variables. The purpose of this paper is to perform just such a synthetic analysis, using variables which Hadden and Borgatta argue are crucial for describing community structure as a point of departure. We wish to consider these variables through additional analyses which permit their use in a consolidated form more meaningful for theory building.1 RECAPITULATION For readers unacquainted with the American Cities volume, a brief review is in order. Hadden and Borgatta performed a factor analysis of 65 variables measuring various dimensions of contemporary American community structure (largely those variables reported in The City County Data Book-1962). Using the results of the factor analysis as a basis, the authors develop a set of 12 variables which they argue are crucial for the analysis of urban structures: 272

DECEMBER, 1969 (01) total population (25) percent single dwelling unit (02) density (12) median income (I) deprivation index (04) percent nonwhite (09) percent foreign born (08) median age (03) percent population increase 1950-60 (18) percent same house, 1955-60 (19) percent migrants (II) education center The deprivation index (I) and educational center (II) are "indices."2 For each 644 American cities, decilized scores are constructed according to the city's position on each of the variables, and these scores for each city are available for analysis. In addition, the authors present matrices correlating all 65 variables with each other for the following categories of cities: (a) all cities (b) large cities (c) medium-sized cities (d) small cities (e) towns (f) central cities (g) suburban cities (h) independent cities THE STRUCTURE OF AMERICAN COMMUNITIES As the variables are presented by Hadden and Borgatta, they are simply a list, albeit a list derived by an explicit and orderly procedure. The variables would have more utility and gain more acceptance if it were possible to relate them to some theoretical ideas about community structure developed in the community literature. A perusal of current literature on communities suggests a minimum of four variables are receiving attention-size of community, socio273

URBAN AFFAIRS QUARTERLY economic composition of the community, the racial composition of the community, and its status on a continuum of maturity/growth. These four may be used to consolidate the Hadden-Borgatta set of twelve. Such a consolidation would provide a parsimonious set of variables which could be used as a beginning step in community analysis, and in no way precludes the use of other variables for specific purposes. Size of Community. Size is always referred to as an important variable in community analysis. Large communities have issues and problems different from those in small communities, as a comparison between Sayre and Kaufman's Governing New York City (1960) and Vidich and Bensman's Small Town in Mass Society (1962) clearly shows. For community systems, total population (01) is conventionally used, and can be used here. Socioeconomic Composition. The class composition of the community is a variable of great importance. Banfield and Wilson in City Politics (1965) point to a critical difference between lower- and middleclass political ethos, and use this distinction as a fundamental point of departure in their political analysis. Presthus in his comparative analysis of Edgewood and Riverdale in Men at the Top (1964) notes that a difference in class composition of the two communities is of fundamental importance in accounting for different political and social styles. Hadden and Borgatta's variable of median income (12) is a good measure here, although it only indirectly captures differences in style of life which are implied here. Racial Composition. Angell's classic study of community moral integration (1957) used racial composition as part of an index of heterogeneity. More recently it received attention in an article by Wilensky and Lebeaux (1958). Hence, it should be included even if recent disturbances in American communities had not occurred. However, the riots and urban discontent are metropolitan problems very related to the nonwhite population, and they make this variable even more important. Maturity/Growth. Communities develop over time, and part of the difference in their character and social organization may be attributed to their position on a maturity/growth continuum. On the one hand, some cities are young and new, and are characterized by increasing populations. On the other hand, some cities are old and are characterized by 274

DECEMBER, 1969 population loss. This seems to capture the popular distinction between urban and suburban communities. The maturity/growth dimension is an implicit one in Agger, Goldrich, and Swanson's comparison of the political styles of four communities in The Rulers and the Ruled (1965). However, the relationship of a community's position on a maturity/ growth continuum with other variables has not been systematically explored to the same extent as size or class composition. No single variable used by Hadden and Borgatta is a direct measure of the maturity/growth notion, but a combination of some of their variables may serve, and this combination reduces the total number of critical variables necessary for initial analysis. Mature urban communities may be characterized by older residents (median age, 08), a greater proportion of whom are foreign born (09), who live in apartments and other more densely populated areas (density, 02), and who have a greater residential stability (percent same house 1955-60, 18). On the other hand, suburban communities might be characterized by a greater proportion of migrants (19), a population which grew between 1950 and 1960 (03), and a greater proportion of single-dwelling units (25).3 In sum, there is a convergence between some of the concepts developed in recent literature and the empirical indicators developed by Hadden and Borgatta. Of the four characteristics of community structure taken from the literature on community, Hadden and Borgatta have singe operational measures for three: total population, 1960, is taken as a measure of size; median income is appropriate as a measure of socioeconomic composition; and percentage nonwhite is a measure of racial composition. The concept of maturity/growth utilizes a combination of seven separate variables. Maturity of a community can be operationalized by considering its proportion of foreign born, median age, and density. Growth and developmental characterizations of a community, on the other hand, can be operationalized by noting the presence of migrants, population increases, and single-dwelling units. We shall now reexamine the data presented by Hadden and Borgatta in various ways to show that the reordering and consolidation of variables has empirical support and allows meaningful interpretations. METHODOLOGY The Design of Analysis The design of analysis involves a reexamination of the correlation matrices. Since the critical variables are supposed to represent inde275

URBAN AFFAIRS QUARTERLY pendently important measures of community structure, the null hypothesis would predict the absence of any important patterning of intercorrelation among the variables. The testing mechanism, if it can be called that, is simply to look for a form or pattern of interrelationship in the matrix. An early discussion of this approach was given by Louis Guttman (1954) in his paper, "A New Approach to Factor Analysis-The Radex." More recently, James C. Lingoes of The University of Michigan Computing Center commented on this form of analysis: "The point to be made here is that there is much more to a correlation matrix than its correlations; that is, the form or pattern existing among them, an aspect largely ignored by the factor analytic technique" (1964: 2). Lingoes calls this technique "molar correlation analysis," using the term of M. E. Jones (1960). According to Lingoes, "the molar analyst looks for some organizing principle which accounts not only for the pattern of intercorrelations, but would also organically relate the factors among themselves, by such concepts as complexity, assimilation, inclusion, etc." (1964: 3). A pattern, for purposes here, is developed using the notion of similarity. If variables are similar to one another, or are different measures of the same dimension, they will relate more strongly among themselves than to other variables, and they will behave similarly in their relationship to variables external to the matrix. The exact form of this analysis will become clear as we look at the first table. TABLE 1 THE INTERCORRELATION MATRIX OF COMMUNITY COMPOSITION VARIABLES, 644 CITIES, 1960a 12 04 09 08 02 18 19 03 25 (A) Size (01) total population -02 14 15 07 27 03 -10 -07 -20 (B) Socioeconomic Class (12) median income 48 27 17 08 36 04 31 08 (C) Racial Composition (04) % nonwhite -31 -15 -01 -07 -04 -15 02 (D) Maturity/Growth (09) % foreign-born 51 49 36 -28 -11 -64 (08) median age 28 50 46 -37 -44 (02) density 25 -27 -21 -56 (18) % same house, 1955-1960 -80 47 -30 (19) % migrants 47 32 (03) % population increase 1950-1960 43 (25) % single dwelling units SOURCE: J. Hadden and E. Borgatta, American Cities (Chicago: Rand McNally, 1965), pp. 112-184. a Educational center and deprivation index omitted. 276

DECEMBER, 1969 FINDINGS On the basis of the preceding discussion of the structure of American communities, four major variables are listed in their appropriate groupings-size, socioeconomic class, race, and maturity/growth. Four variables which define maturity and the three variables which define growth have a high intragroup correlation (medians:.42 and.43, respectively), and a high negative intergroup correlation (median: -.40). Further, each of the twelve correlations in the intercluster matrix is negative (from -.11 to -.80). The pattern analysis which was just discussed now becomes clear through this example. The null hypothesis must be rejected. Rather than a random pattern of interrelationship, these variables show a theoretically predictable and consistent pattern of interrelationship. Thus, we can provisionally conclude that rather than being seven independent variables, they measure opposite ends of a maturity/growth dimension. The maturity/growth dimension appears to be reasonably independent of the other variables. It has relatively low median correlations to racial composition (-.15) and socioeconomic class (+.17). The relationship to community size is patterned with size being positively correlated to maturity (median: +.10) and negatively correlated to growth (-.10). We shall comment on this pattern in a moment. At this point, the strength of the correlations is not really sufficient to consider that maturity/ growth is an independent dimension. Socioeconomic Class and Racial Composition These two variables, empirically, might well be considered as representative of a single dimension. In the matrix for "all cities" (Table 1) they have a correlation of -.48 which is even higher than the median level we accepted for the two ends of the maturity/growth continuum. In addition, they behave in opposite ways in most instances, further supporting the notion that they are at either end of a class dimension. Class relates positively to both maturity and growth while percentage nonwhite relates negatively to both. (As we shall see in Table 3, there is a high correlation between a community's decile score on the deprivation index and the decile score for percentage nonwhite, +.68.) Thus, on empirical grounds, we cannot support the notion of their independence.5 However, with full recognition of this convergence, we wish to leave them as separate variables, although other analysts might wish to combine them. It appears that the factor of race may take on a force 277

URBAN AFFAIRS QUARTERLY independent of class, and the current rhetoric of black power suggests this to be the case. While all may recognize that concepts of class in American communities have a heavy overlay of race, there seems to be some policy justification for keeping them as separate dimensions for some purposes. Neither of these variables is related particularly strongly to community size (percentage nonwhite: +.14; median income: -.02), so that as a single dimension, or as separate dimensions, they may be considered independent of size. Controlling for Community Size Size is often thought to be an overweening factor in community analysis, although the particular pattern of correlations observed in Table 1 does not suggest that it is as powerful as one might think. We see it as one dimension, as important as the others, but not of absolutely paramount importance. To satisfy an uncertain feeling on the part of some readers, we shall look at the correlation matrices within size groups. The patterns we have noted should not be fundamentally altered by such submatrices, a demonstration which is expecially important because of the patterned relationship of maturity to size which was previously noted. The data are displayed in Table 2. TABLE 2 THE INTERCORRELATION MATRIX OF COMMUNITY COMPOSITION VARIABLES, LARGE (N+79) AND SMALL (N+150) CITIES, 1960a Small Cities (50-75,000) N+150 01 12 04 09 08 02 18 19 03 25 (A) Size (01) total population -15 12 -05-03-10-01 -09-07-05 (B) Socioeconomic Class (12) median income 14 -59 27 08 07-08 10 42 08 (C) Racial Composition (04) % nonwhite 08 -49 -37-24-15-14 03-1611 (D) Maturity/Growth (09) % foreign born 35 27 -30 68 48 26 -20-13-70 (08) median age 19 03 -07 50 35 44 4241-54 (02) density 57 12 10 59 43 25 -31-21-53 (18) % same house, 1955-1960 16 04 04 30 39 49 -85-58-36 (19) % migrants -20 -01 -23 -26-30-52-85 53 36 (03) % population increase 1950-1960 -12 -02 -22 -19-3646-64 75 45 (25) % single dwelling units -31 -11 -01 -6947-7545 54 51 Large Cities (150,000+) N=79 SOURCE: J. Hadden and E. Borgatta, American Cites (Chicago: Rand McNally, 1965), pp. 112-184. a Education center and deprivation index omitted. 278

DECEMBER, 1969 First, the convergence pattern by which we established the maturity growth cluster is completely maintained in both large and small cities. However, looking at the within-size group correlations suggests that the relationship of race and class in a community may depend somewhat upon the size of the community, but is not defined by size. in particular, race continues to exhibit its negative relationship to the maturity/ growth dimension with three small exceptions. In large cities, it is positively related to density (+.10) while in small cities it is positively related to percent migrants and percent single dwelling units (+.03 and +.11, respectively). This suggests that while nonwhites are associated with ghettos in mature cities, they are also a component of growth in small cities.6 It is thus important to realize that increased racial proportions may have different effects in different communities. Socioeconomic class retains its negative correlation to percentage nonwhite in large and small cities (-.49 and -.59, respectively) but not its entirely positive association with the maturity/growth variables in small cities. However, socioeconomic class in large cities is positively correlated with maturity and negatively correlated with growth. As can be seen, the correlations tend to be small. These differences between race and class variables further argue for seeing them as separate dimensions. In sum, size exerts a moderate effect only on the relationship between class and race and their interactions within particular communities. Since we previously argued that they should be separate dimensions, these small adjustments require no adjustments in the dimensions. A Radicalization of the Data Set If we argue that the four-variable scheme is as useful for many purposes as the original twelve-variable Hadden-Borgatta scheme, we should be able to show one final point-that the variables do not alter their relationships in important ways under a severe radicalization of the data set. The development of ad hoc collections of communities is not uncommon in community analyses, and the reduced scheme should prove as useful as the full variable set under field conditions. Data are presented in Table 3 for 154 communities in the United States. They were chosen for research purposes extraneous to the present effort. The correlations presented are correlations of decile scores. Since the code scores for the 154 cities in question were punched from the data in the Hadden-Borgatta volume, we have an opportunity to observe the location of the two index variables, depriva279

0^ \O t^ O 0 SO O e, <0 OA t C,4 0 00 0'0. - > 9 > p O 0t: o o u ^'^ C^ c r —. ~. Cd, { b |. CD | O! cc r' o oo %OW eq0 00 e)', e eq 0 e* 2 ), v,O O_ 0 0V. 4.< 0^m 3 o S^ o ^ o ^ t t~~0: 0 0 0 -i,- <0 | —a 0I 0.-. =Q H U 0= A nn ^ os48 a..- a, E 0 2 jE~ ffi 3 ~ 6 280 ^o ^^U c, 51 o Cz^ Ur$* 280c~~

DECEMBER, 1969 tion index and educational center, in the scheme. Deprivation index is essentially a measure of lower classness and hence is located with median income, with which it has the strong correlation of -.70. Educational center seems to be related to community growth and is included with those variables which measure growth. The reader will note that, even with such a strenuous radicalization of the data set, the same pattern which we have been discussing is upheld. Given such a severe test of the correlation pattern, we can have relative confidence that it is a pattern of some stability. External Relationships One might well ask, at this point, the extent to which these four variables are useful in looking at other components of community structure. Relationships between the index variables and selected community characteristics are displayed in Table 4. Nine variables are selected for examination: (20) percent labor force unemployed (22) percent in white collar occupation (26) percent units built 1950 or later (27) percent units sound (33) percent units with two or more autos (36) age of city (38) per capita live births (39) per capita deaths (65) per capita governmental expenditures These variables were selected on an arbitrary basis. There is confirmation of the previous analysis in that the distinction between maturity and growth is retained. If the variables which define maturity relate positively to a variable, those which define growth relate negatively to it, with only one exception. The exception is variable 27, percentage units sound, to which both maturity and growth measures relate positively. The variables which define community maturity show a positive relationship to age of city (median: +.34), while the variables which define growth show a negative relationship to it (median: -.46). The concept of community maturity can now be extended to the age of the 281

utO ^'1' 0~ ~-1 cf, ~-., t 49 0I B * >- ^ ON q 1 o 0 et- oe ~ t o e b te - r-. s.t N f. 00, o (! - a o>. ^ -,t w 9 Z 0 Z o, od 0 Qg ~ t 9 ed ^ ^ ^ ^ ^ 9 0 T 0 cn9C< E <S Z OQ I2 ~'; N Z I IOH a C'S- D, |:o o Co..a o. e 2| S B E. 8 8:'~Y _ e t < o o 2o 82 rf ( N w 3 -. _5 _ c C, n O

DECEMBER, 1969 community itself. Race and socioeconomic status tend to maintain their antithetical posture. For example, socioeconomic class shows a negative relationship to both per capita birth and death rates, while race shows positive relationships to both. What is interesting is the relatively low relationship power of population size. Its median correlation to all variables (disregarding sign) is.09 while socioeconomic status has a median of.36. In fact, size has the lowest median in the array. While one would expect somewhat lower correlations between size and per capita measures, a.09 is quite low, given the degree to which size is discussed. It suggests that we should perhaps look more carefully at size in the future to be sure that it is not obscuring other relevant concepts. Because these variables just discussed were included in the original Hadden-Borgatta analysis, one might think that this array is just a repetition of the factor analysis. To show the integrity of the pattern, we can look at some dependent variables which were not included in the original analysis at all. These data are presented in Table 5. The reader's attention is called to the fact that the correlation data here are for 154 cities in the United States. Data are presented for the amount of money raised in private welfare campaigns, 1964 (Community Chest/United Fund); the per capita amount of county expenditures for public welfare, property tax and intergovernmental revenues; MOP ratio; and percent voting for the Democratic party in the 1960 presidential race. Again, the variables which define maturity and those which define growth generally behave in opposite ways. The communities characterized by growth apparently raised less money for the United Fund, both totally and per capita, than older, more mature communities. Also, they spend less per capita on county welfare, and have lower property taxes than older communities. The MOP ratio is clearly distinguished here. Mature communities have a greater concentration of managers, officials, and proprietors. This pattern of relationships suggests a reevaluation of Hawley's finding. It may be the dimension of maturity which is related to urban renewal. Growing communities are negatively related to the proportion of persons voting for the Democratic presidential candidate in 1960. (These results are doubly interesting because the decile scores were used, rather than the actual values of the variables and the N is 154 rather than 644.) In these data socioeconomic class and race do not behave in exactly opposite ways, as they tended to do in the other tables. Both variables are positively correlated to private welfare, income in aggregate, and per capita terms. Both variables are negatively correlated to per capita 283

0 — O " 0 0' 0 eq.oI eQ S S o I * a S I I ) $ OQ o H W o o O~ten~t ON Q-4 I <ocan 9~~~~~~~~~< ^* -l S o O 0\ e e e Q c 0,, O tn e W 9E 9 0 9 u-4C4 n S I *^n ~ 0 0e^ i e 00c0 e -D -a 6 % a | X |...,i 0 |c I~ o ~1 > O en O N _z _ 0 r. O C0 O C,,4 0 " 0 oN n n o s, o cr3o C OC30 t~, I.o -c-n. |0o 0 z g g; o ~,>=E t &. ^c0' 0_ ~ _ o OC f.SS ~, ON 24 m < O v0e 00 eOfNst 284~ S m S) aoE 8c 284~55

DECEMBER, 1969 county welfare expenditure. Both variables are moderately positively correlated to percent voting Democratic in the 1960 presidential election. In short, looking at a set of intercorrelations completely independent from those used in the original analysis we come to the same conclusions about the maturity/growth dimension, although we view a different pattern for socioeconomic status and racial composition. DISCUSSION AND IMPLICATIONS The correlation data presented generally support the notion of four variables. Throughout many transformations of the data set, the maturity/growth cluster retains its integrity. However, some questions may be raised about the clarity of the distinction between class and racial composition. In certain respects, as we noted previously, these variables act in ways which suggest a single dimension, with nonwhite being (on the average) at the bottom of the class hierarchy. However, such a combination would, at this point, be premature. While it is true that nonwhites may be concentrated at a particular position in the class hierarchy, using them to define that hierarchy introduces a conceptual confusion, rather than conceptual clarity. Further, recent events in Sources for Table 5 SOURCE: J. Tropman, "A Comparative Analysis of Welfare Councils." Ph.D. dissertation, University of Michigan, 1967. a The first 6 columns in this table are the intercorrelations of decile code score and various dependent variables. The last column, % voting democratic, is the intercorrelation between actual values, tabulated on a county basis, and a dependent variable. Since percentage same house, 1955-1960 is not available on a county basis, it was not included. As index variables, educational center and deprivation index are not available. b Directory, 1965 (New York: United Community Funds and Councils of America. 1965). c Local Government Finances and Employment in Relation to Population-1957, State and Local Government Special Studies 45 (Washington, D.C.: Government Printing Office, 1957). d This ratio was taken from a study by Mrs. Ann Hudson, a doctoral candidate at The University of Michigan, and was computed by her. It is the measure used by Professor Hawley in his study "Community Power and Urban Renewal Success" (A.J.S. 63, 4 [January 1964]). It is defined as the ratio of persons employed in the MOP (managers, officials and proprietors) group to all other employed persons. It is interpreted that the lower the ratio, the greater will be the concentration of power. (In other words, the fewer the MOP's, the smaller proportion of the total labor force they will be. So, the ratio is computed by taking: #MOP = MOP ratio. Total employed labor force 285

URBAN AFFAIRS QUART'ERLY American cities suggest that the nonwhite population may have a distinct influence on city life, apart from class dimension. Thus, for the moment at least, it seems appropriate to keep them distinct. Apart from this issue, the four-variable scheme should be useful to urban analysts to provide an overview of the dimensions of any particular community, or as a background measure in comparative community analysis. The maturity/growth cluster need not be measured by all seven variables; one can be chosen, depending upon the available data and individual preference. The development of the four-variable scheme from the HaddenBorgatta data in no way vitiates their effort. In fact, it confirms the utility of their work in several ways. For one thing, we were able to show that their empirical analysis could be meaningfully related to concepts current in the community literature. In this respect, their factor analysis could be interpreted as an hypothesis test of crucial variables developed by others. In addition, the fact that there is some constancy of variable performance throughout various transformations of the data set confirm their conclusion that the unit of tabulation is less critical to community analysis than has hitherto been thought (Hadden and Borgatta, 1965: 185ff). The importance of this conclusion should not be underestimated; it means that urban researchers need not, in all cases, limit their analysis to variables collected over the same unit. In a broader sense, it confirms notions we all have about sociological communities, as apart from cities, counties, and so on. In conclusion, we should like to suggest further lines of research which flow from this particular investigation. The first of such questions deals with the degree to which these community characteristics have in fact the linear relationships required by a correlation analysis. In the absence of other evidence, assumption of linearity is not an improper one; however, it should not go unchecked. The independence of these four variables might simply be relationships of a curvilinear nature. One could argue, for example, that size has a curvilinear relationship to class such that the higher class communities are in the middle range of size. It may be that racial composition has a curvilinear relationship to size such that the smaller and larger communities have relatively high proportions of nonwhites as compared to communities of middle range. Other hypotheses are of course possible. The point is to consider whether or not new assumptions would improve our understanding of the data. The possibility of curvilinearity relates to the second point-the convergence of characteristics. People in policy and research might like to know if these characteristics cluster in any way favorable to the 286

DECEMBER, 1969 development of community types.7 The existence of types in these data is unclear. There is a mild tendency for more mature communities to be larger, and have more nonwhites, but this represents only the barest of convergences. Nonetheless, there is no reason, a priori, to reject the possibility of community types. Finally, we might ask about the theoretical completeness of the operational measures of the sociological concepts discussed here. Consider size as a case in point. Often implicit in the variable of size of a community are such notions as the complexity of the community or the growth over time of a community. We should be exactly clear on what the effects of size are alleged to be, and if other independent variables are implicit, develop operational measures for them. Or again, the notion of socioeconomic class is measured by only one variable. We have not included the informational resources of communities, in the sense of coupling in an educational variable. Perhaps this needs to be done. In short, this is just a beginning. SUMMARY We have reexamined the Hadden-Borgatta indicators of community composition from the perspective of a definition of community which implies four, rather than twelve variables. Analysis of the correlation patterns of the twelve variables with each other supports the reduction to four-size, socioeconomic class, racial composition, and maturity/ growth. Indicators of community socioeconomic composition and community racial composition, considered separately here, perform in a way which might suggest their combination on an empirical basis. However, there seem to be sufficient theoretical and policy reasons to keep them as separate dimensions, at least for the moment. Further work in the analysis of community characteristics deals with exploration of possible curvilinearity of relationships and additional theoretical development concerning the concepts implicit in the particular operational variables and the development of community types. NOTES 1. Interestingly enough, the author discovered, after this paper had been completed, that Professor Wyatt Jones of the Florence Heller School, Brandeis University, was also engaged in a reexamination of these findings and had come to similar conclusions, although he used a different approach. 287

URBAN AFFAIRS QUARTERLY 2. The numbers in parentheses refer to the number assigned to the variable in the Hadden-Borgatta volume (1965: 73). The index variables have no number in the volume, and are here identified with Roman numerals. 3. The authors note this confluence themselves, but do not carry the point any further. They comment concerning percentage same house, 1955-1960, percentage migrants, and percentage population increase that "taken together, these three variables provide some basis for assessing the historical status of a city with regard to mobility and growth (1965: 72)." 4. This methodology is explored in more detail in John E. Tropman (1967: Ch. 2). 5. Wyatt Jones' "class" factor was defined by percentage nonwhite and median income. 6. Small cities had the second highest percentage growth, 1950-1960 (50%); in first place were suburbs (70.2%). Hadden and Borgatta (1965: 107). 7. This line of investigation is being developed by Wyatt Jones in his forthcoming paper. REFERENCES AGGER, R., et al. (1964) The Rulers and the Ruled. New York: John Wiley. ANGELL, R. C. (1957) "The moral integration of American cities," in P. K. Hatt and J. J. Reiss, Jr. (eds.) Cities and Society. Glencoe, Ill.: Free Press. BANFIELD, E. and J. Q. WILSON (1965) City Politics. Cambridge, Mass.: Harvard Univ. Press. GUTTMAN, L. (1954) "A new approach to factor analysis-theradex," in P. F. Lazarsfeld (ed.) Mathematical Thinking in the Social Sciences. Glencoe, Il.: Free Press. HADDEN, J. and BORGATTA, E. F. (1965) American Cities, Their Social Characteristics. Chicago: Rand McNally. JONASSEN, C. T. (1967) Review. Amer. J. of Sociology 72, 4: January. JONES M. E. (1960) Practice as a Process of Simplification. Pensacola: U. S. Naval School of Aviation Medicine. (mimeo) LINGOES, J.C. (1964) New Computer Developments in Pattern Analysis and Non-Metric Techniques. Ann Arbor: The Computing Center, University of Michigan. (mimeo) PRESTHUS, R. (1964) Men at the Top. New York: Oxford Univ. Press. SAYRE, W. and H. KAUFMAN (1960) Governing New York City. New York: Russell Sage Foundation. TROPMAN, J. E. (1967) "A comparative analysis of welfare councils." Ph.D. dissertation. Ann Arbor: University of Michigan. VIDICK, A. and J. BENSMAN (1962) Small Town in Mass Society. Garden City, N.Y.: Doubleday. WILENSKY, H. and C. LEBEAUX (1958) "Class and race in a changing city," in Industrial Society and Social Welfare. New York: Russell Sage Foundation. 288

3.B. THE AMERICAN STRATIFICATION SYSTEM John E. Tropman The University of Michigan Work on this paper was supported in part by a grant from the Horace H. Rackham School of Graduate Studies, The University of Michigan, a grant, 425-CL, from the Social and Rehabilitation Service Administration, U. S. Department of Health, Education, and Welfare, and a contract from the Office of Juvenile Delinquency and Youth Development, JD 68-15. 289

INTRODUCTION Stratification has been thought of primarily as a measure at the individual level, of particular persons. As such, it has been a variable of immense power, so much so in fact that almost no person or organization collecting data on individuals would omit at least some measure referring to the position of a person on some dimension of stratification, What has been largely ignored in the literature on systems other than individual, however, is the fact that groups, organizations, and communities, in a fashion similar to individuals, have a position in a stratification hierarchyo Hence, we can speak of group, organization, or community rank on some dimension of the stratification system. Interestingly enough, the assumption that groups and organizations at least, and to some extent communities have such a position in their own stratification system is easy to find in common parlance. One hears frequent reference to prestigious groups or organizations, and some communities are considered "prestigious" as well. Organizations have social rank as a part of their organizational character. Consider the commonly used concept of prestige universities or law firms, as opposed to organizations which are sometimes called "poor" or in the extreme case, "fly-by-night." In short, the position of a group, organization, or community on the stratification hierarchy is a key factor in the relationship of other units to it, of subunits within it, and of its presentation of self in the environment generally, Yet we are only beginning to secure information in the professional and disciplinary literature on this point. Part of the reason is because certain types of research-for example, the power structure studies-have not been thought of as within the preview of stratification. Hence the work of Spinard,* Mills,** and Fowlert is only recently being considered in this light. The work of Presthust in comparative community analysis clearly has community stratification as part of its explanation for inter-community differences, And Banfield and Wilson, in City Politics, make the distinction between middle class politics and lower class politics the key differentiating factor in their analysis. *William Spinard, "Power in Local Communities," in R. Bendix and S. Mo Lipset, Class, Status, and Power, 2nd edo (New York: The Free Press, 1966). **Co Wright Mills, "The Middle Classes in Middle Sized Cities," Ibid. tIo Fowler, "Local Industrial Structures, Economic Power and Community Welfare," Ibid. Re. Presthus, Men at the Top (New York: Oxford, 1964)o FE. Banfield and J. Q. Wilson, City Politics (Cambridge: Harvard University Press, 1962)o 290

Stratification, then, is clearly important. But, there have been conflicting theories about what, exactly, is the criteria to be stratified,* Money, prestige, power, and other items have all been advocated in the literature. Yet, though overlapping, the various measures of class, status, or power certainly do not correlate perfectly, and have come to take on something of a life of their own. Further, each of the measures offered by a particular author tends to be seen as the key and independent stratification variable, with all the others dependent and derivative. The evidence seems to us to point to a multifaceted system, with five dimensions —money, occupational position, status, power, and information. How these fit together becomes a point of critical importance, not only for stratification theorists, but for policy makers and social interveners. Let us consider each of the dimensions, ECONOMIC SYSTEM One of the main differentiating factors on a vertical scale has been money. Weber, for example, differentiated classes on the basis of the source of money. Today, income is commonly accepted as an indicator of class. The higher the income, the higher the class. Although there is not a well established convention of usage, the very word class itself tends to refer to a man's ranking in the economic system, and it is particularly unclear whether class refers to a man's position in the financial structure or his position in the occupational structure. These are, of course, related, but not identical, Warner, however, in his discussion, used the term class, but in fact did not use it in reference to economic variables, Having rejected those, he used the term to refer to a sort of reputational "status" variable. At the community level, it can be measured by Median Income. OCCUPATION It has recently become clear that a man's position in the occupational sector, which is variously called occupational prestige, occupational status, and so on, is a dimension of stratification which is closely related to class as defined above, but certainly not identical to it. This is the correct sense in which the Marxian notion of class should be used, Marx was really an occupational theorist; for him what was crucial was not how much a man made in terms *Discussion often revolves around the difference between the criteria with which stratification takes place, and the units which are stratified. This topic deserves extensive discussion by itself. For purposes here, the criteria are those items which are used to discriminate among the units (amounts of money, power, etc.) and the units are the items which occupy stratified positions (persons, organizations, etc.). 291

of dollars, but rather his position vis-a-vis the labor marketo The most recent work in this area is the Blau-Duncan volume, in The American Occupational Structure. The percent white collar could be an operational indicator here. STATUS Status might refer to prestige in and of itself. This is the kind of variable Warner was interested in his study of the American community, It is reflected in the work of E. D. Baltzell, in his study of The Protestant Establishment. Generally, style of life is important here with the style being evaluated and ranked in the system upon a "prestige" dimension. Those styles of life apparently most finely articulated and developed are the most prestigious. This variable is equivalent to the status variable mentioned in Weber's original discussion of "class, status, and power." Status of a community might be measured by percent housing that is sound. POWER The final variable in the Weber discussion, power, referred, as he originally developed it, to the political party. However, it has become clear that people might well be rated on a dimension of power apart from purely political office holding, although this may well be a central aspect of the dimension. Dahrendorf's discussion of "power class" is a case in point, Neibeur's discussion of moral man and immoral society focuses exactly on the position of the powerless individual as he attempts to deal with an "immoral" stateo Etzioni's book on the comparative analysis of formal organizations uses the power of the organization as its most salient dimension, These are but a few of the many works which focus on the power dimension. In the current rhetoric, there are notions of student power, Black power, flower power, and so on, all of which suggest, though colloquially to be sure, an increasing salience of the power dimension0 In fact, a number of community organization strategies such as the Alinsky strategy, as exemplified in his book, Reveille for Radicals, are explicitly based on a power framework0 Despite the currency with which the concept is used, there has been no attempt, with the exception of Hawley's proposal on the degree of concentration, of managers, officials, and proprietors, to provide a measure of power in a community. In part, this has been due to the varying conceptions of power itself, and the way in which measures of power were articulated with disciplinary interests, Therefore, a slight digression is appropriate in order to present the operational indicator of power used here, The key concepts are reviewed elsewhere, but suffice it to note that we can take Parson's comment that power is like money as a point of departure. This can lead one to think of power, not in terms of overcoming resistance, but as the successful pursuit of goals. Hence, we define power as the level of efficacy within the system, efficacy referring to the ease with which the unit 292

itself, or its constituent elements, secure their desires. From this perspective, the greater the resistance, the less efficacious is the pursuit of goals, and, hence, the less power there is in the system. One can thus define violence as the absence of power. Violence is presumably used when all else has failed as a procurement device. If this line of reasoning is correct, then one could look to the level of expenditures for violence management made in any system as an indicator of power, taking the reciprocal of that figure. Therefore, provisionally, we are using the reciprocal of police expenditures per capita as the indicator of the level of power in the system, Phrased another way, this measure simply suggests that, as the money spent for violence control in the system goes up, power decreases, the efficacy of and within the system is on the wane, and could theoretically culminate in the control of everyone by police. As yet, this thinking is in the beginning stages. However, the line of thought holds some promise, particularly in the notion that there can be power growth, just like economic growth. Further, there can be "power recessions" just like economic recessionso INFORMATION We need to include a dimension not in Weber's original thought. It has been recently developed by Slavistoga in his discussion of social differentiation, viz., the ranking on an informational variable. Perhaps it always has been the case. The mayors of the palace in medieval times were people who had control over a good amount of information. The secretary and other persons who are by virtue of their physical position in an office, centers for informational transfer, have developed legendary influence in many formal organizations. And in general, as society becomes increasingly complex, increasingly differentiated and diversified, the possession of specific kinds of informational bits becomes an increasingly valuable component. The role of the expert, the role of the consultantis increasing daily in its importance in the system. The best way to reflect the information variable is by years of education, We recognize that this is not the best measure because it does not pick up types of informational inputs that operate in the system apart from it. However, it does seem to be a relatively good, clear, and an unambiguous indicator, and at least we know more or less what it will tell us and what it won't, THE STRUCTURE AND INTERACTION OF STRATIFICATION VARIABLES It appears to us that there is little point in debating the merits of one dimension being predominant over the other, It appears quite clear(and there are many indications in the literature that the trend is moving in this direction) that the stratification system is, in fact, multidimensional, and that each of these dimensions, at least, is one of the important dimensions in that 293

system. So, from the start, we will simply say that money, occupational prestige, status, power, and information are important "goodies" in the system, and that people seek each of them, and people have differing amounts of each of them. This poses at once then the question originally posed by Lenskio What is the degree of rank crystallization in the system? That is, to what extent do people occupying one position in one rank system occupy a similar position in a second rank system? Let us assume for purposes of discussion that each of the systems is divided into deciles. Hence, a person can have none of a characteristic, and be coded zero, or can be in the top 10 percent and be coded 9. Hence, we have five rank systems, each of which runs from zero to nine. The question posed when one asks about rank crystallization, then, is the degree to which a person who occupies, for example, a five in the power dimension is also likely to occupy a five on the occupational prestige dimension, on the status dimension, on the economic dimension, and on the informational dimension. At this point, there is not much evidence available but a number of studies in the literature suggest that the intercorrelation among some rank variables is around ~4, In our research, we have had results of.36 which correspond with those reported in the literature, using ecological correlations rather than individual ones. Hence, it seems quite clear that an important component of the American stratification system is the separation of a variety of dimensions. and the ability of a person not simply to be "upwardly mobile," but to be upwardly mobile along a varied set of tracks. A CONICAL STRUCTURE The foregoing remarks pull together and articulate the trends which were evident in the literature in a variety of places, What we would like to do now is take the next step, in a sense, and suggest that the American stratification structure is shaped like a cone, narrow and compact at the bottom, wide and flat at the top. The geometry of the cone suggests a somewhat different conceptualization of the rank crystallization problem because it implies that the correlation values may not be the same at the top as at the bottom, and indeed, we are rather convinced that this is the case, For example, it seems much more likely that a person who is in the first decile with the lowest 10 percent of education is much more likely to be low on income, prestige, and power than a person who has a 9 in any dimension is to be high on all the others, In general (and it is recognized that it is overstated), we propose that the correlations will be inverse to the rank. In other words, if the correlations were run decile by decile, in the first decile, those in the zero group, the correlation among low status physicians would be.9, It would be inversely such that the correlation among those in the top decile in any dimension would be between 0 and olt Hence, we're suggesting that the value of,4, which has come up in the literature, is something of an average between a very low degree of intercorrelation at the top of the cone and a very high degree of intercorrelation at the bottom of the coneo We hope to test this hypothesis out in future researcho 294

But it has a number of implications. For one thing, it is quite clear that one variable can be substituted for another operationally for a lower class person, but for a middle and upper class person that may be increasingly less so; and it may also be clear as Lenski suggested that the degree of rank crystallization is itself an important variable. Under this hypothesis, rank crystallization is high only in the lower class, and it is this latter implication which we feel has very serious ramifications for intervention programs. -.- - ~~.. Hypothetical Decile Dimensions __________________________Mean Economic Occupation Prestige Power Information Rxy 9 9 0 9 9 0 8 8 8 8 8 1 7 7 7 7 7 2 6 6 6 6 6 3 5 5 5 5 5 4 4 4 4 4 4 5 55555 6 3 3 3 3 3 6 2 2 2 2 2 7 1 1 1 1 1 8 O O O O 0 9 SOCIAL CLASS AND SOCIAL INTERVENTION This analysis of stratification in the American system has property of interest to observers of social intervention programs. Every one of them, class, occupational prestige, education, and style of life, has been advocated and used in its programmatic implications for the adjustment of social problems. Thus, the approach to the solution of social problems is one of developing power. We have used homemaker programs to develop style of life more in line with middle class values. Educational programs, of course, have been a main force variable in the system, and a constant point of frustration and attack. Certainly, the financial variable is one that has historically been, to us, the provision of money. The occupational prestige variable has been a root to providing money and greater prestige through job trainingo In short, the reader will instantly note that one way of looking at social problems, one way of defining many of the main social problems that we currently face today, is to simply say that they are (result from) stratification deprivation. In other words, the way in which people talk about social problems viewed from the perspective of this analysis is simply that a social problem is found among that group of people that is acutely deprived on any dimension, Social interventionists have differed in the degree qf salience which they have accorded each dimension, and their intervention programs have differed accordingly. 295

In sum, we have suggested that there are five different dimensions along which units are stratified. Since each unit is involved in each of the dimensions, the degree to which he (or it) occupies the same position in the rank system of each of the dimensions becomes a critical point of interest. In particular, we have hypothesized that low positions on rank hierarchies are more likely to be correlated than middle or higher positions, making the American stratification system conical in structure. This hypothesis will have to be tested out in empirical research, but seems to be theoretically justified. 296

35C. SOCIAL STRATIFICATION, SOCIAL INTERVENTION, AND COMMUNITY COMPETENCE John E. Tropman Roger M. Lind The University of Michigan This is a revised version of Chapter X in Delinquency Planning and Community Competence, unclassified report, Office of Juvenile Delinquency and Youth Development, Department of Health, Education, and Welfare, 1969o It was supported by Contract JD 68-150 297

INTRODUCTION As the decade of the 1960's closes, social planners can look back on the development of the first attempts by the Federal Government to develop local planning and action talent directly in municipalities. The decade began with the President's Committee on Delinquency and Youth Crime, which gave way to the Poverty Program, which in turn appears to be yielding to the Model Cities program. These latter programs are still in existence, and evaluations are beginning only now to come in. The purpose of this paper is not to provide an "evaluation" of these programs. That has been done elsewhere.* Our interest here is to explore in some detail one of the assumptions which provided part of the rational for at least the delinquency program and the poverty program, viz., that municipal institutions were crumbling, and that cities were becoming unable to deal with the problems they faced. One of the solutions, therefore, was to help communities regain their executive integrity, or, as people in the delinquency program would put it, develop community competence. Basically, we wish to argue here that to understand the notion of institutional decay, the idea of developing community competence, and the problems which those intervention programs had which used these notions as points of departure, one must understand the system of social stratification in America, and the way that this system interacts with formal organizations, communities, and social problems. Throughout, we shall use the delinquency program as a point of reference. The assumption implicit in the delinquency program, and the poverty program, was that there was a fundamental flaw in the institutions serving the poor in the American city, and that institutional innovation and radicalization needed to occur before meaningful change on the community level could take place. This assumption rested not only on the empirical observation that institutions which served the lower classes were not performing as they should, but also on the theoretical notion that these same institutions were subject to bureaucratization and developed inflexibility as a matter of course. Gradually, subtly, means were being substituted for ends, staff was becoming increasingly parochial, procedures were becoming pathologically complicated, and clients were being badly treated. This inventory of bureaucratic decay was anticipated *Peter Marris and Martin Rein, The Dilemmas of Social Reform (New York: Atherton, 1967); Roger Mo Lind and John Eo Tropman,"Delinquency Planning and Community Competence," unclassified report, Office of Juvenile Delinquency and Youth Development, 1969; D. P. Moynihan, Maximum Feasible Misunderstanding (New York: Free Press, 1969). 298

by Max Weber and later discussed by Robert Merton in his classic essay.* The assumed validity of this notion has been a fundamental link in a series of points made in the past and present about the nature of the problem which social reformers face, and the nature of the conceptual issue which social scientists must address. It is for this reason that we must clarify as much as possible the issue of institutional decay, and the attendant problems, as well as the relationship of such decay to community competence. COMMUNITY INCOMPETENCE AND INSTITUTIONAL DECAY Community incompetence and institutional decay are not the same thing, although they are related, It is impossible for a community to be very competent when the institutions which are important structural members are rigid, conservative, time-bound, proud, and haughty, alienating rather than welcoming clientele. But such institutional decay is not the only aspect of community incompetenceo Correction of this condition may not be enough to solve the problem of community competence. Strong institutions do not necessarily make a strong city. In fact, strong institutions may weaken community competence because the institutional representatives think they can speak for the city and may in fact try to do so. Although strong institutions may not be an unmixed blessing from the viewpoint of developing community competence, the requirements of their own task performance call for a dynamic strength which minimizes the growth of institutional rigidity. SOCIAL CLASS, INSTITUTIONAL RIGIDIFICATION, AND COMMUNITY COMPETENCE It does seem to us, however, that there is an underlying variable which strongly affects both community competence and the integrity of institutions, one which generally has been ignored by people who are concerned with the development of social programs, despite its obviousnesso Hence, the following proposition: As the proportion of lower class people in a community increases, institutions will become rigid and communities will become increasingly incompetent. *Robert Merton, Social Theory and Social Structure (New York: The Free Press, of Glencoe, 1957), P. 195 ffo 299

Institutional rigidity increases in inverse proportion to the class position of those served by the organization: the most rigid institutions are the ones which deal with lower class people. The problem is compounded because the client group, due to its deprivation, is not able under normal circumstances to exert much influence over the organization. This is not true for their middle class counterparts. The increase in the proportion of lower class members of a community means that its resources are diminishing. Hence, it not only has a group of people who are less able to exert influence than their middle class counterparts, but it has a very rigid institutional structure, more so than a middle class community, and it has less resources than would exist in a middle class community. For this set of reasons, social change is extremely difficult. The direction in which this analysis will lead us should be clear at this point. It is not sufficient for us arbitrarily to attend to the problem of institutional rigidity, without understanding its causes and correlates, nor is it enough to attempt to solve this problem in order to open up routes toward mobility for poor people. For one thing, the institutions will resist very strongly, a fact which has been clearly documented in a number of studies of OJD activity. Secondly,, the lack of resources in a community means that the institutions will fight for every small increment of resources they can obtain. This makes cooperative, consensual types of arrangements extremely difficult to produce simply because, under conditions of scarce resources, competition becomes more severe and more wide-spread. Hence, it appears to us that the whole problem of community competence hinges in a sense upon the distribution of deprivation in the community. As the proportion of lower class persons increases, the community will simultaneously manifest (a) the kinds of problems that require community competence to solve and (b) the absence of that competence, It does seem that one common characteristic of both rigid institutions on the one hand, and low-competence communities on the other, is that they both deal with or are forced to process lower class individuals. This commonality explains the spurious association between community incompetence and institutional rigidity. SOCIAL CLASS AS THE CRITICAL VARIABLE In the social system called the United States of America, both currently and over time, lower class status has been something of a pariah position, It has been avoided by those who could do so, through a variety of mechanismso There has been a great emphasis on upward mobility and progress, and a very negative connotation attaches to "downward mobilityo?" (As we shall see there are a number of dimensions along which upward mobility might be achieved.) It appears the virtually every social interventionistic or reform-type program has 300

at its base an attempt to change the class position of the recipient. Even psychotherapy, rather than being an attempt to adjust people to their societal position, as some of its critics argue, might better be described as an attempt to make them sufficiently fit to return to the mobility contesto Indeed, a theme running throughout most of the dilemmas and discussions in reform movements, in social agencies, and in various helping programs, reflects concern about problems of class-related behavior. Examples of such behavior may include violence in the prison system, malperformance in the public school, inability to verbalize in a psychotherapeutic situation. In sum, the status contest, the mobility game, is a fundamental game in the United States played not only by individuals and families, but also by organizations and communities. THE ORGANIZATIONAL MOBILITY RACE Organizations are no less class-interested than persons. They seek to improve their own reputation by dealing with "a better class of people." Organizations which deal with lower class clients experience great difficulty dealing with them because of the middle class orientation of the organizations' employees, and often seek to avoid them0 This suggests a fundamental proposition. Organizations receive prestige from the people who use them0 For this reason, an organization will seek to increase the power, wealth, education, or status of its clients, customers, or people affiliated with it. A high prestige organization is one frequented by high prestige people, such as the high-powered law firm, the top-rated university, the top industrial firm, the high-status social agency. Conversely, those organizations which are generally regarded as inadequate in the system are those organizations which deal exclusively or primarily with lower class people, organizations such as the welfare department, the police department, the prison system, the public mental hospital, and so on, Norton Long expressed this thought regarding particular departments in business organizations: "The satiation of businesses with the fruits of successful political activity led to a withdrawal from overt and intense participation in the field of local politics, the activity of tending the local political fences became a specialized department in the corporate hierarchy and was no longer a main concern of top management. The position in the hierarchy of the man responsible for taking care of the corporationts local political problems became down-graded and with the intensified disesteem of politics and politicians, especially at the local level, his post suffered from the bad odor of its clientele0 The courts and a successful lobby with the state legislature permitted a growing neglect if not contempt for local politics. Suburbanization meant a physical unconcern for all but company related local services, and the politics of top managementVs suburban residence 301

represents scarcely anything more momentous than the politics of the country clubo Thus, the replacement of robber baron, resident owners, and leading families by a withdrawn managerial elite has left vacant positions in the local structure."* An organization will try to improve the class level of its clientele unless it is prohibited by law from so doing~ One example of such a prohibition is the rule in public housing which states that if tenants earn over a certain amount of money, they must move out of the dwelling. This limits the clientele of the Public Housing Authority to low-income groups. A similar rule obtains in public welfare, Hence, in a double sense, it is the "poor organization" that deals with poor people. The organization is poor because of its clientele, and it is poor because of few resources as well. One solution for this double bind is organizational mobility. A private social agency exemplifies such mobility very well0 Originally, social agencies developed to meet the needs of the poor. However, as the social work organizations professionalized, they left-and some say ran from-the poor.** Organizational mobility occurs for at least two reasons0 One is that clients of greater prestige and wealth permit the organization access to more resources in the system, either directly in terms of fees paid or contributions made or indirectly by reputation0 Secondly, because individuals working for the organization benefit if the organization moves up,they are interested in pushing the organization as a vehicle for their own individual mobility or an enhancement to it. The individual gets a second-order prestige from the organization, The organization which cannot move upward suffers from several things0 For one thing, it is often a "poor organization" in the explicit sense of lack of resources0 Every social welfare program suffers from this lacko The system simply will not produce adequate resources for poor people0 It is not so much a question whether the middle class will pay for charity for the poor; the middle class pays for "charity" for everybody via the tax system, military, farmers, and so on0 Unfortunately, the poor have no rubric under which they can hide, no legitimizing myth under which they can claim resources from the system0 And the organizations that deal with them reflect thiso Indeed, these organizations are frequently so poor and so understaffed that they cause worse problems than the ones they are designed to resolve0 *Norton Long, "Corporation Satellites,"t in Edward S. Mason (edo), The Corporation in Modern Society (Cambridge: The Harvard University Press, 1961), po 212. **Richard A. Cloward and Irwin Epstein, "Private Social Welfare's Disengagements from the Poor: The Case of Family Adjustment Agencies," in Mayer N. Zald (ed.), Social Welfare Institutions (New York: John Wiley and Sons, 1965). 302

PROTECTIVE MANEUVERS OF LOW STATUS ORGANIZATIONS Organizations which are constrained to deal with the poor and cannot be mobile develop defensive reflexes which enable them to appear to maintain their organizational integrity in the organizational status race. Generally, they try to develop middle class oriented staff and middle class oriented modes of approach. By this we mean that organizations which deal with the poor can be expected to put a strong emphasis on professionalism. This is the case in the two professions of social welfare and education, particularly in the areas of those professions which service the poor, namely public welfare and ghetto education. There is a strain between the need for resources and the development of professionalism. The recent unionization activities of these two professions may represent a reversal of the attempt to develop professional status while dealing with the poor; perhaps the feeling is one of having been so tainted by this association as to require decent pay for such service. In any case, as a defensive measure, organizations which deal exclusively with the poor tend to use a middle class mode of approach, compensating for their feeling of deprivation in terms of clientele. But this approach tends to separate the organization from clientele very sharply, and to place substantial social distance between the client and his organizational "mentoro" Naturally, this is an organizational goal, not a client goal. But there is very little the client can do to change it because of his lower class status. One of the characteristics of lower class status is the convergence of negatives on a whole series of dimensions. Indeed, if one assumes that there are several independent and independently meaningful dimensions of stratification in the American system, then the poor are deprived on all of them, It appears that there are at least five (and perhaps many more) substantive areas within which people can be ranked, if one considers the elements of power, education, money, occupational prestige, and social status. The question, then, hinges not only on the location of any person or organization on any single dimension, but the overall pattern of position. A lower class person is likely to be characterized by the absence of money, of education, of power, of occupational prestige, and in general, the absence of anything that is considered of value in this system. It is a pariah state. Such people are often not treated with respect by organizations, something which has long been a sore point with them. This particular concurrence of negatives puts them in the worst possible position for controlling an organizational bureaucracy. The difference is immediately clear if we again use the example of schools and social agencies. Suburban parents have little trouble controlling their schools. The principal knows what they expect, and he endeavors to provide it in order to avoid difficulty with them. The teachers are predominantly middle class and so already generally in sympathy with the goals and life styles of 305

suburbanites. Suburban schools are generally of high quality, are desired goods in this system, and add many thousands of dollars to land values in an area. The type of school district in which a family will relocate is an important consideration in the move of a middle class family. The middle class by a variety of means produces the kind of school system in which they are interested, There is no evidence that lower class people are less interested in the school system and its benefits for their children, but the quasi-professional functionaries who operate the schools have been unresponsive to their circumstances and desires. That is institutional rigidity. They simply do not have the status, the money, the power, the education, or any combination of these needed for effective control of the bureaucracy, Bureaucracies certainly need to be controlled and directed, and there is good evidence that to some degree this can and does happen. However, not every constituent group has the skill to do it, and it is here that the problem lies. It is from a stratification perspective that we must consider social intervention programs. SOCIAL STRATIFICATION AND SOCIAL INTERVENTION Social intervention programs, including those developed with the support of the OJD/YD-President's Committee and Ford Foundation, are essentially middle class missions to the poor, albeit considerably more sophisticated today than the evangelical loin cloths of the past. This increased sophistication does not mean, however, that the academics and intellectuals who formulate the programs have enhanced their understanding of the nature of the problem, nor does it increase the ultimate success of the program, Virtually every intervention program makes a crucial assumption which practically guarantees its failure, vizo that the recipients of the program suffer on only the dimensions addressed by the program. Advocates of the development of power appear to feel that either all the other dimensions are irrelevant, or that people can secure whatever they need if they have adequate power. A similar attitude can be observed on the part of those who advocate money paymento To paraphrase the "material wealth" solution to problems of poverty, "Well, the problem of poverty can be solved simply with money, and there is no point in fussing around with anything elseo" This view is limited despite its face validity. People who have money, whether a little or a lot, require intelligence and capability for money management. Without those skills, the amount of money provided may well be irrelevant, In short, most of the intervention programs about which we know were designed assuming that people had, for example, a moderate amount of education, power, occupational prestige, but were temporarily out of money, So money provided from the public fisc was to "tide them over" until they got "back on their 304

feet." In this sense, the dispensers of public funds, the developers of power, the providers of education, were operating under a limited and harmful assumption, If a person is deprived on several dimensions, helping him on a single dimension is simply insufficient. This, it seems to us, is the critical pointo The poor, then, suffer from what might be termed "stratification deprivationo" By this phrase we mean to indicate that they are very low on every and all dimensions of stratification, a fact which must be taken as the central point of departure for every intervention strategy. In addition, as we have previously suggested, the problem of increasing stratification deprivation in our central cities is the cause of both institutional rigidity and lack of community competence. Hence, it is quite clear that a person who wishes to design an intervention strategy must take into account these interrelationships. For example, a person who wishes to build neighborhood power through the formation of neighborhood organizations must start with the understanding that the people with whom he is working are not only low in power, but also in education and information. This suggests something which has been affirmed in a large number of local organizing efforts and in the OEO and JD project themselves: that lower class people do not use middle class organizations wello There is probably no way to make this point without sounding critical, which is not our intent. But it simply appears to be a fact of life that we cannot expect to build organizations as we do in the middle class, with people who are experienced in the use of organizational roles, in delegating authority to professionals, etco If an organization is to be designed, it must be designed with these facts in mind. The same is true with programs such as job training, Why, for example, should a person take a job that will provide him with a minimum of money, and little prestige? Here again, the educational factor comes into play. So does the power factor. Can people of low power really secure good opportunities? Probably not, at least not unaided. In short, what we wish to suggest is that any strategy which begins on a single dimension and does not consider the fact that the targets are simultaneously deprived on a number of dimensions will fail, It is an inaccurate design, and the treatment is not so much inadequate as it is misplaced. It is this misplacement which leads us to a note of caution. The particular nature and intertwined interdependence of problems which people,deprived on a number of dimensions,suffer has meant that the intervention programs designed to remedy problems on a single dimension might well have worked if only one problem existedo So we wish to caution those who are writing and thinking about social intervention not to dismiss too quickly a number of attempts which have been made in the past. These attempts may well be worthwhile, provided that they are taken on their own terms and are appropriately addressed to a problem set for them, The juvenile delinquency program is a case in pointo Our findings, as well as the findings of Marris and Rein* all indicate that the effect of these *Marris and Rein, op. cit. 305

programs has apparently been negligible. This may well be true, and yet this should not necessarily sound a death knell for programs of this type. Although a number of existing programs did not take the structure properly into account, and were doomed to failure from the start,they nevertheless may well be reasonable and efficacious programs in their own right, addressed to the appropriate target. The notion of the development of social power among a power-deprived group may be used as a case in point. In lower class groups where this approach has failed, it has been largely because people were so deprived on other dimensions that even if they succeeded in getting power, they would not be able to use it efficaciously, and in fact, other simultaneous deprivations prevented them from successfully amassing power in the first place. The strategy of power development, which we may refer to as the Alinsky strategy, has in contra-distinction been quite successfully used by studentsO Students are a group of people who are not deprived on a whole series of dimensions, but, in terms of their relationship to the university, can be said to suffer a severe power deprivation, They have set about to rectify this in recent years. Their ability to move rapidly into positions of influence on faculty committees and administration committees, to create a situation where they are now almost automatically consulted on a variety of issues (although not as much as they would like), suggests that where a group which has had some success in other areas in deprived on a power dimension, the recouping of the power in an excellent strategy. SUMMARY In this paper, we have attempted to assess the notions of institutional rigidity and community competence via analysis of the stratification system of the United States, We have suggested that there are five dimensions to that system at least: income; occupational prestige; status; education; and power. People can be differently ranked on all of the dimensions. We have proposed that it is far more likely that people at the lower end of any one dimension will be ranked at the low end of all other dimensions, than it is that people who are at the upper end of any dimension will be at the upper end of all other dimensionso This configuration of low statuses might be called poverty crystallization. This conceptual scheme serves, then, a defining function for social problems We tried to show how those particular configurations of events which are called social problems can really be seen as'stratification deprivation,"T Simultaneously, it became possible to view most intervention programs as attempts to reduce stratification deprivation by raising people on a particular dimension, We argued that the contemporary middle class concept of a program for the poor operates under the assumption that the poor exist on some middle level of dimensionality on each dimension except for one, where they have a low 306

place. Hence, only moderate help need be given because support from the other programs acts to pull a person back up to moderate status. We said that exactly the converse was true-that the poor are low on all dimensions and that help in one must not only suffice for that dimension, but must also overcome the dragging effects of low status on other dimensions, It appears clear to us that a program design needs to take into account these simultaneous deprivations, rather than attempting to operate on a single dimension. The multi-varied dimensionality of the stratification deprivation problem further suggests to us that we should not be overly hasty in dismissing programs which appear to have failed previously to solve what we have called generally "social problems." Indeed, it is quite logical that they would fail, particularly if they were addressed to a single dimension in a multi-variate situation. Hence, this analysis not only suggests delaying the rejection of current programs which appear to be problematic, but reexamining other programs which have already been dismissed, 307

UNIVERSITY OF MICHIGAN 3 9016 03695 6913