THE UNIVERSIT Y OF MICHIGAN SCHOOL OF EDUCATION Final Report A PREDICTIVE STUDY OF EMPLOYABILITY AMONG THE WITH. THE CALIFORNIA PSYCHOLOGICAL VISUALLY IMPAIRED INVENTORY Bernard A. Bast ORA Project 033690 supported by: U., S,. DEPARTMEJT OF LABOR MANPOWER ADMINISTRATION GRANT NO.. 91-24-70-14 YWASHINGTON,. -D. C. administered through: OFFICE OF RESEARCH ADMINISTRATION ANN ARBOR March 1971

IA-es j4x, \ ") J,J -) (I; k,.O ^

ACKNOWLEDGMENTS The formulation and execution of this study has depended upon the invaluable assistance of dedicated persons both at the University and in the community. I express deep appreciation to my thesis Committee Chairman Professor Geraldine T. Scholl, and Committee Members Professors Percy Bates, Anna S. Elonen, and M. Clemens Johnson. I am grateful to Professor Donald R. Brown for his assistance in the initial stages of the dissertation, but whose presence on the committee was terminated by a sabbatical leave abroad. I express warm gratitude to my supervisor and colleague at the Rehabilitation Institute, Detroit, Michigan, Dr. James V. Tattan, who has assisted me throughout my graduate career with unsparing contributions to my professional development. I am grateful to Mr. Lyle Thume, Director of Services for the Blind, Rehabilitation Institute, Detroit, Michigan, and to staff members of the State of Michigan, Office of Services for the Blind and the Metropolitan Society for the Blind, Detroit, Michigan, for their gracious efforts to locate subjects for the study. I am both grateful to and admiring of the subjects in the study who volunteered countless efforts to the conduct of the study. I express gratitude to Dr. Howard Rosen, his staff and review panel of the Manpower Task Force, United States Department of Labor, without whose financial assistance this study could not have been iii

undertaken. Finally, I thank my wife, Gretta, for her patience during my graduate work and for her assistance with the item analysis in this study. iv

TABLE OF CONTENTS Page LIST CF TABLES vii LIST OF FIGURES x ABSTRACT xi I. INTRODUCTION 1 II. RELATED LITERATURE 4 A. Psychological Implications of Blindness 4 B. Some Psychoanalytic Viewpoints on Blindness 6 C. Personality Testing of the Blind 8 1. Projective techniques 8 2. "Objective" tests 9 D. The California Psychological Inventory 11 1. General description 11 2. Theoretical aspects of the CPI 14 3. Research applications of the CPI 15 E. Factor Analyses and the CPI 32 F. Summary 36 III. DEVELOPMENT OF THE PROBLEM 37 A. Limitations of the CPI 38 1. Some limitations of "objective" personality inventories 38 2. Response sets and the CPI 42 3. The appropriateness of a "sighted" test 45 B. Statement of the Problem 45 1. Standardization phase 46 2. Cross-Validation phase 47.IV. METHOD 48 A. Subjects 48 1. Standardization sample 48 2. Cross-Validation sample 54 B. Instruments 57 1. Standardization sample 57 2. Cross-Validation sample 58 v

LIST OF TABLES Table Page 2.1. CPI Scales Related to Achievement Motivation 19 2.2. CPI Scales Related to Occupational Training 20 2.3. CPI Scales Related to Academic Achievement 22 2.4. CPI Scales Related to Personal Functioning 28 2.5. CPI Scales Related to Cross-Cultural Applications 31 4.1. Sources of Subjects 51 4.2. Occupational Groupings of Employed Standardization Subjects 52 4.3. Years of Employment and Income of Standardization Subjects 53 4.4. Cross-Validation Subjects Classified According to Age and Current Vocational Status 56 5.1. Comparison of Employed and Unemployed on the Scales of the CPI 67 5.2. Distribution of Scores on Employment Key Scale for Employed and Unemployed Groups in the Standardization Sample 71 5.5. F Values for Each Scale Prior to Entry into the Discriminant Functions 72 5.4. Stepwise Entry of Variables with Corresponding F Values at Each Step 74 5.5. Discriminant Functions at Each Step, Variable Weights, and Constants 75 5.6. Correct Classification of Subjects by the Discriminant Functions Generated with the Two-Variable Equation 76 vii

LIST OF TABLES (Continued) Table Page 5.7. Distribution of Scores on the Employment Key Scale for the Cross-Validation Subjects 79 5.8. The Discriminatory Power of the Employment Key Scale with a Suggested Cutting Score at 12 80 5.9. Predictive Accuracy of the Two-Variable Equation with the Cross-Validation Subjects 81 B.1. Current Occupations of Employed Subjects in Standardization Sample 105 B.2. Age of Standardization Subjects 106 B.3. Type of Visual Impairment of Standardization Subjects 106 B.4. Years of Academic Education of Standardization Subjects 107 B.5. Functional Vision of Standardization Subjects 107 B.6. Functional Mobility of Standardization Subjects 108 B.7. Marital Status of Standardization Subjects 108 B.8. Race of Standardization Subjects 109 B.9. Area of Residence of Standardization Subjects 109 B.10. Type of Visual Impairment of Cross-Validation Subjects 109 B.11. Years of Education Reported at Time of Evaluation of Cross-Validation Subjects 110 B.12. Reported Vision of Cross-Validation Subjects 110 B.13. Verbal I.Q. (WAIS) of Cross-Validation Subjects 111 B.14. Race of Cross-Validation Subjects 111 B.15. Marital Status of Cross-Validation Subjects 112 viii

LIST OF TABLES (Concluded) Table Page D.l. Formula for Estimating the Mahalanobis Distance Between Two Populations 126 D.2. Within Groups Correlation Matrix 128 D.3. Employment Key Items 129 D.4. Comparison of Scale Means Across Standardization and Cross-Validation Samples 131 D.5. Comparison of Scale Standard Deviations Across Standardization and Cross-Validation Samples ]32 D.6. Cross-Validation Subjects' Actual Group Membership, Predicted Classification, Squared Mahalanobis Distances, and Associated Probability of Predicted Group Membership for the Two-Variable Equations 135 D.7. Cross-Validation Subjects' Actual Group Membership, Predicted Classification, Squared Mahalanobis Distances, and Associated Probability of Predicted Group Membership for the Five-Variable Equations 136 ix

LIST OF FIGURES Figure Page 5.1. Comparison of employed and unemployed Standardization subjects on the CPI. 68 D.1. Projection of "new" scores onto discriminant function yl. 127 D.2. Comparison of mean scale scores of working and nonworking Cross-Validation subjects. 133 D.3. Mean scale scores of college students in the CrossValidation sample. 134 D.4. Grand mean scale scores across all groups. 37 x

ABSTRACT The purpose of the study was to determine the extent to which the California Psychological Inventory (CPI) is an effective instrument for identifying personality trait differences between groups of employed and unemployed blind, and whether such differences could be used to predict employability among the blind. This study consisted of Standardization and Cross-Validation phases. The Standardization sample included 51 visually impaired males who were employed full time for at least 2.8 years in occupations not requiring a college degree and 50 visually impaired males who were unemployed for at least 2.7 years. The CPI was administered by means of a tape recorded presentation of the items for card sort according to their agreement or nonagreement with each item. An Interview instrument was developed and administered to each subject for the purpose of comparing the employed and unemployed on selected demographic characteristics. Initial analysis of CPI data indicated that the unemployed scored significantly lower than the employed on nine scales. When the scale score differences were analyzed by a Discriminant Analysis for two 2 groups, a Mahalanobis D or generalized difference between the groups was significant at the.001 level of confidence suggesting the dissimilarity of personality traits between the employed and unemployed. An Employment Key scale was developed with 21 CPI items which best xi

discriminated between the employed and unemployed. With a suggested cutting score at 12, 90% of the employed and 72% of the unemployed Standardization subjects were identified. Stepwise Discriminant Analysis identified the most efficient combination of CPI scales for predictive classification of subjects. Two and five variable equations were developed. The most efficient used the Employment Key and Tolerance scales with the following coefficients: Employed = -11.459 +1.698Ek -.116To; Unemployed = -4.734 +.596Ek +.243To. These discriminant functions are for use with raw scores; whichever equation produces the larger value determines the subject's predicted classification, employed or unemployed. These equations classified correctly 940 of the employed and 82% of the unemployed in the Standardization sample. These findings were applied to 15 unemployed and 11 employed Cross-Validation subjects with the following results. The Employment Key identified 8 2% of the working and 7530 of the nonworking subjects. " The discriminant functions established on the Employment Key and Tolerance scales classified 82% of the working and 80% of those not working. These findings provisionally suggest personality traits which may differentiate the employed from the unemployed. The employed seem to have stronger feelings of self-worth, to be less inhibited by excessive fear, and to have a greater capacity for both adaptive behavior and commitment to goal directed activities. The employed may be more permissive and accepting of the personal beliefs of others as well as less xii

authoritarian and self-centered than their unemployed counterparts. In general, the results are promising and may provide a useful approach to employment prediction as well as identify certain relationships between personality traits and employability. The CPI would appear to be a useful instrument with the visually impaired. For those concerned with developing the manpower of the visually impaired it should be emphasized that the Employment Key scale items and the discriminant functions presented can be conveniently used on an individual basis. Two important considerations pertain to the application of these findings. The limitations of the relatively small samples and the unanticipated difficulties encountered in obtaining unemployed subjects should be recognized. Additional analyses for data collected in this study were suggested and implications for further research were discussed. xiii

CHAPTER I INTRODUCTION Barriers to employment commensurate with an individual's potential may arise from long term unemployment, social disadvantage, minority group membership, displacement by automation, or various physical and emotional impairments. The ranks of both the unemployed and underemployed waste manpower and decrease the Nation's productivity. Personal hardship and the economic drain of public assistance are additional by-products. It is well known that body disablement creates for the individual a special set of problems. These problems are unique for each individual in that their existence depends upon the degree of body dysfunction, its limitation on normal functioning, and individual's perception of his impairment, and how he adapts to it. While body impairment itself may be vocationally handicapping for certain jobs, difficult questions arise as to which other factors, or combinations thereof, contribute to successful employment for some persons while not for others who have comparable physical limitations. Economic deprivation from either under-employment or unemployment is a chronic problem among the visually impaired. Scholl (1969), studying a group of 126 visually impaired with a mean IQ of 106.7, found the mean annual income for 1967 to be $5,500, with 12 subjects receiving $500 to $1,000, and three receiving $20,000 or more. A 1

2 total of 34 subjects was receiving some kind of public assistance. Workers with the blind attribute the high rate of under-employment and unemployment to a variety of factors, in addition to visual impairment, which in and of itself may limit employment opportunities for the blind. These include: low intelligence, poor adjustment to visual handicap, poor travel ability, over-protective home situations, inadequate job training facilities for the visually impaired, an insufficient number of job opportunities, discrimination by employers, and a system of Social Security Disability benefits and Public Assistance which tend to increase dependency by jeopardizing payments when efforts to work are made. Many of these factors represent influences that are beyond the control of the individual; some,.however, are directly related to personality. A knowledge of personality characteristics which are more directly related to successful employment may alert rehabilitation agency personnel to those potential personality problems which may present barriers to full employment. Most rehabilitation workers include in the psychological evaluation of the client a systematic assessment of personality functioning as well as intellectual ability, aptitude and interest patterns. It is possible that systematic personality measurement can contribute to client understanding relative to such factors as readiness for vocational training, personal adjustment training, and the capacity for work responsibilities. Such information may be particularly

35 helpful in interdisciplinary settings which focus on the preparation of blind persons for employment. The purpose of this study was to explore whether a systematic assessment of personality functioning of the visually impaired could identify relationships between certain personality traits and potential for employment.

CHAPTER II RELATED LITERATURE Research related to personality and personality assessment of the visually impaired in presented in the following sections: psychological implications, psychoanalytic viewpoints, and personality testing. The final section reviews research on the California Psychological Inventory. Psychological Implications of Blindness The psychological implications of blindness are in general derived from studies of impairment itself. Raskin (1962) takes the position that blindness in and of itself does not produce maladjustment but that it does present problems in practical living. The correlates of maladjustment are derived from the attitudes of others, particularly parents and society. This view is consistent with that of Cowen, Underberg, Verillo, and Benham (1961). Lowenfeld (1955) maintains that blindness creates problems sui generis only in certain areas of cognitive functioning and in mobility. Underberg, Verillo, Benham, and Cowen (1961) found no systematic differences in personality attributes on seven global measures of adjustment for blind and nonblind. Scott (1969), Cutsforth (1962), and Freedman (1965) discuss the effects of the interpersonal attitudes of the family upon the personal 4

5 ity development when a blind individual lives in a social world which is unaccustomed to the condition of blindness. MacFarland (1966) suggests that the intensely visual orientation of our society leaves the blind decisively isolated from easy social interaction and practical convenience. Rusalem (1950) and Cowen, Underberg, and Verillo (1958) studied social attitudes toward the blind. Significant correlations were found between negative attitudes to blindness and various antiminority, and proauthoritarian attitudes. Krause (1962) and Carroll (1961) postulate a relationship between the psycho-social dependency needs of the blind as acquired from the family setting and future rehabilitation progress. Bauman (1954; 1963) takes the position that the difference between being self-supporting and being dependent is related to the attitudes of the blind toward themselves, others, and toward blindness itself rather than to such physical facts as the amount of vision, health, or education. Cole and Taboroff (1955) and Freedman (1966) discuss social and environmental aspects which they feel bear upon the personality growth and development of the blind individual. The personal response to physical loss, including visual impairment, and the adjustive efforts to permanent disablement are discussed by Dembo, Leviton, and Wright (1969) and Wright (1960). In summary, it appears that personality adjustment among the blind is more related to such factors as home background, parental and

6 community attitudes, and attitudes toward self than to the blindness itself. The reader is referred to Lowenfeld (1955; 1963) for an extensive review of the literature regarding the psychological problems of visually impaired children, and to Goldberg (1969) for a relatively current, comprehensive review of research on blindness. Some Psychoanalytic Viewpoints on Blindness There have been set forth some psychoanalytic views regarding blindness. Freud (1959) notes that the eye as the faculty of vision provides a link between the ego and consciousness and the external world. Blank (1957, 1958a, 1958c) discusses the symbolic significance of the eye and vision relative to unconscious conflicts in psychosexual and ego development in both the congenitally and the adventitiously blinded. He stresses the importance for workers to recognize the possibility of unconscious conflicts arising from their own relationships with blind clients (Blank, 1958b). Cholden (1952; 1958), in discussing the adjustment process in adventitious blindness, stresses the need for personal internal reorganization before the person is able to awaken to the fact that he is, after blindness, a different person with a different body image. He stresses the role of the ego-recovery capacity to deal with the shock period, emotional withdrawal, and mourning. Fries and Woolf (1953)'stress the impact of environmental experi

7 ence occurring before the differentiation of the ego and the id and compare this influence with the importance of constitutional factors. Fraiberg and Freedman (1964) studied arrested ego development of certain deviant blind children whose personalities remained frozen on the level of mouth-centeredness and nondifferentiation. Fraiberg, Siegel, and Gibson (1966) suggest the importance of the role of vision in early ego development and the function of vision in facilitating and insuring the autonomous functions of the ego. Nagera and Colonna (1965) discuss the differences and similarities between blind and sighted children relative to ego and drive development, regression and fixation points, tolerance of anxiety, reaction to frustration, and potential for sublimation. Sandler (1963) suggests there is an irreversible regressive pull toward self-centeredness with a basic retardation in ego development in blind children. She suggests the likelihood of a lesser drive for mastery and a retreat from the external world which may have a profound effect on later stages of development. Burlingham (1961) suggests that the blind child remains in a prolonged state of dependency upon objects which substitute for the function of the eyes. Burlingham (1965) notes that motor immobilization is a mode of self-protection for young blind children and suggests that this motor restraint is displaced to other ego functions thereby resulting in blindisms and other repetitive behaviors. Further, the process of verbalization is divided between certain meaningless words of

8 the sighted based on memorized definitions and word symbols developed through their own personal experience. In general, the psychoanalytic view stresses the important role of vision in ego development and functioning. Personality Testing of the Blind Research in the area of personality testing with the blind is not extensive. This section summarizes relevant studies on the use of projective and "objective" tests with the blind. PROJECTIVE TECHNIQUES Wachs (1966) evaluated the following projective techniques which have been or could be used with the blind: the Auditory Apperception Test (ATT), the Gramphomotor Projection Technique, the Insight Test (IT), the Twitchell Allen Three Dimensional Personality Test, and the Rotter Incomplete Sentences Blank (ISB). He suggests that better scoring methods and normative data for the blind be developed for the instruments. In addition, he proposes that a qualitative, thematic scoring of the ISB would be more useful with the blind than the objective scoring method. This position is supported by Dean (1957). Wright and Trotter (1968) suggest a battery of tests that may be useful in

9 personnel selection. A Sentence Completion Test, developed for this project, correlated significantly with job hierarchy and salary. Lebo and Bruce (1960) and Raskin (1962) review the status of projective techniques for the blind and indicate that such measures have not "caught on" and the lack of normative data renders them of provisional value. "OBJECTIVE" TESTS Paper and pencil type personality inventories have been used with blind subjects. However, there are certain reservations expressed regarding the use of "sighted" norms, the significance of the adjustments necessary for test administration, and the meaning of certain items to persons blinded early in life. Hayes (1949) pioneered the testing movement with the blind including the use of personality inventories. However, he is best known for his work in the measurement of intelligence. The Minnesota Multiphasic Personality Inventory (MMPI), developed by Hathaway and McKinley (1943), has been used with the blind. Wachs (1966) reviews some modifications of administration to blind subjects. Cross (1947) used a braille version of the MMPI and found that the blind differed from sighted norms on only 20 items of the original 550. Dean (1957) likewise found "normal" MMPI profiles with his blind subjects. Bonk (1955) recorded the MMPI for card sort and identified MMPI profiles which differentiated between various occupational groups of the blind.

Bauman (1950; 1954) developed the Emotional Factors Inventory (EFI), an eight scale, true-false inventory of 170 items which purports to measure adjustment to blindness and excludes items which may have different implications for a blind person. Items are reportedly derived from an analysis of statements of rehabilitation workers about the personality qualities and difficulties of their clients. An Adolescents Emotional Factors Inventory is also available (Bauman, Platt, and Strauss, 1963; Bauman, 1964). Dean (1957) compared judges' ratings of subjects' adjustment to blindness with scores obtained from the MMPI, EFI, Rotter Incomplete Sentences Blank, and the Sargent Insight Test. None of these measures agreed with adjustment as determined by judges' ratings in his group of 53 rehabilitation candidates. However, he felt the MMPI is applicable to the blind without modification and that the need for separate norms was not indicated. His EFI and MMPI protocols suggest that the blind are not paranoid or depressed as a group. He reported that the EFI did not discriminate adjustment of the subjects in his design, and' that some items showed too much variability suggesting the need for more refinement to be useful for individual prediction. Further, he recommends the Sargent Insight Test for cautious use with the blind. The use of the Bernreuter Personality Inventory and California F Scale are not recommended by Wachs (1966) for use with the blind on the basis of the deviant results reported by Greenberg and Jordan (1957).

11 Sources of information on tests and testing techniques for use with blind adults and adolescents have been developed by Bauman and Hayes (1951) and Dishart (1959; 1960). Although there appears to be interest in the area of personality testing of the blind, there is little systematic or definitive research on the relationship of such tests to performance criteria. Studies summarized in this section seem to be more concerned with factors such as modifications necessary for administration to the blind and comparisons of the blind on "sighted" norms than with the application of test results to predictive criteria. In addition, it appears that one widely accepted personality measure, namely, the California Psychological Inventory, has not been systematically studied with blind subjects. Prior experience with the CPI in a rehabilitation setting led the investigator to consider this instrument to study the question of whether a standardized "objective" personality test could discriminate between employed and unemployed visually impaired subjects. An additional reason for utilizing the CPI was the extensive literature available concerning its application. The following section summarizes some of the research related to the California Psychological Inventory. The California Psychological Inventory GENERAL DESCRIPTION The California Psychological Inventory (CPI) is a paper-and-pencil

12 personality assessment device developed for the comprehensive, multidimensional evaluation of "normal" persons (Kelly, 1965). The test includes 480 items which yield 18 scale scores. Answer sheets are available for hand or machine scoring. A special test profile sheet has been published on which the 18 scale scores may be plotted. A copy of this profile sheet and a listing of the standardized scales, scale abbreviations, and classes of scales from the Manual may be found in Appendix C (Gough, 1957). The inventory can be administered either individually or in groups.* The standardized instructions require the subject to read each item, decide whether he agrees or disagrees with what is stated, and then mark true or false on the answer sheet. The subject may omit items if he prefers not to answer. Testing time is about 45 minutes. Items may be read aloud for poor readers (Bennett and Rudoff, 1957), explained if questions are asked, and the inventory may be completed in different sessions, with or without supervision (Gough, 1968c). The inventory has been used on a mail-out/mail-in basis (MacKinnon, 1962). The inventory has been used from the junior high school level (Keimowitz and Ausbacher, 1960; Pierce, 1961), through high school (Gough, 1964a; Snider, 1966), at the college level (Aiken, 1963; Johnson and Frandsen, 1962; Cersen and Parker, 1966), across educational levels (Schendel, 1965), in adult life (Goodstein and Schrader, 1965; Gough, 1966a) andin old age (Schaie, i959). *Personal communication with Harrison G. Gough, April 30, 1970.

15 The CPI author does not imply that the 18 scaled dimensions of the CPI are independent but groups them into four categories (Kelly, 1965). The scales in Class I share a common emphasis on feelings of interpersonal and intrapersonal adequacy designeted as "Measures of Poise, Ascendency, and Self-assurance" (Scales Do, Cs, Sy, Sp, Sa, and Wb). The scales of Class II are concerned with social norms, values, and the disposition to observe or reject such values. They are described as "Measures of Socialization, Maturity, and Responsibility" (Scales Re, So, Sc, To, Gi, and Cm). The scales of Class III bear on academic and intellectual endeavors and are designated as "Measures of Achievement Potential and Intellectual Efficiency (Scales Ac, Ai, and Ie), Class IV scales (Py, Fx, and Fe) are described as "Measures of Intellectual and Interest Modes." Approximately 200 of the CPI itemsappeared originally in the MMPI. Eleven of the 18 scales were empirically derived, i.e., developed with items which were found to differentiate defined criterion groups such as dominant versus nondominant individuals so defined by independent criteria. Four scales were constructed by gathering items bearing a theoretical relevance to a personality trait and refining them through internal consistency analysis. Three remaining scales were derived empirically to detect tendencies of subject to fake good or bad, or respond in a manner which makes the other scales of doubtful validity (Kelly, 1965).

THEORETICAL ASPECTS OF THE CPI The CPI is designed for the assessment and description of interpersonal behavior. The concepts involved are intended to provide a description and analysis of personality in everyday life and social interaction (Gough, 1965a, p. 294). Gough theorizes that these "folk concepts" emerge from interpersonal behavior, are immediately meaningful, universally recognized, and relevant to the ongoing processes of human behavior in all cultures and societies (Gough, 1968c, pp. 56-58; Gough, 1965a). The CPI has been scaled to describe current behavior (Gough and Heilbrun, 1965) and to forecast behavioral outcomes. Gough (1964c) questions whether positive outcomes as personal achievement can be predicted from measures of clinical disturbance, anxiety, and distress. He proposes that personal adequacy is not synonomous with the absence of pathology but is related to positive achievement-oriented dimensions of personality which are capable of measurement. Since Gough (1968c, p. 65) states that the purpose of each scale is "to predict what an individual will do in a specified context, and/ or to identify individuals who will be described in a certain way," the distinction must be made between this and the more common goal in inventory measurement, that of trait specification. For example, an elevated score on the scale for social status does not demonstrate the presence of a "trait" of social status but rather that one possesses dispositions leading toward such attainment or that his outlook or

15 viewpoint is similar to people of high status. Similarly the occupational scale of computer programmer on the Strong Vocational Interest Blank (Strong, 1959) is not intended to define a unidimensional vocational interest trait of computer programmer. Rather, a high score on this scale indicates the subject's similarity in viewpoint and outlook on a variety of human activities with other computer programmers. While the CPI may seem pretentious in its presentation of 18 scaled variables, the test author (Gough, 1965) sees his system of concepts as an "open" system which can change by the addition, elimination, or interaction among selected scales. Its purpose is to reflect interpersonal adjustment by a sufficient number of variables so that all major forms of behavior can be forecast. The CPI manual (Gough, 1957) offers suggestions for interpretation of the scale scores on the profile sheet: (1) consideration of the overall elevation of the scales in relation to the general norms; (2) concern for high or low scale groups; (3) examination of the interaction between the more extreme scores and whether such reinforce or ameliorate one another; (4) study of the unique features of the profile, and (5) a consideration of an individual's scale scores about his own "personal mean." RESEARCH. APPLICATIONS OF THE CPI It seems reasonable to assume that certain personality characteristics are related to successful employment. The capacity to organize

one's resources into goal-directed, productive behavior which stresses the disciplinary implications of developing job skills and conforming work habits should have such a relationship to the potential for competitive employment. The CPI supposedly taps these personality dimensions. CPI studies on personality characteristics which may be related directly or indirectly to employment potential are reviewed under the following groupings: achievement motivation, occupational training criteria, academic achievement, personal functioning, and cross-cultural applications. For brevity, the CPI scales of consequence identified in the investigations reviewed are reported in the tables following each grouping of studies. It should be noted that an occasional scale becomes relevant because of a negative weighting or low scale score in the pattern of most predictive scales identified by multiple regression equations. For example, a negative coefficient is frequently assigned to the Good impression, or social desirability scale in regression equations. Achievement Motivation The following studies report the relevance of the CPI scales to the measurement of achievement motivation. Table 2.1 summarizes the scales which the authors for each study considered relevant. Leadership is an interpersonal behavior which is an appropriate criterion for measurement with the CPI scales. Several investigators

17 report evidence of elevated CPI protocols being associated with independent ratings of leadership. Liddle (1958) found highly significant correlations between ratings of leadership and total CPI elevations. Johnson and Frandsen (1962) found similar, elevated profiles in their comparisons of 50 student leaders and 50 randomly selected nonleaders at the college level. Carson and Parker (1966) grouped 164 college males into 3 categories of leadership, past presidents of student bodies, former class presidents, and prior social group presidents. Significant F ratios distinguished the first group from the remaining two groups. Gough (1969) developed a leadership index on a group of high school students designated as outstanding in that personal characteristic. On 11 CPI scales 90 males and 89 females scored significantly higher than controls. Biserial correlations of his regression equation with leadership versus nonleadership dichotomies were.45 for males and.44 for females. The Good impression scale received a negative weighting in the equation which excludes the desire to please from the "leadership" equation. Megargee, Bogart, and Anderson (1966) found that Dominance scale scores reflected situations of initiative with simulated industrial tasks. Goodstein and Schrader (1963) empirically developed a managerial scale from items which reliably classified the total managerial group from men-in-general and differentiated personnel at three different

18 levels of management. The scales contributing the most items to the scale are likewise listed in Table 2.1. Rawls and Rawls (1968) studied differences between 30 highly successful and 30 less successful executives with two personality measures, the CPI and a Biographical Information Blank. Significantly higher CPI scores were found on eight CPI scales. It is important to note that the less successful group scored significantly higher on the Selfcontrol and Femininity scales. The biographical information corroborated the findings of the personality scales. Schwartz, Dennerll, and Lin (1968) contend that the employability of persons with epilepsy must be conceived of from a multidimensional viewpoint. They found five CPI scales to be associated with current employment status along with intellectual and neuropsychological measures. Occupational Training Criteria Selected studies concerned with predicting excellence in occupational training are reviewed in this section. Table 2.2 summarizes the relevant scales identified in these studies. Gough, Durflinger, and Hill (1968) found selected CPI scales useful in predicting student teacher effectiveness. Scales selected by a regression equation in their analysis correlated.44 on a CrossValidation sample with higher and lower rated student teachers. Hill (1960) found the CPI scales useful in measuring the relationships

TABLE 2.1 CPI SCALES RELATED TO ACHIEVEMENT MOTIVATION Relevant Scales;y Sp Sa Wb Re So Sc To Gi (......... Reference Johnson & Frandsen (1962) Carson & Parker (1966) Gough (1969) Megargee, Bogart & Anderson (1966) Goodstein & Schrader (1963) Rawls & Rawls (1968) Schwartz, Dennerll & Lin (1968) Do Cs s Cm Ac Ai Ie Py Fx Fe x x x x x x x x x x x x x x x x H \10 x x x x x x x x x x x x x x x x

TABLE 2.2 CPI SCALES RELATED TO OCCUPATIONAL TRAINING....................Reference -_____Relevant Scales Reference Do Cs Sy S S a Wb Re So Sc To Gi Cm Ac Ai Ie Py_ Fx Fe Gough, Durflinger & Hill (1968) x x x x x Hill (1960) x Query (1966) x x x x x Gough & Hall (1964) x x x x Watley (1968) x x x x x Southern & Plant (1968) x x x x R) 0

21 between personality characteristics of student teachers and teacher excellence. Query (1966) used the CPI scales to distinguish "successful" from "unsuccessful" seminary candidates. In the absence of effective measures predicting performance in medical training, Gough and Hall (1964) developed a regression equation with CPI scales having a predictive validity of.66 in an initial sample and.46 in a cross-validating sample of 63 persons. Their equation of "medical promise" is described as stressing unselfishness and consideration for others rather than need achievement, compensatory striving, or scholastic attainment. Watley (1969) followed the educational and career progress of highly gifted students. Those who made the least career progress differed from their opposites by scoring lower on three CPI scales. It is noteworthy that those who progressed least scored highest on the Flexibility and lowest on the Femininity scales. Southern and Plant (1968) found that their group of very bright adults scored significantly higher than the normative samples on three CPI scales. These very bright persons scored lower on the Self-control scale. Academic Achievement Numerous investigators have used the CPI measures to predict success in academic endeavors. Table 2.3 summarizes the scales reported

TABLE 2.3 CPI SCALES RELATED TO ACADEMIC ACHIEVEMENT Relevant Scales Cs Sy Sp Sa Wb Re So Sc To Gi Cm Ac Ai Ie Py Fx Fe Reference Do Bowman & Matthews (1960) Gough (1966b) Gough (1964a) Snider (1966) Gough & Fink (1964) Pierce (1961) Fink (1963) Gill & Spilka (1962) Gough (1964b) Rosenberg, McHenry, Rosenberg & McNichols (1962) x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x X x x x x x x X X X X x x x x x x x X X x X X X X X X X x x x x

TABLE 2.3 (Concluded) Relevant Scales Reference Do Cs Sy Sp Sa Wb Re So Sc To Gi Cm Ac Ai Ie Py Fx Fe Holland (1959) x x x Domino (1968) x x Gough (1968b) x x x x x x Norfleet (1968) x x x x x

as being relevant. Bowman and Matthews (1960) report that CPI testing at the 10th grade level differentiated 55 high school dropouts from 112 graduates who were matched for IQ and socio-economic status. Gough (1966c) studied the power of the CPI scales to identify dropouts from a large sample of high school students. He developed a regression equation which optimally selected the most differentiating combination of scales. Maxwell (1960) found seven CPI scales which significantly forecast graduation versus dropping out from college in a group of 400 male students equally divided between graduates and dropouts. Gough (1964a) studied large samples of high school students from 11 states and made a cross-cultural validation in four Italian schools. His regression equation stressed the relationship of high school achievement to internalized value systems in which self-discipline, control of impulse, and adherence to value are important. The achievement motive, both adaptive and independent, were emphasized. The Intellectual efficiency scale supported the cognitive ability in gradepoint average. The negative weighting on the Good impression scale discourages the undifferentiated desire to please. Snider (1966) used Gough's (1964a) regression equation to identify high school achievers with an overall percentage of 65% correct "hits." Snider's regression equation, coincided remarkably with that developed by Gough (1964a). Gough and Fink (1964) studied the ability of CPI scales to forecast grade point averages among students of average

25 ability. Fourteen of 18 CPI scales correlated significantly with grade point averages when applied to subjects in the cross-validation sample and a.55 coefficient was obtained with an unselected sample. Pierce (1961) identified more positive personality characteristics among high achievers with the CPI. Fink (1963) identified 52 CPI items which differentiated achievers from underachievers at the.01 level of confidence. Gill and Spilka (1962) matched groups of Mexican-American high school students on IQ. Significantly differentiating CPI scales described the achievers as more mature than their underachieving peers. Lessinger and Martinson (1961) found CPI scales useful to assess psychological maturity among gifted students and concluded that chronological age norms are not completely useful for the gifted since they differ strikingly from their contemporaries. Other investigators have found the CPI to forecast success in specific courses. Keimowitz and Ansbacher (1960) report 13 scales which differentiate achievers from underachievers in mathematics. Gough (1964b) found the Achievement via independence scale to be the best single predicter of achievement defined as course grades in large samples of introductory psychology students. His regression equation on a cross-validation sample was.41. Rosenberg, McHenry, Rosenberg, and Nichols (1962) predicted academic success or failure of students in three military courses. They found final grades and seven CPI scales to correlate with.01 significance. Their predictive equation

26 emphasizes both intellectual and motivational factors in determining academic success. Holland (1959) studied finalists of the National Merit Scholarship Corporation and found that CPI scales yielded predictive validities significantly superior to those derived from aptitude scores. These students described themselves on the CPI as more socially introverted, responsible, mature, and conforming to recognized societal standards than their underachieving peers. Domino (1968) studied the interaction between achievement motivation and the demands of the environment, defined as scholastic achievement. He supported his hypothesis that conforming and independent achievement motivation, as measured by the CPI, are related to scholastic achievement. Gough (1968a) used CPI scales to forecast college attendance by high ability high school graduates. He reports predictive validities with his regression equation of.51 and.37 for males and females, respectively, and.52 for unselected subjects in Cross-Validation. Similar to his earlier study, (Gough, 1964a) the Good impression scale received a negative weighting suggesting unconcern for social desirability among high achieving students. Norfleet (1968), studying gifted university women, found five CPI scales which differentiated achievers from underachievers. She notes that her two groups also described themselves differently on the Adjective Check List (Gough, 1960).

27 Personal Functioning The CPI has been found useful as a diagnostic measure of personal functioning in counseling procedures and as an indicator of asocial behavior. The CPI scales involved in these studies are noted in Table 2.4. Goodstein, Crites, Heilbrun, Jr., and Remple (1961) used the CPI in a university counseling center because of its applicability to a relatively stable, nonpsychiatric population. They found it useful for distinguishing personal adjustment cases from vocational-educational problem cases, and for separating both samples from control groups. Levanthal (1966) developed an Anxiety scale with 22 CPI items on a university counseling center population. Additional data (Levanthal, 1968) indicated that persons with high scores on this scale had poorer prognosis in counseling, required more interviews to termination, and improved at a slower rate. He reports negative correlations between the Anxiety scale and certain CPI scales and suggests that high scores on the Anxiety scale are erratic, conflicted, and self-centered. Finney and Van Dalsem (1969) report improved measures on eight CPI scales after group counseling with 68 academically gifted but underachieving high school students. The counseled students were rated by teachers as more cooperative in the classroom and absent from the classroom less than the controls. There were no observed differences nor improvements in grade-point averages nor in the scores of the California Study Methods Survey.

TABLE CPI SCALES RELATED TO 2.4 PERSONAL FUNCTIONING Reference Goodstein, Crites, Heilbrun, Jr. & Rempel (1961) Levanthal (1966) Levanthal (1968) Finney & Van Dalsem (1969) Peterson, Quay & Anderson (1959) Gough, Wenk, & Rozynko (1965) Stein, Gough & Sarbin (1966) Gough (1966c) Relevant Scales Do Cs Sy Sp Sa Wb Re So Sc To Gi Cm Ac Ai Ie Py Fx Fe..~~ ~ ~ ~ ~~ Ai, I e Py,F F x x x x x x x x x x x x x x x x x x x x ro oo x x x x x x x x x x x x x x x x x x x x

29 The Socialization scale of the CPI has been used successfully in identifying asocial behavior. Peterson, Quay, and Anderson (1959) used the Socialization scale of the CPI with 239 inmates of a boys' training school and a group of nondelinquent controls. The scale correctly identified over 70% of the members from both groups. Significant differences were found between delinquent recidivists and first offenders, and between "good citizens" and "disciplinary problems." Gough, Wenk, and Rozynko (1965) report that a combination of both demographic data along with CPI scales provide the best prediction of parole outcome in their sample. Stein, Gough, and Sarbin (1966) found three main dimensions in the 54 items of the Socialization scale, each of which differentiated significantly between delinquents and nondelinquent males. Gough (1966a) compared large samples of nondelinquent males with institutionalized delinquents. Seventeen of the 18 CPI scales differentiated significantly between the two groups. The Socialization scale was most discriminating. A six variable regression equation produced a point-biserial coefficient of.65 on delinquent and nondelinquent CrossValidation groups. Screening efficiency with the equation reached 90% and 82% in identifying delinquents and nondelinquents, respectively. A number of CPI studies focusing upon social maturity and socialization have been carried out cross-culturally and are reviewed in the following section.

30 Cross-cultural Applications Translations of the CPI into other languages have made possible cross-cultural studies with this instrument. Gough (1968b) suggests that asocial behavior is a cultural universal found wherever people live and work together. Although the definitions of asocial behavior may differ, the psychological dispositions underlying social conformity and adaptation may be constant and psychologically measurable. Further, Gough (1968b) discusses methodological issues encountered in translation and cross-cultural measurement. Table 2.5 identifies the scales concerned in the cross-cultural studies reviewed below. Initially it should be noted that some cross-cultural investigations with the CPI were discussed above in the contexts of other criteria (Gough, 1964a; Gill and Spilka, 1962; Gough, 1966c). Gough (1964c) reports that he was able to predict the achievement motive in Italian schools from personality appraisals with the CPI. Mizushima and DeVos (1967) found that despite cultural differences between American and Japanese youth there were remarkable similarities in social attitudes related to delinquent activity in both cultures. Japanese delinquents scored significantly lower than nondelinquents on three scales. Gough, DeVos, and Mizushima (1968) used Gough's (1966c) equation with a Japanese sample. A predictive coefficient for the delinquent versus nondelinquent dichotomy was.69 and the authors identified a cutting score which classified 88.6~/ of the subjects correctly.

TABLE 2.5 CPI SCALES RELATED TO CROSS-CULTURAL APPLICATIONS Relevant Scales Sy Sp Sa Wb Re So Sc To Gi Cm x x x Reference Gough (1964a) Gill & Spilka (1962) Gough (1966c) Gough (1964c) Mizushima & DeVos (1967) Gough, DeVos & Mizushima (1968) Gough (1965b) Gough & Sandhu (1964) Gough, Chun & Chung (1968) Do Cs _ Ac Ai Ie Py Fx Fe x x x x x x x x x x x x x x x x H x x x x x x x x x x x x

32 The Socialization scale of the CPI was translated into eight languages in ten countries and administered to large numbers of delinquent and nondelinquent persons. Using an optimum cutting score of twentyeight, 780 of the males and 85o of the females were correctly classified according to the delinquent versus nondelinquent dichotomy (Gough, 1965b). Gough and Sandhu (1964) report that translations of the Socialization scale into Hindi and Punsabi were given to delinquents and college students in India. Behavioral ratings correlated.70 with Socialization scale scores. Gough, Chun, and Chung (1968) found that a Korean translation of the CPI Femininity scale differentiated females from males at the.001 level of confidence in Korean adolescents. The effectiveness of the CPI in cross-cultural studies supports the likelihood of its potential for application to handicapped populations such as the visually impaired. It should be noted that although not all the studies cited above bear directly upon employability or employment readiness, these studies are helpful in outlining some of the diversified applications made with this instrument and suggest the kinds of information which may become available through its use. Factor Analyses and the CPI Factor analytic studies of the CPI merit some discussion both in

33 relation to test profile interpretation and psychometric test theory. Initially it should be noted that Springob and Struening (1964) report a personal communication from Gough relative to four factor analyses he conducted prior to publication of the test and arrangement of the Profile Sheet. Springob and Struening state further that Gough supported his first two scale groupings on those findings, while the scales in Classes III and IV were not intended to be factorially grouped. Gough (1968c) states that his Class III and IV groupings are for diagnostic convenience. The available studies suggest that the reliable variance measured by the 18 scales of the CPI can be explained by four or five dimensions. It is important to note that the available factor analyses were made on restricted samples of student populations with varying factor analytic techniques. Consequently, extended comparisons of results between studies are difficult. These valuable but limited findings neither establish the dimensional structure of the CPI scales nor encourage unguarded generalization. Mitchell and Pierce-Jones (1960) identified four orthogonal factors with a somewhat different combination of scales than those presented by Gough. Their analysis on a college population found the strongest factor to include the CPI scales of Self-control, Good impression, Achievement via conformity, Well-being, Tolerance, and Responsibility. This Factor I was designated Adjustment by Social Conformity. Crites, Bechtoldt, Goodstein, and Heilbrun (1961) confirmed

these results and designated this factor Good impression, after the scale with the highest loading. This factor presumably represents the social stereotype of a stable individual who functions in a socially acceptable manner. Springob and Struening (1964). found close agreement as these scales emerged from a rotated factor matrix while studying a high school sample. They designated this factor Self-control since that scale is almost a pure measure of that factor. Nichols and Schnell (1963) approximated this factor on college students and designated it Value Orientation. Notice is made that in these studies the Well-being and Achievement via conformance scales are factorially drawn from Gough's Class I and III groupings. Crites, Bechtoldt, Goodstein, and Helibrun, Jr. (1961), similarly to Mitchell and Pierce-Jones (1960), found five of Gough's six Class I scales to represent a second factor. These scales are Dominance, Capacity for status, Sociability, Social presence and Self-acceptance. Their designation of this factor was Dominance, the scale with the highest factor loading. Mitchell and Pierce-Jones (1960) named this factor Social Poise or Extroversion. Springob and Struening (1964) claim to have obtained similar findings while studying a college sample. Nichols and Schnell (1963) approximated this factor and titled it Person Orientation. Crites and associates (1961) contend that these two factorings measure distinctly different modes of adjustment. The first factor defines an attitude of compliance, cooperation, and adaption of self

35 to environment. The second factor underlies adjustment to changes in external reality rather than the self, stressing a disposition to direct and manipulate the environment and others. Further comparisons of the factor studies provide diminishing agreement among themselves. For example, the controid analysis of Mitchell and Pierce-Jones split the third factor of Nichols and Schnell into two. They entitled the first division Super-Ego Strength because of Gough's descriptions of the characteristics associated with those scales loading this factor, Communality, Socialization, Femininity, and, again, Responsibility. They identified a fourth factor which comprised the scales of Achievement via independence, Flexibility, Intellectual efficiency, and, again, loadings on Tolerance, Social presence, and Capacity for status. This grouping was designated Capacity for Independent Thought and Action. Springob and Struening also found a factorial association between the Flexibility and Achievement via independence scales in their Factor C. The studies reviewed above indicate that 18 independent aspects of personality are not measured by the 18 scales. Gough suggests the "open-ended" quality of the inventory and refers to the forecast of a certain behavior either by a single scale or a meaningful combination of scales (Gough, 1968c). Crites and associates (1961) contend that it is economical and efficient to use a reduced set of six reference scales which exhibit homogeneous correlation matrices for three groups of subjects. These scales include the measures of Dominance, Good

56 impression, Intellectual efficiency, Flexibility, Femininity, and Communality. These investigators feel that these six reference scales predict most of the reliable variance in the other CPI measures and provide more concise definitions of the variables assessed by the CPI. Summary This chapter reviewed some psychological implications of blindness and the role of vision from the psychoanalytic view. Studies on personality measurement of the blind were found to be limited in scope and almost devoid of systematic attempts to relate test results to prediction of performance criteria. Selected studies concerned with the California Psychological Inventory as a measure of interpersonal and intrapersonal functioning, social maturity, and achievement motivation were reviewed. The demonstrated usefulness of the CPI with various nontest criteria suggest its value as an instrument for exploring whether personality traits, as measured by the CPI, will identify employment potential among the visually impaired.

CHAPTER III DEVELOPMENT OF THE PROBLEM Research reviewed in the previous chapter suggests that the CPI may be a suitable personality measure for use with a visually impaired population. It is a relatively easy test to take,in that its "folk concepts" do not engender undue apprehension. It can be administered to blind subjects by volunteer reader or tape recording and card sort for auditory presentation (Bennett and Rudoff, 1957). The CPI is scaled for measuring interpersonal traits which have a broad relevance to personal and social adequacy, social maturity, achievement, and goal-directed behavior. A strong recommendation for the CPI is that it is extensively researched on a wide variety of populations. This immediately provides opportunity for comparisons between blind and sighted populations. CPI profiles of blind persons should provide the practitioner with meaningful descriptions of his clients and how each compares with the sighted. CPI profiles of the blind should provide more definitive information relative to the implied contention that there is a personality configuration which is characteristic of blind persons. Another recommendation for use of the CPI with the visually impaired is that the CPI profile analysis offers descriptions of personality traits in everyday concepts which are readily meaningful and comprehensible to the user. While more subtle implications may be 37

38 apparent to the skilled interpreter, the main thrust of the scales provides descriptions of interpersonal qualities which are both readily communicated to and understood by the client in counseling sessions. Finally, preliminary impressions of the CPI gained by its use with various disability groups at the Rehabilitation Institute, Detroit, Michigan, have been encouraging. There are certain limitations which should be noted. These are discussed in the following section. Limitations Of The CPI The following questions may be raised regarding the suitability of using the CPI in this study: limitations of "objective" personality inventories, response sets and the CPI, and the appropriateness of employing a "sighted" test with a visually impaired population. SOME LIMITATIONS OF "OBJECTIVE" PERSONALITY INVENTORIES Certain methological issues are involved in the attempted measurement of personality traits by inventory-type measurements. Initially, there was slow progress in this area. The literature suggests some reasons for this. Hathaway (1965) proposes that such may be due to the lack of a convincing theory of personality, undeveloped psychometric methodology, and difficulty with developing a satisfactory diagnostic system of personality traits. The "objective" designation of paper and pencil inventories, in

39 contrast to projective techniques, infers an objective scoring system under universally reproducible situations with a standardized treatment of test data. Yet, the term "objective" raises an interesting question, whether such tests are truly objective (Bass and Berg, 1959). The relationship between personality itself and inventory measurement poses issues as to what is meant by personality. The definitions of personality tend to vary with theoretical biases (Hall and Lindzey, 1957, pp. 1-28). Nunnally (1967, p. 470) suggests the measurement of personality traits be addressed to three broad classes of traits: Social characteristics, Motives (needs or drives) and Adjustment versus maladjustment. Hathaway (1961) suggests that personalities do not exist in a void but in relation to person-environment interaction. In this sense he proposes that personality tests are not culture-free, and that normative data depends on some culturally accepted value system for a basis of personality measurement. Another fundamental concern to personality inventories is the question of a developmentally-changing personality or a constant personality structure. Hathaway (1965) argues that cultural values obviously change and the person also. He suggests that there are both intrinsic and social aspects to personality which are measurable from the focus of what the person is and as observers see him. He posits that either some factors of personality are stable and others

40 variable, or that all factors have both stable and variable aspects. This middle posture respects the position of psychoanalytic viewpoint which supports the determination of personality structure at an early age and the requisite of intensive psychotherapy to effect changes. It likewise considers the position that personality changes with daily experience and is modified by age and learning experiences which allow for man's adaptability to changing environments. He does not see personality measures as witnessing an absolutely constant personality structure, and if such did, personality measurement would seem to be of little help in evaluating persons in times of situational stress. At the same time, personality inventories which are too sensitive to momentary change could be vitiated by momentary states and belie the recognition that individuals do indeed demonstrate some stability of personality traits. Hathaway (1965) stresses that the primary contribution of the personality profile is in its predictive value. Nunnally (1967, pp. 472-513) offers a review of the methods to find and measure general traits of personality. He defines personality traits as a "measurable dimension of behavior," whicrn ~s measurable either dichotomously or in finer gradations. He sees traits as varying in generality from specific habits, such as smoking rather than nonsmoking, to very broad dimensions of behavior, such as extroversion as opposed to introversion. He discusses personality traits relative to the nomothetic approach which strives to measure Personality charactI~~~~~~~~~~~~~

eristics of all people in terms of profiles of quantifiable traits. The ideographic or personalized point of view denies factors among personality characteristics and stresses that personality traits are so individualized that each is unique to the individual possessing it. The methodological choice of inventory items in personality test construction is an important consideration. Content validation or face validity refers to the selection of inventory items made on the basis of their relation to self-perceived or self-reported problems or characteristics (Hathaway, 1965); Anastasi, 1968). Empirical criterion keying is another approach whereby a criterion dimension is defined and inventory statements are assembled according to their capacity to identify those persons demonstrated as possessing the criterion dimension. The criterion dimension must be shown to characterize those persons by some procedure independent of the test instrument (Cronbach, 1960; Gough, 1957, 1968c). Others have approached item selection by defining constructs on the basis of personality theory and preparing items which bear on traits or dispositions particularly relevant to those constructs (Cronbach, 1960). A final method of selecting inventory items has been through internal consistency or factor analysis. In this method items are selected on the basis of their relationship to each other and to a theoretical relationship to some personality trait (Nunnally, 1967,

pp. 210-211; pp. 476-479). In spite of the limitations currently inherent in "objective" personality assessment, this technique remains important for the visually impaired because projective techniques for the blind have not been investigated (Wachs, 1966). RESPONSE SETS AND THE CPI An overview of personality inventory measurement requires consideration of the possible influence of consistent responses determinants or "sets" which result from item or inventory style rather than specific item content. The question arises whether self-report inventories are indeed dominatedby such response styles as acquiescence and social desirability. This issue was initially studied in detail by Edwards (1953; 1957) and by Jackson and Messick (1958; 1960). Investigations relative to response set with the CPI may be particularly relevant at this point. Jackson (1960) reports evidence suggesting that acquiescence and social desirability may account for a considerable amount of variance of the CPI. Dicken (1960) reports trait-simulation experiments with naive college students and psychologists who were not particularly well versed with the CPI. He concludes the CPI is a sufficiently subtle instrument for resistence to differential bias such as social desirability and the detectability of simulation. A cutting score on the validity scale, Good impression, identified biased records with

a high degree of efficiency among naive college students. The psychologists were able to simulate their scores distinctively without changing their validity scores sufficiently to permit efficient detection. He states the CPI appears more subtle by all criteria considered than the Edwards Personal Preference Scale (Dicken, 1959). Canter's (1963) simulation studies of the CPI are likewise supportive of the position that the CPI is not susceptible to undetectable distortion. Canter used some subjects who were not functioning well such as groups of poorly and better adjusted alcoholics and ward aide applicants instead of college students. He found that under "fake good" instructions all subjects could improve their test profiles but the subjects' actual life adjustment limited their capacity to improve their test pictures. He suggests that while conscious simulation is detectable, those who obtain higher Good impression scores have a potential for the interactional expectancies of socially adequate behavior (Gough, 1960a). He suggests that this is a meaningful aspect of the "real" personality and not a response set. Similar support is furnished by Lichtenstein and Bryan (1966) who used the Marlowe-Crowne Social Desirability Scale (M-C SDS) and found that high scorers on the M-C SDS were better adjusted than low scorers. They found eight CPI scales which yielded significant, replicated correlations. The scales of Good impression and Self-control were strongly related, followed by Achievement via conformity, Responsibil

44 ity, and Well-being. Consistent with Gough's (1957) expectations, the M-C SDS is most strongly associated with measures of socialization, maturity, and responsibility. Corroborative evidence is furnished by Dicken (1963b) who was unable to improve the validity of nine CPI scales when "correcting" for the response biases of social desirability and acquiescence over six independent samples. Dicken (1963a) validated the Dominance and Intellectual efficiency scales with convergent and discriminant validity techniques. When social desirability was controlled by means of a 32 item CPI social desirability scale, the overall discriminant validity was not improved. Pumroy (1962), using the Edward's Social Desirability Scale (Edwards, 1957) with the CPI, found that only five of the 18 CPI scales correlated with the SD scale in a positive direction. He indicated some correlation was to be expected because more positive adjectives are associated with elevated CPI scales. He contends with Gough (1957) that the empirical criterion-keying of most CPI scales facilitated more "subtle" items and minimized the effect of social desirability. He concluded that social desirability is involved with the CPI but it is not as important a factor as with many other personality tests. In summary, the current research suggests that the response sets of social desirability and acquiescence do not pose a serious problem with the CPI. Further, conscious simulation of the CPI is felt to be detectable and may lend support to the identification of certain

45 personality characteristics. THE APPROPRIATENESS OF A "SIGHTED" TEST A further question arises regarding the suitability of a "sighted" test for use with a visually impaired population. Bauman and Yoder (1966) take the position that a visually impaired population can be better evaluated with test items which are not only chosen specifically for blind subjects but which have a content validity corresponding to the varying problems presumed to confront blind people. Consequently, the variables felt to be measured by such items lend themselves to scales purporting to measure "adjustment to blindness" variables. This position is accompanied by the assumption that blindness presents the visually impaired with a potential for unique differences in personality development which requires an equally unique instrument for measurement. If one is not prepared to accept this assumption, an alternate option suggests the use of a widely standardized instrument, such as the CPI, which has been used with a variety of behavioral dimensions relating to interpersonal effectiveness, social maturity, and achievement potential. This study was undertaken on this assumption. Statement Of The Problem The purpose of this study is to evaluate the usefulness of the CPI in predicting the employment status of visually impaired males of

46 working age. The study involves two phases, Standardization and Cross-Validation. STANDARDIZATION PHASE Statistical treatment of the Standardization data was undertaken to meet the following three objectives. 1. To determine whether any CPI scales discriminate between the two groups of Standardization subjects who are differentiated by an employed-unemployed dichotomy. The null hypothesis (Johnson, 1967) can be stated as follows: H = The employed and unemployed subjects of the Standardization sample will not differ on CPI scale dardization sample will not differ on CPI scale scores. 2. To determine by empirical item-analysis whether any CPI items significantly discriminate between the employed and unemployed subjects. The null hypothesis can be formulated in this manner: H = The employed and unemployed subjects of the Standardization sample will not differ on the direction of item response. If the null hypothesis is rejected, the significant items can be assembled into a separate experimental scale, designated the Employment Key, and analyzed in conjunction with the standardized CPI scale scores. 3. On the basis of results obtained in Steps 1 and 2, a

47 determination can be made whether any scale or pattern of scales efficiently predicts the dichotomous criterion of employment versus unemployment among the Standardization subjects. CROSS-VALIDATION PHASE In order to determine the validity of the findings derived from the CPI data of the Standardization subjects, the findings will be applied to Cross-Validation subjects. CPI data from a separate group of subjects was obtained during previous administrations. The procedures employed and the results obtained are reported in the chapters that follow.

CHAPTER IV METHOD This chapter presents the experimental design of the study. It compares the essential demographic characteristics of the employed and unemployed subjects for both the Standardization and Cross-Validation samples. The instruments used for interviewing and testing the Standardization sample are discussed. Sources of data. for the CrossValidation subjects are identified. An account of the procedures includes the administration of the instruments, the analyses of the data, and the application of the findings to a separate group of CrossValidation subjects for evaluation of the findings of the Standardization procedures. Subjects STANDARDIZATION SAMPLE In order to determine whether the CPI can be effectively used to identify personality differences between the employed and the unemployed blind, two groups of subjects were selected: a substantially employed group and a chronically unemployed group. These groups will comprise the Standardization sample. This study was designed to focus upon the employed blind who had no specific professional training. Therefore, college graduates were

49 excluded with the exception of six who were employed in an occupation which did not require a college degree. Criteria for Selection The total Standardization sample included 101 males between the ages of 21 and 61, with a visual impairment of such severity that they were eligible for vocational rehabilitation services, Public Assistance (Aid to the Blind), or disability retirement benefits. The sample included two sub-groups: 51 subjects employed continuously in a "full time" capacity, i.e., 40 hours or more per week, during the past two years, and earning a minimum of $200 per month; and, 50 subjects unemployed for at least the past 2.5 years. In order to focus upon the employment-unemployment dimension and avoid possible variance from other sources, the following subjects were excluded: (1) those who had an additional, employment handicapping condition, such as orthopedic, severe diabetic, emotionally disturbed, mentally retarded, or similarly impaired which may markedly limit employment potential, particularly in combination with blindness; and (2) the recently blinded, i.e., those who sustained visual impairment since 1967. Selection of Subjects Subjects for the Standardization portion of the study were selected from the following sources: 1. Names in the file of the State of Michigan Division of

50 Services for the Blind; a letter was sent to a number of clients in the "inactive" files requesting their cooperation. Interested subjects were asked to sign a Release of Information form giving their consent to have the investigator initiate contact. A copy of this form may be found in Appendix A; 2. Eligible subjects included in the Scholl, Bauman, and Crissey (1969) research project; 3. A list of names furnished by the Vocational Consultant at the Metropolitan Society for the Blind, Detroit; 4. The subjects who volunteered the names and phone numbers of acquaintances, or in some instances actively worked to locate subjects; 5. Clients recently evaluated in the Psychology Department at the Rehabilitation Institute, Detroit. It was necessary to utilize this latter resource in order to obtain the 50 unemployed subjects. This group, when tested, was receiving extensive evaluations in the areas of personal adjustment, mobility, employability, vocational selection, and communication skills. Table 4.1 summarizes the sources of subjects. Much difficulty was encountered in identifying the group of unemployed subjects. Initially the unemployed group was intended to sample a relatively young, chronically unemployed population embracing persons

51 Soul TABLE 4.1 RCES OF SUBJECTS Employed -N.% - I J I. Source Michigan Services for the Blind Scholl, Bauman, & Crissey Research (1969) Metropolitan Society for the Blind, Detroit Subjects in the Current Study Rehabilitation Institute, Detroit Totals Unemployed N %0 10 20 3 6 14 27 5 10 5 6 5 10 24 47 25 50 0 0 51 100 12 24 50 100 who were either reluctant to initiate a, vocational training program for which they were judged to be capable, or unable to complete a program, or trained for work but unable to assume employment. Many unemployed blind persons were ineligible for this study because of multiple handicaps which perhaps rendered them unemployable for reasons other than blindness. Others were at or near retirement age. Most striking, however, was the repeated experience that the unemployed were no longer at the last recorded phone number or address despite relatively recent referral information. This experience may suggest that a proportion of the unemployed blind are, for whatever reasons, a transient segment of

52 the population whose economic hardship may have forced them into an unstable life-style, or vice versa. Finally, it was observed that the unemployed were conspicuously less willing to participate in the testing. Their reluctance may have been related to an unfounded fear that in some way the testing might jeopardize their Social Security Disability payments or Public Assistance funds by challenging their status as "unemployable. " Table 4.2 provides a general breakdown of the major occupational group divisions of the employed subjects in the Standardization sample. A breakdown by specific occupation may be found in Appendix B (Table B.1). TABLE 4.2 OCCUPATIONAL GROUPINGS OF TEE EMPLOYED STANDARDIZATION SUBJECTS Professional and. managerial 4 8 Professional and managerial 4 8 Clerical and sales 18 55 Service 4 8 Agriculture, fishery, forestry 0 0 Skilled 4 8 Semiskilled 17 33 Unskilled 4 8 Totals 51 100 Description of Standardization Subjects Table 4.3 presents information relative to the years of continuous employment of Standardization subjects at the time of interview along

53 with their current income. The years of continuous unemployment is noted for the unemployed group. Data relative to their income were not available. TABLE 4.3 YEARS OF EMPLOYMENT AND INCOME OF STANDARDIZATION SUBJECTS Mean Range Employed Years of continuous employment 14.3 2.8 to 39.7 Annual income for 1969 $6,357 $2,876 to $14,100 Unemployed Years of continuous unemployment 9.8 2.7 to 35.4 Annual income not available Tables B.2 to B.9 in Appendix B present demographic data comparing the two groups in the Standardization sample. The following statements summarize the major differences: 1. The unemployed group was somewhat older (Table B.2). 2. The employed group included more who were congenitally blind or lost vision early in life (Table B.3). 3. The employed had a higher level of education, more completing high school (Table B.4). 4. The employed included more who were able to read some print while the unemployed tended to be dependent upon braille and auditory media (Table B.5). 5. In general, the unemployed reported their travel to be limited

54 to their neighborhoods in contrast to the employed who claimed the ability to travel to new and distant places (Table B.6). 6. The degree of vision did not seem related to travel ability (Table B.6). 7. Separation and divorce are more prevalent among the unemployed (Table B.7). 8. The unemployed sample included a higher percentage of blacks (Table B.8). 9. The majority of both groups resided in urban areas (Table B.9). CROSS-VALIDATION SAMPLE In order to test the predictive validity of the CPI scales and items which may significantly discriminate between the employed and unemployed groups in the Standardization sample, an independent group of subjects was identified on whom the initial findings could be tested. This sample is designated as the Cross-Validation sample and includes subjects who were administered the CPI earlier and whose current vocational or educational status could be determined by follow-up contacts. Criteria and Selection of Subjects The Cross-Validation sample was drawn from a, group of 56 males who were administered the CPI as part of their evaluation in the Psychology Department of the Rehabilitation Institute, Detroit, between February 1, 1968, and December 1, 1969. The initial date of this period was selected because it was when the CPI came into use at the Institute.

55 The closing date was used as a cut-off so that a minimum of 8 months would elapse between evaluation and follow-up. This interim should provide sufficient opportunity for clients to commit themselves either to productivity, i.e., employment, technical training, college programs, or to continuing unproductivity. The current status of 56 males was determined by information furnished by Institute staff, personal contacts by the investigator, and the Detroit Office of Services for the Blind. Of this number 11 were employed, 15 were unemployed, 17 were college students; those ineligible for further consideration were two high school students, four who were severely physically disabled, and four who were deceased. A separate criterion group was suggested by Rehabilitation Institute staff to classify three employed subjects. Their employed status was regarded as being uniquely dependent upon extensive interagency cooperation and intense vocational counseling. Description of Cross-Validation Subjects Table 4.4 and Tables B.10to B.15 in Appendix B provide descriptive data on the Cross-Validation sample. The following statements summarize the characteristics of the group: 1. The employed and unemployed do not differ significantly in age (Table 4.4). 2. Proportionately more congenitally blind are in the employed group in constrast to a greater number of recently blinded in the

56 TABLE 4.4 CROSS-VALIDATION SUBJECTS CLASSIFIED ACCORDING TO AGE AND CURRENT VOCATIONAL STATUS Employed,olege Age Employed Unemployed Separately Students Class ified 17-21 21-25 26-30 31-40 41-50 51-61 Totals Mean 1 4 1 1 2 2 11 33.o90 2 4 5 1 4 1 15 32.266 0 1 2 0 0 0 TO TO 13 1 2 0 1 0 17 Mean difference: Employed vs. Unemployed = 0.824 t = 0.465 n.s. *A separate classification for these subjects was suggested by Rehabilitation Institute staff because their employment was regarded as dependent upon extensive interagency cooperation and intensive vocational counseling. unemployed group (Table B.10). 3. The employed attained a higher level of education, more completing high school (Table B.11). 4. The advantage of partial vision appears to be associated with employability to some extent in this sample (Table B.12). 5. The employed have a somewhat higher IQ than the unemployed (Table B.13). 6. More blacks are included in the unemployed sample (Table B.14). 7. Proportionately more single persons are in the employed group (Table B.15).

57 Instruments STANDARDIZATION SAMPLE Two instruments were used to gather data on the subjects in the Standardization sample: the California Psychological Inventory and an Interview Form. California Psychological Inventory The standardized form of the CPI was administered by tape recording and scored according to the standardized scale keys. Prior consent was obtained in writing from Consulting Psychologists Press to use the scales and the Inventory in this research. A listing of the CPI scales, their abbreviations, and scale classes may be found Appendix C. Interview Form An interview form was designed to gather personal and vocational data including geographical location, marital status, presence or absence of other disabilities, type and age of onset of visual impairment, academic education, vocational training, vocational history, current earnings and hours worked per week if employed, duration of continuous employment or unemployment to date, and measures of functional vision and mobility. A copy of the Interview form and set of Coding Instructions are included in Appendix C. The measure of functional vision used in the Scholl, Bauman, and Crissey (1969) study was utilized for this study. The measure of functional mobility was developed by

58 the investigator. Both measures rely on the subject's self-report. CROSS-VALIDATION SAMPLE Demographic data and CPI protocols for the Cross-Validation sample were available from the files of the Psychology Department, Rehabilitation Institute, Detroit. Retrieval and coding of these data were conducted by this investigator. Verification of the current employment status of each subject was made through the current files of the Comprehensive Services for the Blind at the Rehabilitation Institute, by inquiry at the Detroit Office of Services for the Blind, and/or a telephone contact to the subject by this investigator. No additional measures were employed with this group. Procedure ADMINISTERING THE INSTRUMENTS TO THE STANDARDIZATION SAMPLE This investigator contacted each subject to arrange for the interview and testing. Each session was conducted by this investigator. The length of the session ranged from one hour forty-five minutes to three hours depending upon the subject's varying concerns. The tape recording of the CPI was one hour and 17 minutes in length. Each subject was paid $5.00 for his cooperation. The interview and testing were accomplished between March and June, 1970.

59 California Psychological Inventory To provide a standardized presentation of the Inventory items to each person, the entire format of 480 items was tape recorded for presentation to each individual. A wooden rack of 480 IBM cards, keypunched from 1 to 480, was placed on the table. The subject was instructed to take the top card each time and place it in either the "true" or "false" box before him. Either box was identified with a brailled label and a, heavy ink notation. The "true" box on the subject's right was indicated for items with which he agreed, or which were true of him. The "false" box on his left was designated for items with which he disagreed or which he regarded as not true of him. He was further instructed that every tenth card was notched and that before every 10th item on the tape, he would hear: "This card should be notched." He was instructed to notify the investigator if the notched card and the taped statement did not correspond. This method of administration provided an efficient and comfortable presentation of the items. Modification of the ordinary conditions of testing to-an oral presentation was found by Bennett and Rudoff (1957) to yield equivalent results.* After the testing the investigator recorded the "true" and "false" card responses on standardized answer sheets. Separate coding of each card was used to reduce the possibility of recording errors. Each answer sheet was scored *Personal communication from Harrison G. Gough after the above procedure was described to him.

60 and verified according to the standardized scales. Interview Form Data on the Interview form were recorded after the CPI testing. The data were coded according to the Coding Instructions to facilitate later processing. ANALYSIS OF DATA FROM THE STANDARDIZATION SAMPLE After completion of the testing of the Standardization Sample all coded data from the Interview form and CPI protocols were recorded on coding sheets to facilitate keypunching for computer processing of the data. The entire recording process was verified by an independent checker. A total of six errors was identified and corrected. An error count of every 5th CPI scoring sheet was made by the independent checker to determine the error rate of summing the individual scale scores. No errors were found. The 18 variables were keypunched for each subject. Preliminary Evaluation Initial group differences were established by analysis of group means and their corresponding t ratios as computed on the Olivetti Programma 101-. A second test of the overall significance of between group differences was computed by means of the Mahalanobis D statistic as provided by the Discriminant Analysis-Two Groups (BMD04M) program (Dixon, 1968).

61 The Generalized Distance or D statistic searches for that linear combination of variates which will "maximize" the "between"-group differences relative to the "within"-group differences. This linear combination reduces the multiple variables to a univariate measurement and the overall, maximum discrimination between the two groups is computed on the single discriminant variate. The advantage of this generalized distance measurement between the groups is superior to a series of t tests since simple t tests between sample means lack efficiency in that they are unable to take into account covariances or interrelationships of variables and assess their relative power in determining sample differences (Anderson, 1966, p. 169). The formula for estimating the Mahalanobis distance between two populations may be found in Table D.1 of Appendix D. Development of the Employment Key An item analysis was made for each of the 480 items in the CPI. This was accomplished by making a 2x2 chi square contingency table for each item. The scored direction of each item, true or false, was weighted +1 or 0 in association with a subject's employed or unemployed status. The chi square values were then calculated on the Michigan Terminal System of The University of Michigan Computing Center on an IBM 560 Model 67-2. A 2x2 chi square program was developed and filed. Chi square values were processed from a terminal. Those items reaching a.01 level of significance were combined

62 into an experimental Employment Key scale and a plastic scoring sheet was made to correspond to the items. Each CPI protocol in the Standardization sample was scored with this Employment Key. Employment Key scale scores were then recorded on the coding sheets along with the 18 standardized CPI scale scores. This 19th variable was keypunched on each subject's IBM data card. This process was likewise verified for each subject to eliminate clerical errors. Stepwise Discriminant Analysis The method chosen for a more comprehensive evaluation of the 19 personality variables was that of linear discriminant function which is provided by the Stepwise Discriminant Analysis, BMD07M, computer program (Dixon, 1968). This multivariate prediction technique is used for purposes of calculating linear discriminant functions on variables, such as test scores, which presumably separate the members of two dichotomous groups. The efficiency of the discriminant functions is then tested by determining how effectively the subjects can be classified into their actual group membership on the basis of their test scores alone. In this model the intent is to determine how effectively the actual group membership of employed and unemployed subjects can be predicted on the basis of their scores obtained on the standardized CPI scales and the newly developed Employment Key scale. The variables for the discriminant functions are selected according to their respective capacities to account for the unexplained

variance at each step. The feature aspect to this stepwise entry of variables is the elimination of overlapping whereby two or more variables may in effect account for the same variance. For example, it is possible that more than one CPI scale accounts for some or all of the same variance represented by a certain personality trait dimension. At each step the discriminant analysis is repeated and the most discriminating variable is entered into the equations or set of discriminating functions. This technique accounts for the most rapid and effective reduction of the variance between two groups. The computer program generates linear discriminating functions in a stepwise fashion. Linear discriminating functions are depicted as Y = a+ X + a X +...+ a X where Y = scores on the discriminant 11 22 + k Xk function; X1, X2,...,Xk = raw scores on variables; and, al, a,...ak are applied to raw scores of each person in each group. Figure D.1 in Appendix D depicts how such linear functions discriminate two groups. After weights al, a2,...,ak are applied to each persons raw scores, the new scores can be projected on a line Y. The scores for each person on line Y can be located as frequency distributions. The location of scores on Y serves to condense the discriminatory information on the variables entered. The program performs the multiple discriminant analysis in a stepwise manner, entering or removing a variable from the set of discriminating variables according to specific criteria (Dixon, 1968, p. 214a). This provides selection of a variable, its optimal weighting, and a

constant which comprise the discriminant function or predictive equations. The fundamental criterion for entry of a variable into the equations is determined by the amount of variance explained by it. This is a special application of multiple regression analysis whereby optimal weights are obtained which maximize the F ratio of the variance between means on Y variance within groups on Y Details of this process are discussed by Nunnally (1967, p. 392). The computer program provides a classification matrix at each step indicating the correct "hits" as determined by the correspondence between the discrimination of the equation and the known criterion group membership. Finally, it establishes the probability of each person coming from his group on the basis of his profile scores. In order to provide a confirmation of the significant variables differentiating the employed and unemployed groups on the basis of their test data, a second analysis of the Standardization data was made. This was accomplished with a different mathematical model, Stepwise Regression Analysis, BMD2R, computer program (Dixon, 1968). This alternate statistical procedure identified the same discriminating variables with identical F values and followed an identical stepwise pattern in explaining the variance. APPLICATION OF FINDINGS TO THE CROSS-VALIDATION SAMPLE With predictive models the initial findings should be tested with

65 additional, independent samples to demonstrate potential for prediction beyond the original sample. To accomplish this, a follow-up of the Cross-Validation subjects relative to their current vocational-educational status was made in June, 1970. Their CPI protocols, which recorded their responses while involved in a previous vocational evaluation program, were classified according to their current vocational status. Each CPI answer sheet was re-scored with the Employment Key to obtain each subject's score on this experimental scale. This 19th variable was keypunched on data. cards along with the values for the subjects' 18 standardized scale values. This provided data parallel to that gathered from the Standardization subjects. The Stepwise Discriminant Analysis program provides a ready opportunity for admitting test data of new subjects for predictive classification according to the discriminant functions established on the initial group of subjects (Dixon, 1968, p. 214e). This facilitates an evaluation of the predictive efficacy of the discriminant functions in a Cross-Validation technique. It should be noted that if computer facilities are unavailable, the predictive classifications can be performed through manual computations by inserting the raw scale scores of new subjects into the discriminant function equations established on the original data. The results obtained from these procedures are presented in the following chapter.

CHAPTER V RESULTS Results will be reported in two major sections, the Standardization sample and the Cross-Validation sample. The Standardization sample includes a report of the individual scale differences and a generalized test of significance between the employed and unemployed groups. Secondly, it includes an Employment Key empirically developed by an item analysis. Finally, it reports a stepwise discriminant function analysis for the purpose of developing a formula for prediction. The Cross-Validation sample section presents the classificatory power or predictive results obtained when the discriminant functions, established on the Standardization sample, were applied to the CPI scores of a separate group of subjects who were administered the CPI previously and whose current employment status was established on follow-up. Standardization Sample EVALUATION Table 5.1 compares the employed and unemployed groups on the 18 standardized scales. Significant t ratios at the.01 level of confidence were found on nine scales: Dominance, Capacity for status, Social presence, Well-being, Tolerance, Communality, Achievement via 66

67 TABLE 5.1 EMPLOYED AND UNEMPLOYED ON THE SCALES OF THE CPI COMPARISON OF Employed Unemployed Dif CPI Scales N = 51 N=50 M Ra N -- 50M Ratio M SD M SD Dominance 29.4 6.0 25.7 6.4 3.7 2.90* Capacity for Status 18.8 4.3 15.6 4.6 3.2 3.60* Sociability 24.8 5.1 22.8 5.5 2.0 1.86* Social Presence 33.5 6.0 30.2 5.9 3. 2.74* Self-acceptance 20.5 3.8 18.8 4.1 1.7 2.17 Sense of Well Being 36.4 4.9 33.4 6.0 3.4 3505* Responsibility 30.4 4.9 28.1 5.0 2.3 2.31 Socialization 35.9 6.2 33.4 5.6 2.5 2.11 Self-control 30.6 6.5 28.3 7.1 2.3 1.63 Tolerance 19.6 6.0 16.4 5.5 3.2 2.78* Good Impression 20.6 5.3 19.2 5.8 1.4 1.28 Communality 26.6 1.6 25.0 2.7 1.6 3.57* Achievement via Conformance 27.3 4.8 24.5 4.6 2.8 3.02* Achievement via Independence 16.6 4.1 13.3 4.8 3.5 3.74* Intellectual Efficiency 35.7 6.0 32.5 3.6 3.2 2.56 Psychological Mindedness 11.0 2.9 9.2 2.4 1.8 3.26* Flexibility 6.5 3.7 5.3 3.6 1.2 1.71 Femininity 18.2 4.0 18.4 3.7 -0.2 -0.34 *P <.01 conformance, Achievement via independence, and Psychological-mindedness. Figure 5.1 reports these data in graphic form. 2 The Mahalanobis D statistic was used to test the overall distance or generalized difference between the two groups dichotomized on employment versus unemployment. The Discriminant Analysis-Two Groups 2 (BMD04M) program established a D value of 6.6403. To test the hypothesis specifying no difference in mean values between groups on the 18

PROFILE SHEET FOR THE 4wza Aoym JAr: MALE Notes: Do Cs Sy Sp Sa Wb Re So Sc To Gi Cm Ac Ai le Py Fx Fe 1- -35 90 -an -55 --45 -30 Employed Group (N=51) - Unemployed Group (N=50) 70 - -30 -40 -30 -;50 - - -:0 _: 0-3 - - - - 45 -- - -40 - - - - -: o — 5 - ^r35- -25 -o -2535 -25 -45 30-35 -- 2 _: -:30 ^~~405 - 15 30 -45 --."45 -45 - - -4 0.- _. _ ~o T~ 50 -40 - - - - - — 35 - -25 - -30 4: - -40 - -40 -35 - 30 - 25 -20 -35: - - - - 40 -10 -- 5s Z-35 1,1 _:* X -^ '., _ —.~-= — ^-: —^.^^ -o_-/'/~-~ ',,.,_~:./%-__ /, -% - ~. _: ~-3.^\ _^ ^.^ //- '~ - 2 - /-,o~-/'-,5 ~~-,, -,... -, — 35 ) ----^ — \ o, -25 ~-/",.. — ~ / - 30 - v - r — 15 _:0 -15 _ -70 -60 -50 -40 (A 1 0 Q - 15 - -- -. - - /:: - -- - - 5-' -25. -2 -Io -5 -25 -30 — 20 - - - - - - -20 -25 -20 - - - - 2 Is 0 2 - - - - - - 0 - -2 - 0 -:: -..-2 - -15 - - - - - - - - 0 -* - _ _ _ ' 0 _ __ _ _ _ _ _ _ _ _ _ _ _ __ _ _ - __ _ 30 - - - _ _ 5 2 0- - - _ _o_ - 0 5:1 - - - - - _ -10 - -10 - 0 - - - -20 - - -- -2 - - - - 0 0 - -- 201 -10 0 I0 -20 - 0 - - 10 - - - - -,: - _5 10 - _ - 15 - 10 — 0 5 --- - 30 -20 - 10 - 0 Do Cs Sy Sp Sa Wb Re So Sc To Gi * * * - * Cm Ac Ai le Py Fx * * * * Fe *t test significant at.01 level Figure 5.1. Comparison of employed and unemployed Standardization subjects on the CPI.

69 CPI scale scores, this value was converted to an F value of 7.219 according to the formula (Rao, p. 247) N1 N2 (N + N2 -p-l) 1 2 1 2 P(N1 + N2)(N1 + N2 - 2) With 18 and 82 degrees of freedom, according to the formula p and (N + N - l-p), this value is sufficient to reject the null hypothesis of no difference in mean values for the two groups at the.001 level of confidence. The formula for estimating the Mahalanobis distance between two populations may be found in Table D.1 of Appendix D. The importance of this statistic lies in its ability to reduce a number of variables to a single measure of discrimination. The high degree of confidence attained by this statistic lends support to the contention that visually impaired persons who differ in respect to employment versus unemployment also tend to describe themselves differently on the CPI. Table D.2 furnishes a within groups correlation matrix to provide comparisons of the scale values of the Standardization subjects. EMPLOYMENT KEY An item analysis was performed on each of the 480 CPI statements of the 101 Standardization subjects. This provided a measure of association between the employed versus the unemployed groups and the elerespective frequencies of their directions of response to each item, true or false. Twenty-one items reached significance at the.01 level of

70 confidence. Table D.3 in Appendix D furnishes a listing of these items, corresponding chi square values, direction of scoring, and the location of these items in the standardized CPI scales. Another 51 items reached a significance at the.05 level of significance but were not considered further in the analysis of the data. The 21 items of.01 level of confidence were next combined into a separate scale designated as the Employment Key (Ek). Each CPI answer sheet was rescored with this key. A preliminary indication of the discriminating power of this scale between the employed and unemployed groups is suggested by the t ratio of 9.21 between mean differences which is significant at the.001 level of confidence. Table 5.2 provides a comparison of the distributions of raw scores obtained by the employed and unemployed groups on the Employment key. A suggested cutting score for the Employment key scale and its discriminatory power as an independent scale will be presented later in this chapter in relation to both Standardization and Cross-Validation subjects. STEPWISE DISCRIMINANT ANALYSIS' This analysis of the Standardization data was made with the 19 predictor variables, the 18 standardized CPI scales and the Employment Key scale scores. The corresponding F values with 1 and 99 degrees of freedom revealed significant differences between the criterion groups on 14 scales. The Employment key, Achievement via independence, Capacity for status, and Communality scales differentiated the employed

71 TABLE 5.2 DISTRIBUTION OF SCORES ON EMPLOYMENT KEY SCALE FOR EMPLOYED AND UNEMPLOYED GROUPS IN THE STANDARDIZAT ION SAMPLE t Score 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 N M SD M1 - M2 Employed Unemployed..... -................ L~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1 1 5 9 8 4 9 3 6 2 2 1 51 14.84 2.56 = 5.68; t = 1 2 1 8 2 5 4 4 7 4 5 2 1 2 2 50 9.16 3.55 9.21, P <.001 from the unemployed groups at the.001 level of confidence. The Psychological-mindedness, Well-being, Achievement via conformity, Dominance, Tolerance, and Social presence scales differentiated the employed groups with.01 level of significance. The Intellectual efficiency, Responsibility, Self-acceptance, and Socialization scales

72 reached significance at the.05 level of significance. These results are presented in Table 5.3. TABLE 5.3 F VALUES FOR EACH SCALE PRIOR TO ENTRY INTO THE DISCRIMINANT FUNCTIONS CPI Scales Dominance Capacity for Status Sociability Social Presence Self-acceptance Well Being Responsibility Socialization Self-control Tolerance Good Impression Communality Achievement via Conformance Achievement via Independence Intellectual Efficiency Psychological-mindedness Flexibility Femininity Employment Key F Value 8.428** 12.951*** 3.484 7.503** 4.742* 9. 205* 5.358* 4.457* 2.661 7.768** 1.651 12. 731** 9.167** 13. 980*** 6.587* 10.675** 2.852 0.116 84.139*** N = 101 * <.05 ** <.01 *** <.001 Discriminant Functions In order to establish the most efficient pattern of scales for differentiating the employed from the unemployed, the Stepwise Discrim inant Analysis program repeated numerous analyses of the data in a

73 stepwise manner. At each step that variable was entered into the equation which had the largest F value, gave the greatest decrease in the ratio of within to total generalized variances, and had the highest multiple correlation with the groups when partialed on the previously entered variables. This method identified at each step which variable accounted for the most variance in combination with the previously entered variables, and evaluated the capacity of the discriminant function at each step to classify subjects according to their dichotomous groups of origin. Table 5.4 provides a comparison of the F values established during the first five steps of the stepwise discriminant analysis. After that point the rapid reduction of F values of those variables not yet entered into the equation accounted for little, if any, of the unexplained variance. Likewise, additional steps increase the possibility of chance covariance. It must be stressed that the maximum efficiency in classifying subjects was found at the second step with the entrance to the equation of the Employment Key and Tolerance scales. However, because of sample limitations and the exploratory nature of this study, consideration will also be given to the five-variable equations. Table 5.4 furnishes a comparison of the initial F values prior to the stepwise computations of the discriminant functions. At each step the asterisk indicates the point of entry of at variable into the equations which accounts for the most variance when partialed on the variable(s) already entered into the equation. After each step the

74 TABLE STEPWISE ENTRY 5.4 OF VARIABLES WITH CORRESPONDING F VALUES AT EACH STEP Variable Initial Step 1 Step 2 Step 3 Step 4 Step 5 Entered F Ek To So Ie Ac Do 8.428 0.929 1.137 0.917 o.168 1.406 Cs 12.951 3.964 0.288 0.195 o.o46 0.028 Sy 3.484 3.277 1.670 1.276 0.148 1.031 Sp 7.503 8.547 3.930 2.076 0.832 0.708 Sa 4.742 l.06 1.6005 0.559 0.031 o.492 Wb 9.205 3.869 0.076 1.397 0.924 1.618 Re 5.358 4.594 0.015 1.006 0.451 1.438 So 4.457 0.539 4.434 *4.434 Sc 2.661 0.642 1.261 0.036 o.oo 0.075 To 7.768 23.970 *23.970 Gi 1 651 o.646 0.259 0.011 0.013 o.417 Cm 12.731 1.075 0.114 0.007 0.209 0.053 Ac 9.167 o. 006 2.024 0.607 2.641 *2.641 Ai 13.980 8.666 0.188 o.008 0.026 0.022 Ie 6.587 14.559 2.080 2.176 *2.176 - Py 10.675 1.977 0.020 0.003 0.092 0.000 Fx 2.852 2.795 1.115 0.110 0.285 0.075 Fe. 0.116 2.617 1.889 0.696 0.086 0.063 Ek 84.139 *84.159....... _.....: decrease of F values can be noted. Table 5.5 should be consulted in conjunction with Table 5.4. Table 5.5 specifies the discriminant func tions and constants computed at each step. These values provide vari able weightings in equation for use with each subject's raw scores. These values may be thought of as regression weights. This model provides discriminant functions for each dichotomous group at each step. A subject's group classification is determined by whichever equation in combination with his raw scores produces the larger value. Gough (1962, p. 5538) describes the discriminant function analysis as an

75 TABLE 5.5 DISCRIMINANT FUNCTIONS AT EACH STEP VARIABLE WEIGHTS, AND CONSTANTS.Functions Variable(s) Entered Employed Unemployed Group Group Step 1 Employment Key 1.5531 0.945 Constant -11.366 - 4.328 Step 2 Tolerance - O.116 0.243 Employment Key 1.698 0.596 Constant -11.459 - 4.734 Step 3 Socialization 1.050 0.936 Tolerance - o.636 - 0.219 Employment Key 2.197 1.040 Constant -28.914 -18.599 Step 5 Intellectual Efficiency 0.935 1.051 Socialization 1.027 0.910 Tolerance - 1.197 0.849 Employment Key 1.685. 0.465 Constant -35.903 -27.425 Step 5 Achievement via Conformance 0.436 0.292 Socialization 0.928 0.844 Tolerance - 1.184 - 0.840 Intellectual Efficiency 0.730 0.913 Employment Key 1.762 0.516 Constant - 37129 -27.974 alternative to the multiple-regression equation, so developed for use when the prediction problem is to place a subject in one of the two or more discrete or unordered classes. From data reported in Table 5.5 the following pairs of equations

76 may be written for use with raw CPI scale scores. Two-variable equations: Employed = -11.459 +1.698Ek -.116To Unemployed = - 4.734 +.596Ek +.243To Five-variable equations: Employed = -37.129 +1.762Ek -1.184To +.928So +.730Ie +.436Ac Unemployed = -27.974 +.516Ek -.840To +.844So +.913Ie +.292Ac The discriminant functions generated at the second step provided a two-variable equation which classified correctly 94% of the employed subjects. This represents a total of 48 "hits" out of a possible 51. The equations were slightly less efficient with the unemployed. Eightytwo percent were correctly classified from this group with 41 "hits" from a possible 50. Table 5.6 presents the efficacy of the discriminant functions with the two-variable equation. TABLE 5.6 CORRECT CLASSIFICATION OF SUBJECTS BY THE DISCRIMINANT FUNCTIONS GENERATED WITH THE TWO-VARIABLE EQUATION Actual Employment Classification Employment Employed Unemployed Status N % N, Employed 48 94 3 6 Unemployed 9 18 41 82

77 Use of the five-variable equations produced identical discriminating power (940) with the employed. With the unemployed the efficacy decreased slightly with forty "hits" observed for a 20 loss. However, permitting the program to include five variables in the equations explains further variance between the employed and unemployed groups, and simultaneously identifies additional personality traits which distinguish the two groups. Cross-Validation Sample A Cross-Validation technique is useful to determine whether the predictive findings of an empirical study are useful beyond the subjects on whom such findings were initially developed. To provide such a test in this model, the earlier CPI test results and current vocational status of an independent group of blind subjects were used. Specifically, this Cross-Validation procedure is addressed to the question of whether the Employment Key scales and the discriminant functions presently developed for use with the CPI could have predicted the current employment status of subjects who were administered the CPI previously during their vocational evaluation program. In addition, the use of Cross-Validation subjects underscores an important aspect to this predictive model in that their te st scores were obtained while they were unemployed and receiving concentrated rehabilitation services. Consequently, this precludes the possibility

78 of their test scores being influenced by employment experiences which might contribute to test score differences between the employed and unemployed. A comparison of means and standard deviations of the CrossValidation subjects on the 18 CPI scales and the experimental Employment Key scale may be found in Tables D.4 and D.5 in Appendix D. Figure D.2 provides a graphic comparison between the scale means of the working and nonworking subjects in the Cross-Validation sample. Table 5.7 presents the distribution of scores on the Employment key scale for the Cross-Validation subjects. The Employment Key discriminated between the employed and unemployed groups at the.05 level of confidence. It should be pointed out that the mean value of the college student group on the Employment Key closely resembles those obtained by the employed groups of the Standardization and CrossValidation samples. This suggests that the college students tend to describe themselves in a manner similar to that of the employed subjects in this study. Table 5.8 presents the predictive efficiency of the Employment Key when it is used as an independent scale. A cutting point at 12 is indicated on the table. The accuracy of its discriminating power to identify the employed and unemployed in the Standardization and Cross-Validation samples is presented. The two-variable equations predicted employment among the working Cross-Validation subjects with 82% accuracy by identifying nine of the eleven working subjects. The effectiveness of these equations in

79 TABLE 5.7 DISTRIBUTION OF SCORES ON THE EMPLOYMENT KEY SCALE FOR THE CROSS-VALIDATION SUBJECTS Employed.. College Score Employed Unemployed Extensive Student Students Efforts 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 1 1 1 1 1 2 2 1 2 1 1 1 2 2 4 1 1 2 1 1 1 2 4 3 5 2 1 1 1 1 1 N 11 15 3 17 M 13.363 10.733 7.0 13.294 Employed vs. Unemployed: M1 - = 2.630; t = 2.172, P <.05 predicting unemployment among the nonworking subjects was 80O as twelve of the fifteen unemployed were correctly identified. Table 5.9 presents the predictive accuracy with the two-variable equations. As found with the Standardization subjects the increase of variable entries to five did not increase the predictive efficiency over the two

80 TABLE 5.8 THE DISCRIMINATING POWER OF THE EMPLOYMENT KEY SCALE WITH A SUGGESTED CUTTING SCORE AT 12 'Employed. Unemployed Score Standardiza- Cross-Valida- Standardiza- Cross-Validation Subjects tion Subjects tion Subjects tion Subjects 21 20 19 18 17 16 15 13 12 11 10 9 8 7 6 5 4 3 1 5 9 8 4 9 3 6 2 2 1 1 1 1 1 2 2 1 1 1 2 1 8 2 5 4 4 7 4 5 2 2 1 _ _ 2 4 3 2 2 2 1 N Correct Classification: 51 50 15 46 (90o) 9 (82%) 36 (72%) 11 (73%) variable equation. Its predictions with the nonworking group produced identical efficiency but was slightly less efficient with the working group, classifying eight of eleven subjects correctly. The additional criterion group, which was identified as comprising the three subjects who were employed but only "after extensive

81 TABLE 5.9 PREDICTIVE ACCURACY OF THE TWO-VARIABLE EQUATION WITH THE CROSS-VALIDATION SUBJECTS Predicted Ac tua l Actual Employment Classification Employment Employment Employed Unemployed Status N % N g N N 1 Employed 9 82 5 20 Unemployed 2 18 12 80 counseling and unique interagency cooperation," was analyzed in the following manner. One subject was classified correctly as employed, while the remaining two were incorrectly classified as unemployed. This kind of error in prediction suggests both the importance of considering all other pertinent information about the client as well as the value of well planned case services, despite a poor prognosis based solely on test scores. For exploratory purposes the CPI test protocols of those 17 CrossValidation subjects presently attending college were analyzed with the two-variable discriminant functions. Twelve of the 17 subjects were classified statistically as "employed" which suggests that their Employment Key and Tolerance scale responses appear to resemble those of the employed Standardization group. Figure D.3 in Appendix D presents the CPI scale means for the college students, including the Employment Key. It should be observed that when classification of college students was made only with the Employment Key, one additional student was classified as "employed."

82 Tables D.6 and D.7 in Appendix D present summary statistics for each Cross-Validation subject. These tables list each subject according to his a priori group membership, i.e., actually working or nonworking. It furnishes the predictive classification according to the discriminant functions, the squared Mahalanobis distance from the predicted group, and, in parentheses, the associated probability that the observation came from that group. Figure D.4 presents the grand means for all groups on the CPI scales, including the Employment Key. In summary, the evidence of significant scale differences and the generalized distance between the two groups is sufficient to reject the null hypothesis that the employed and unemployed subjects of the Standardization sample would be undifferentiated on the basis of their CPI scale scores. Further, the individual and collective significance of the Employment Key scale items is sufficient to reject the second null hypothesis that the employed and unemployed subjects of the Standardization sample would undifferentiated by their respective of item responses.

CHAPTER VI DISCUSSION The findings of this study, while preliminary, suggest that the CPI shows promise as a standardized personality inventory as one source of information for evaluating the vocational readiness of visually impaired male adults. The generalized differences between the employed and unemployed of both samples reflect their dissimililarity on personality traits as measured by the CPI. The employed subjects of this study scored systematically higher than the unemployed on all CPI scales except the Femininity scale. This finding suggests that this scale may measure aspects of behavior which may not be related to personality traits associated with employability. The nine CPI scales which significantly differentiated the employed from the unemployed suggest that the CPI may be identifying important differences in personality traits. Certain items are particularly strong discriminators between the employed and unemployed groups. When these items are combined into a single scale their collective differentiating power is surprisingly strong, both as a scale accounting for a disproportionately large amount of the variance in relation to the standardized CPI scales, and as an independent screening measure with a designated cutoff point. Similar uses of cutoff points with individual CPI scales have been 83

effectively used with discrete populations. Examples are those dealing with the severity of asocialization as measured by the Socialization scale (Gough, 1965b) and the presence of anxiety (Levanthal, 1966; 1968). The identification of the Employment Key items and discriminant functions in this model provide a convenient method for sharing these findings with other investigators who may want to study their validity and usefulness in other evaluation-type settings. The application of the Employment Key items and discriminant functions to additional CPI protocols should provide opportunities for making individual predictions even when computer facilities are not readily available. An interesting aspect to the identification of the optimally discriminating variables, which distinguish the employed from the unemployed, is the opportunity to advance personality trait sketches which may characterize the dichotomous groups. This seems warranted because the CPI scales have been intensively researched and adjectival descriptions provided for each scale and groups of scales (Gough, 1957). The interpretation of profiles on the strength of individual scales, and diagnostic implications on the basis of patterns and combinations of scales in interaction are discussed by Heilbrun, Daniel, Goodstein, Stephenson, and Crites (1962). Apart from the discriminatory power of the Employment Key and its usefulness as an independent scale for prediction, some behavioral traits may be provisionally suggested by examination of the scale

85 items. (See Employment Key items, Table D.3 in Appendix D.) The following description must be considered speculative until additional research can relate the scale to independent trait criteria. Tentatively, the employed seem to have stronger feelings of selfworth, confidence and adequacy (items 154, 311, 338, and 452) as well as an absence of excessive fear (items 79 and 452). These positive selfesteem characteristics of the employed are accompanied by a posture of flexibility and capacity for adaptive behavior in thinking and social attitudes (items 47, 257, and 377) along with an open-minded, nonjudgmental approach to another's point of view (items 41 and 128). The employed further seem to possess a certain capacity for decision making (items 13 and 383) which may underlie a capacity for commitment to goal-directed activity. The association of higher self-esteem characteristics among the employed blind is also reported by Scholl, Bauman, and Crissey (1969) who found work-related criteria to be associated with more favorable self-evaluations. Schwartz, Dennerll, and Lin (1968) found employment among epileptics associated with CPI scales which are felt to reflect social poise and self-assurance. Bauman (1954) reports that the subjects of her employed and generally well adjusted group were better integrated into their social milieu. The Tolerance scale was selected by the stepwise analysis as the most discriminating variable after the Employment Key. This scale

86 provides additional opportunity for sketching provisional personality differences between the employed and unemployed in this study. Gough (1968c) reports that the Tolerance scale was designed to be a subtle or indirect measure of the authoritarian personality syndrome as assessed by the California F scale (Adorno, Frenkel-Brunswik, Levinson, and Sanford, 1950). The Tolerance scale items were selected on the basis of their correlation with this scale and the California E (ethnocentrism) scale. Initially it should be noted that the subjects of this study scored consistently below the mean for the national reference group (Gough, 1957). This may suggest that the visually impaired of this study tend to be slightly more rigid, less permissive, and less accepting of others and their personal beliefs than people in general. Since the unemployed in this study scored systematically lower on the Tolerance scale than did the employed, there appears to be evidence which suggests that the employed persons of this study tend to be more permissive, accepting and open-minded, and less authoritarian and self-centered than their unemployed counterparts. When the computer program continued to a five-step analysis of the data, the Socialization, Intellectual efficiency, and Achievement via conformance variables were joined to the variables of the two-step equations. The predictive efficiency was not improved with the entry of the additional variables and one unemployed subject in the Standardization group and one working subject in the Cross-Validation group

87 were misclassified. This slight decrease in classification accuracy may be due to chance covariance. The advantage in identifying additional significant variables provides a more complete accounting for the variance between the groups and offers additional personality trait information which may distinguish the employed from the unemployed on behavioral characteristics. The entry of the Socialization scale into the set of discriminating variables suggests that the employed persons resemble those who tend toward a greater degree of social maturity, are more able to accept rules and authority, and are not given to rash or thoughtless behavior. The emergence of the Achievement via conformance scale is noteworthy in that it stresses the cooperative aspects of concern for others while striving for personal achievement. Persons scoring high on this scale are felt to work well in situations having clear cut rules and regulations. The Intellectual efficiency scale, which is selected as the next most discriminating variable, is regarded as a measure of personal effectiveness, essentially intellectual, which facilitates task accomplishment in an intelligent, resourceful manner. In summary, it seems that the employed subjects in this research possess more of the above characteristics, and to a greater degree, than the unemployed subjects. The above discussion attempts to deal with those personality characteristics presumed to differentiate the employed from the un employed in this study. It is important to note that no attempt was

88 made to contrast these criterion groups on a well adjusted-poorly adjusted continuum. In her study Bauman (1954) distinguishes three groups: those employed and generally well adjusted; those not successfull in employment but generally well adjusted; and those not successful in employment and generally poorly adjusted. However, inspection of the CPI protocols and interview impressions gained during this study suggest that employability may not necessarily be synonomous with good adjustment. This may suggest that Bauman's first group should be subdivided to include those competitively employed but not "generally well adjusted." Such a category would seem to be apropos to the relationship between personality functioning and employment in the general population. To postulate such a category might contribute to a more effective evaluation of both personality data and interview material in vocational assessment and planning with the blind. It is suggested that future research be addressed to the question of whether employability among the blind is related to some very discrete personality factors which may be quite separate from those felt to reflect "good adjustment" among the blind. The practical implications of these findings should be interpreted with caution and understood in the context of this research and its inherent limitations. The use of these findings with clients should be made with the circumspection required for the use of all test results. It must be noted that not every subject in this study was correctly classified according to his test results. Furthermore,

89 there are other considerations which may bear upon a blind person obtaining employment. These are independent of his personal functioning, such as fluctuating opportunities in the job market, local employer bias, variable agency policies, the availability of training resources, and the deviating policies of early industrial retirement and the structure of Social Security Disability which may make job retraining and low level employment less than feasible. Another consideration affecting employability may be the degree of vision possessed by an individual. Partial vision may increase one's capacity to perform adequately on a particular job. This point may find support in the fact that the employed of both samples in this study reported themselves as having usable vision. The findings of this study present interesting research opportunities for those charged with the responsibility of developing manpower among the visually impaired. The current findings suggest that personality traits are important variables bearing on employability among the blind. Further, these results indicate that the CPI shows promise for measuring those traits related to the dimension of employability. The lack of systematic research with personality measures for the blind (Wachs, 1966), and the apparent absence of studies investigating the associations between personality traits and performance criteria may be contributing by default to underemployment and unemployment among the blind. Because the CPI was developed as a general-purpose inventory for

90 the prediction of a variety of personality criteria (Hase and Goldberg, 1967), its use provides information about an individual on a number of behavioral dimensions. Because of its diversified research applications, some of which were discussed above, it seems to have particular relevance to such goal directed behavior as the prediction of success in job training and in maintaining employment. Workers who are involved with the determination of job readiness and employment potential in an evaluation-oriented setting do not simply "sort out" the most promising clients for job placement. For many subjects employability lies far beyond the evaluation procedure. Responsible manpower development calls for identification of the individual's needs for personal adjustment training, his current capacity to accept work responsibilities, or his readiness for vocational training. The CPI should be considered a useful instrument for plumbing out information relative to these questions as well as for providing diagnostic implications for on-going personal counseling when indicated. A precautionary correllary should follow a discussion of a predictive model such as this. Granting for the moment the continued validity of these findings, two outcomes might occur. When a subject responds very favorably to the CPI scales and items, counseling and agency personnel could relax efforts, assuming in error that such a person will achieve vocational success with minimal assistance. At the other extreme a client may depict himself very poorly on the CPI, thereby stimulating the destructive assumption that whatever counseling

91 and agency assistance is offered will be fruitless. The possibility of these dangers demonstrate the importance of evaluating all other information such as is available by interview, case history, and other test data (Elonen and Cain, 1964). Excessive reliance upon any one source of data is precarious and potentially detrimental to the client. A case in point is the reminder of the erroneous classification by this model of two of three employed Cross-Validation subjects who were designated by Rehabilitation Institute personnel as unique in that they were employed but only through extensive counseling and interagency cooperation. If these three subjects maintain successful employment, their success will underscore the inherent weakness of relying solely on test results. The example of these three employed subjects bears on another point in that their separate classification represents the employment of the "unemployable" and illustrates something of the distance which can be bridged by concentrated, in-depth services in manpower development when prognostic indications are far from favorable. In general the findings of this research suggest that the CPI can contribute meaningful data to that body of psychological, social, vocational, and educational information required for the thorough assessment of vocational potential in manpower programs with the visually impaired. In particular, the findings of this study suggest that workers with the blind should strive for even better placement of clients with "higher employability scores" than they would have in

92 the past without such supportive test data. Workers with the blind should also designate those with "lower employability scores" as candidates for extensive interagency cooperation and intensive vocational counseling. Limitations Of The Study The findings presented in this research should be considered in relation to certain limitations. Initially, the small sample sizes in both the Standardization and Cross-Validation phases emphasize the provisional value of the Employment Key and discriminant functions. In addition, the subjects were limited to one geographical area. Their identification and cooperation with testing were of such complications that a randomized sampling procedure was impossible. Consequently, there are unanswered questions relative to possible differences between those who were identified and who did in fact cooperate as opposed to those who may not have been identified or who did not cooperate. While a limitation may lie in the arbitrary criteria used for the admission of employed and unemployed subjects to the Standardization sample, this approach did seem to offer a parsimonious method for determining whether the employed blind differ from the unemployed on personality trait characteristics. Perhaps the most conspicuous bias occurred when unanticipated

93 difficulty was encountered in locating unemployed persons for the Standardization sample. This limitation appears to have implications for further research of this kind as well as for private agency resources who may want to offer their services by personal contact to the chronically unemployed. The approach of this study suggests there may be a "lost" blind population. This is consistent with Scott's (1969) criticism of programs of agencies for the blind. The State Services for the Blind reported that their contact letters to locate former, unemployed clients, were returned with postal notification of having no forwarding address. When this investigator followed up a list of names from a private agency, phone contacts indicated that the majority of subjects had moved. This suggests the presence of a highly transient population among the blind. In the absence of a central registry of blind persons and the current legal restrictions imposed upon the identification of the visually impaired who are receiving public assistance, one concludes that there is an unemployed blind population which is beyond the reach of certain private agencies and research efforts. This inability to locate the unemployed subjects for the Standardization group was circumvented in part by the assistance of numerous blind subjects who enthusiastically volunteered their assistance to identify and/or contact potential subjects in the community. Despite this invaluable assistance, it is presumed that "isolates," the

94 transient, chronically unemployed who do not identify with organizations for the blind or with other blind subjects are in fact not represented in this study. Such persons are deprived of employment counseling services unless they themselves request help from an agency of their choice. The failure to identify relatively young, chronically unemployed males necessitated the inclusion of a number of older males who seemed to be prematurely retired and supported by varying forms of disability insurance benefits, including Social Security Disability. An examination of the computer print-out indicates that these early retirees contributed notably to the inaccuracy of the classification of the unemployed Standardization group, 41 of 50 correct. In contrast the model was able to classify correctly 48 of 51 employed Standardization subjects. This suggests that had the test scores of these "early retirees" been omitted, a stronger test score dichotomy between the two groups may have been obtained. This bias in sample selection is to some extent counterbalanced by the use of the Cross-Validation subjects. Implications For Further Research The findings of this study suggest further research both in relation to additional analyses of the data collected in this study as well as for additional research designs employing the CPI. The

95 following arise specifically from this study: 1. A continuing validation of these findings should be made through periodic follow-up procedures of all blind clients who receive the CPI at the Rehabilitation Institute, Detroit, Michigan. 2. Additional analyses of the current data should be made to determine whether such variables as the degree of vision, age of onset, and the type of onset of visual impairment are associated with certain CPI scales or the employed versus unemployed dichotomy. 3. The current data should be analyzed to determine whether any CPI variables are associated with certain occupations represented in this study. 4. A visual scanning of the CPI items suggests that six items may have slightly different implications for visually impaired persons. These statements should be examined in an item analysis as well as considered in relation to their parent scales to determine whether any should be eliminated from presentation to the visually impaired. The following are suggestions.for further research: 1. Impressions gained from this study suggest that the CPI is a useful instrument for developing the manpower potential of the visually impaired. It appears to be a versatile, widely researched, and conveniently administered instrument. In view of the dearth of systematic personality assessment with the blind, the CPI could be employed to study personality characteristics of the visually impaired and other handicapped groups.

96 2. The CPI should also be considered for the prediction of performance criteria, such as level of employment or academic achievement, among the visually impaired and other handicapped groups. 3. The Employment Key scale and discriminant functions developed in this study can be readily used on an individual basis whenever the CPI is part of a psychological evaluation. The confidence with which predictions may be made with these techniques will depend upon the results of such efforts in various settings. Further use of the Employment Key with subjects in other settings may increase confidence in using it. 4. Since the CPI and the Employment Key items are not uniquely related to the visually impaired, but presumably applicable to some aspect of the employment-unemployment dimension, the application of these findings should be considered for use with other handicapped populations. For example, a request has already been made for the use of the Employment Key scale by another investigator in studies of employability of subjects with seizures.

CHAPTER VII SUMMARY AND CONCLUSIONS Unemployment and underemployment among the blind have been attributed to a variety of factors in addition to the visual impairment itself. This study is concerned with relationships between the personality characteristics of the visually impaired and their suitability for employment. Personality measurement with the blind has not been extensively explored nor systematically related to prediction of performance criteria. The California Psychological Inventory was chosen to determine whether a standardized "objective" personality measure could identify differences between groups of employed and unemployed blind, and, if so, whether such differences could be used to predict employability among the blind. This study consisted of two phases, Standardization and Crossvalidation. For the Standardization phase 51 visually impaired males were identified who were employed full time for at least 2.8 years in occupations not requiring a college degree. Visually impaired persons whose employment was dependent on a college degree were excluded in order to focus on groups with relatively homogeneous educational backgrounds. The unemployed group consisted of 50 visually impaired males who had not worked for at least 2.7 years. All subjects resided in Southeastern Michigan. They were located through

98 agencies serving the blind and by the subjects themselves. The CPI was administered to the Standardization subjects by means of a tape recorded presentation of the items for card sort according to their agreement or nonagreement with each item. An Interview instrument was developed to obtain demographic data to compare the employed and unemployed on selected characteristics. This instrument was administered orally to each subject. An initial analysis of CPI data indicated that the unemployed scored significantly lower than the employed on nine scales. When the scale score differences were analyzed by a Discriminant Analysis for two groups, a Mahalanobis D or generalized difference between the groups was significant at the.001 level of confidence suggesting the dissimilarity of personality traits between the employed and unemployed as measured by the CPI. An item analysis indicated that 21 CPI statements discriminated between the employed and unemployed groups at the.01 level of significance. These items were combined into a scale and designated the Employment Key. Mean differences between the employed and unemployed groups on the Employment Key scale differed with. 001 significance. With a suggested cutting score at 12, the Employment Key correctly identified 90fo of the employed and 72% of the unemployed Standardization subjects. A Stepwise Discriminant Analysis was selected for the purpose of identifying an optimal pattern of predictor variables for

99 discriminating the employed from the unemployed. Two and five variable equations were developed. The most efficient equations used the Employment Key and Tolerance scales with the following coefficients: Employed = -11.459 +1. 698Ek -. ll6To; Unemployed = -4.734 +. 596Ek +.243To. These discriminant functions are for use with raw scores; whichever equation produces the larger value determines the subject's predicted classification, employed or unemployed. These equations classified correctly 94% of the employed and 82% of the unemployed in the Standardization sample. The Cross-validation phase consisted of efforts to test the usefulness of these findings beyond the original sample. Subjects for the Cross-Validation technique consisted of those visually impaired males who previously were administered the CPI as part of their psychological evaluation while receiving Comprehensive Services for the Blind at the Rehabilitation Institute, Detroit, Michigan. Those with additional physical impairments or emotional difficulties were eliminated. Follow-up procedures identified groups of 11 employed, 15 unemployed, and 17 attending college. The findings obtained from the Standardization sample were applied to the Crossvalidation sample. The Employment Key identified 82% of the working and 73% of the nonworking in Cross-Validation. Further, discriminant functions established on the Employment Key and Tolerance scales classified 82o of the working and 80%o of those not presently employed. The results obtained provisionally suggest some personality

100 traits which may differentiate the employed from the unemployed. The employed seem to have stronger feelings of self-worth, to be less inhibited by excessive fear, and to have a greater capacity for both adaptive behavior and committment to goal directed activities. The employed may be more permissive and accepting of the personal beliefs of others as well as less authoritarian and self-centered than their unemployed counterparts. The immediate limitations of this research lie with the relatively small sample sizes for both the Standardization and CrossValidation phases. In addition, unanticipated difficulties encountered while attempting to locate the relatively young, chronically unemployed blind suggests that this population may be transient and frequently out of contact with agencies concerned with assisting the blind to find employment. The arbitrary criteria for sample selection further suggest that the subjects of this study cannot be regarded as representative of the blind in general. It is felt that the sample does represent certain visually impaired who differ essentially on current employment status. While these findings must be recognized as exploratory and provisional, they are promising in that they may provide a useful approach to employment prediction as well as indicate the importance of certain relationships between personality traits and the employability of the blind. For those concerned with manpower development of the blind, it should be emphasized that the Employment Key scale items and the discriminant functions

101 presented can be conveniently used on an individual basis. The validity and usefulness of these findings in other settings will depend upon the results of such applications. In conclusion, the CPI appears to be a versatile personality inventory. It is widely researched and its scales are related to a number of behavioral dimensions. Its usefulness was demonstrated in this study, both with respect to its apparent suitability for personality measurement of the blind as well as for its power to discriminate between the employed and unemployed in this study. While recognizing the limitations of this study on the one hand, there appears to be sufficient strength on the other to propose that the CPI and these findings be used in additional research. Unemployment and underemployment threaten large numbers of persons each year whose work career is interrupted by deteriorating vision or injury, or who reached employment age with visual defect. This loss of manpower to the nation results in economic deprivation and reduced social status to the individuals afflicted. This study proposes that those concerned with the evaluation of vocational potential of the visually impaired consider these findings as an approach to systematic personality measurement of the blind. It is suggested that the CPI receive further study for identifying the relationships between personality traits and such factors as readiness for vocational training, the need for personal adjustment training, and the capacity for work responsibilities among the blind.

APPENDIX A RELEASE OF CASE INFORMATION 102

103 RELEASE OF CASE INFORMATION Date: I hereby authorize the Department of Social Services, Division of Services for the Blind to release any part or all of the information in my case record to Mr. Bernard Bast for his research study which I understand is sponsored by the Federal Government, Department of Labor Grant 91-24-70-14. I further agree to have Mr. Bast interview me and will accept the five dollars ($5.00) which is allowed to this purpose. All information obtained will be held confidential and will not be shared with any agency. Signed by:

APPENDIX B DESCRIPTIONS OF STANDARDIZATION AND CROSS-VALIDATION SAMPLES 104

105 TABLE B.1 CURRENT OCCUPATIONS OF EMPLOYED SUBJECTS IN STANDARDIZATION SAMPLE N Occupation 1 1 1 4 1 8 5 1 1 1 1 1 1 3 1 1 2 2 1 1 2 1 1 1 1 5 4 51 Bread slicer operator (bakery) Bookkeeper Building attendant (City Computer programmers Dairy products processor Darkroom technicians Dictaphone typists Drafting detailer Electroplater Electric motor rewinder of Detroit) Film developer Insurance underwriter Kitchen helper (hospital) Machine operators Musician Piano tuner Porters (nursing home) Packaging small parts (General Motors) Production assembly (General Motors) Production inspection (General Motors) Sales (door to door) Sales (electronic components) Stock receiver (supermarket) Teacher (piano and organ) Varnisher (hotel silverware) Vending stand operators Vending stand and canteen operators Total

106 TABLE B.2 AGE OF STANDARDIZATION SUBJECTS ~Age Employed Unemployed.ge N % N 21-25 5 10 6 12 26-30 12 23 4 8 31-40 15 29 10 20 41-45 5 10 10 20 51-55 7 14 7 14 56-61 2 4 9 18 Totals 51 100 50 100 Mean Age 38.45 yr 43.02 yr Mean Age Difference = 4.62 yr t = 2.02, P <.05 TABLE B.3 TYPE OF VISUAL IMPAIRMENT OF STANDARDIZATION SUBJECTS Type of Employed Unemployed Impairment N % N % Congenital 39 76 16 32 Progressive 8 16 20 40 Adventitious 4 8 14 28 Totals 51 100 50 100

107 TABLE B.4 YEARS OF ACADEMIC EDUCATION OF STANDARDIZATION SUBJECTS Highest Grade Employed Unemployed Completed N % N % 6 and below 1 1 3 6 7-11 8 16 26 52 12 26 51 16 32 1-3 years of college 10 20 3 6 B.A. degree 6 12 2 4 Totals 51 100 50 100 Mean yr education 11.94 10.54 Mean difference = 1.4 yr t ratio = 3.58, P <.01 TABLE B.5 FUNCTIONAL VISION OF STANDARDIZATION Tmnr l rgn~d SUBJECTS Unemployed N % Subject's Preference Braille and auditory media large print but prefers Braille Large print with/ without magnification Large print Ordinary print with Magnification Ordinary print Totals N 27 53 3 6 1 2 5 10 38 76 3 6 0 3 0 6 9 6 51 17 12 100 6 0 50 12 0 100

108 TABLE B.6 FUNCTIONAL MOBILITY OF STANDARDIZATION SUBJECTS Maximum- Degree Maxim D e Travel Aid Employed Unemployed of Travel (Method) N % N f Independenqce....... To new and distant Partial vision; 20 9 13 26 places Cane or dog 14 27 8 16 To and from work Partial vision; 1 2 0 0 Cane or dog 12 24 0 0 In city to conduct Partial vision; 2 4 personal affairs Cane or dog; 3 6 16 32 Person 1 2 In neighborhood No aide; 1 2 1 2 Cane or dog; 7 14 Person 2 4 Totals 51 100 50 100 TABLE B.7 MARTIAL STATUS OF STANDARDIZATION SUBJECTS Marital Employed Unemployed Status N ' N o Single 25 49 18 36 Married -24 47 22 44 Divorced, widowed, separated 2 4 7 14 Divorced, widowed & remarried 0 0 3 6 Totals 51 100 50 100

109 TABLE B.8 RACE OF STANDARDIZATION SUBJECTS Race Employed Unemployed Race...I N %o N % Caucasian 43 84 57 74 Black 8 16 13 26 Totals 51.100 50 100 -. ~ ~....,...... TABLE B.9 AREA OF RESIDENCE OF STANDARDIZATION SUBJECTS Area Employed Unemployed __N O%- _ N %mp Urban 49 "96 45 90 Urbanized area 2 4 3 6 Rural 0 0 2 4 Totals 51 100 50 100 TABLE B. 10 TYPE OF VISUAL IMPAIRMENT OF CROSS-VALIDATION SUBJECTS Employed, ollege Type of Employed Unemployed Separately C ee Students Impairment Classified N oN o _ N N ' Congenital 8 73 7 46 2 67 11 65 Progressive 3 27 5 27 0 0 6 35 Adventitious 0 0 4 27 1 3 0 0 Totals 11 100 15 100 3 100 17 100

110 TABLE B. 11 YEARS OF EDUCATION REPORTED AT TIME OF EVALUATION OF CROSS-VALIDATION SUBJECTS Employed,, ollege Highest Grade Employed Unemployed Separately Sde Completed Classified ~ N % N % N % N % 6 and below 0 0 1 7 0 0 0 0 7-11 3 27 8 53 2 67 2 12 12 7 64 5 33 1 33 14 82 1-3 years of college 1 _ 1 0 0 1 6 Totals 11 100 15 100 3 100 17 100 Mean yr education 11.636 9.866 Mean difference: Employed vs. Unemployed = 1.770 t = 1.683 n.s. TABLE B.12 REPORTED VISION OF CROSS-VALIDATION SUBJECTS Employed. Employed Unemployed Separately lege Degree of Blindness lasie Students Classified N % N % N % N % Total or light perception only 2 18 6 40 1 33 5 30 Partial vision (high or low) 82 60 2 67 12 70 Totals 11 100 15 100 3 100 17 100

11l TABLE B.,13 VERBAL IQ (WAIS) OF CROSS-VALIDATION SUBJECTS Employed, College IQ Range Employed Unemployed Separately Students Classified N g o N go N go N Lo 120 & above 1 9 0 0 0 0 8 47 110- 119 4 36 2 13 0 0 4 24 90 - 109 6 55 8 53 2 67 5 29 80 - 89 0 0 4 27 0 0 0 0 70- 79 o o 1 7 1 33 _o Totals 11 100 15 100 3 100 17 100 Mean IQ 107.363 95.800 Mean difference: Employed vs. Unemployed = 11.563 t = 2.659, P <.05 TABLE B.14 RACE OF CROSS-VALIDATION SUBJECTS Employed, ollege College Race Employed Unemployed Separately Students Classified N % N N % N % Caucasian 10 91 11 73 1 33 13 76 Black 1 4 27 2 7 4 24 Totals 11 100 15 100 3 100 17 100

112 TABLE B. 15 MARITAL STATUS OF CROSS-VALIDATION SUBJECTS Employed, College Marital Employed Unemployed Separately Status Classified N % N % N % N % Single 6 55 33 100 3 100 15 88 Married 4 36 6 40 0 0 2 12 Divorced 0 0 4 27 0 0 0 0 Widowed 1 _ 0 0 0 0 O 0 Totals 11 100 15 100 3 100 17 100..........,.......,,,,

APPENDIX C INSTRUMENTS UTILIZED IN THE STUDY

114 1. Interview Bast: Department of Labor 91-24-70-14 (See coding instructions) (1)* Employment status: Employed (1), Unemployed (2).... (2) Code number of subject.............................. |_I (3) Address Street or route City Urban (1) Urbanized area (2) Rural (3) (4) Date of interview (5) Birthdate Age.......... (6) Do you have any physical or other difficulties, besides visual impairment, which you consider to be a limitation to your employment opportunities? Yes (1), No (2). If yes, describe. (7) Marital status:................................. Single (1) Married (2) Divorced, widowed, or separated (3) Divorced or widowed, and remarried (4) (8) Relationship of current employment status to loss of vision: Type of onset: Congenital (1), Progressive (2), Adventitious (3). *Parentheses indicates that the item is to be coded but not asked directly of the subject.

115 Age of onset (defined as the age when visual loss became potentially handicapping to employment. congenital through two years (1) second birthdate through five years (2) fifth birthdate through fifteen years (3) sixteenth birthdate through twenty years (4) twenty-first birthdate through twenty-five years (5) twenty-fifth birthdate and later (6) (9) Academic education: (Code education received as a sighted person in column labeled Before; code education received as a visually impaired person in column labeled After. ).... Before After See coding instructions. (10) Vocational training: (Code Before and After as noted above, when appropriate.) See coding instructions. Record actual time spent. Technical school __yrs., _months. Describe: Business school __yrs., _months. Describe: Sheltered workshop/personal adjustment _yrs., months. (11) Job history: Job title and duties (beginning with current or most recent job). 1._______ 4.______4 2. 5._____,_ _ 3. 6.

116 7. 8. JOB # AS USED ABOVE BEGINNING & END DATES MONTH YEAI CODES A B I C l2. _ ____________ _ -.I 2. 5. 6. ___. 86. _________.,,. 7. _______________ 8. ____________ A. Income: (00) (01) (02) (03) (04) (05) Less than 501-1000 1001-2000 2001-3000 5001-4000 4001-5000 $500 (06) (07) (08) (09) (10) (11) 5001-6000 6001-7000 7001-8000 8001-9000 9001-10,000 10,001-12,500 (12) 12, 501-15,000 (13)15,001-17,500 (14)17,501-20,000 (15) 2, 001-25, 000 (16)25,001-30,000 (17) 30,001-and above B. Approximate number of hours employed per week: 10 (1), 20 (2), 30 (3), 40 (4), 50 (5), 60 (6). C. Employed as visually impaired. Yes (1), No (2). (12) Income on current job (annual)........ l l I (13) Approximate number of hours employed per week on current job: m 10 (1), 20 (2), 30 (3), 40 (4), 50 (5), 60 (6)....... L (14) Years of coninuous employment to date. l ll (15) Years of continuous unemployment to date. Lll

117 (16) Subject's estimate of vision: (check descriptions that apply) i (1) totally blind (2) light and shadows (3) hand movements (4) count fingers (5) use traffic lights (6) recognize faces (7) perceive smiling faces (8) read newspaper headlines Functional vision: How do you do most of your reading? (Number the top three in the order in which the subject uses them.) (1) Ordinary print without magnification (2) Ordinary print with magnification (3) Large print (4) Braille (5) Talking books (6) Tapes (7) Sighted reader Which methods of writing do you use? Often (1) Sometimes (2) Never (3) (a) Ordinary pen and paper (b) Special large pen (c) Regular typewriter (d) Slate and stylus

118 (17) Mobility: (check appropriate aides and extent of travel) A. To new and relatively distant places. No aide-partial vision (1) Cane or dog.......... (1) Person............... ( 4) Don't go.............. (5) If you travel independently to this extent, what forms of transportation do you use? Plane, train, long distance bus _, suburban bus, taxi. B. To and from work (if employed). No aide-partial vision (2) Cane or dog.......... (2) Person............... (5)_ Don't go............ (6) If you travel independently to this extent, what forms of transportation do you use? Walk, bus, taxi, driver. C. In city or town of residence to conduct your own business (code even if employed). No aide-partial vision (2)_ Cane or dog........... (2) Person................ (5) Don't go.............,(6) If you travel independently to this extent, what forms of transportation do you use? Walk, bus, taxi_, driver D. In your neighborhood. No aide.............. (3) Cane or dog.......... (4) Person.............. (6) Don't go (homebound). (7)

119 2. Interview: Coding Instructions Item Column Variable Code (1) 2 Current employment status Employed Unemployed 1 2 001-999 (2) 3-5 Code number of subject (3) 6 Geographical location* Urban (comprising those living in places of 2,500 inhabitants or more) 1 Urbanized area (comprising closely settled places containing at least 50,000 inhabitants or more) Rural (comprising persons living on farms and places of less than 2,500 inhab.) 2 3 (4) - Date of interview (5) 7-9 Years of age Code actual number from birthdate and date of interview. Do not ask age. Code col 9 in tenths of yrs. (6) 10 Vocationally handicapping conditions other than visual impairment (7) 11 Marital status If yes If none If not known Single (never married) Married Divorced, widowed, or sep. Divorced or widowed, and remarried 1 2 3 1 2 3 *Adapted from: U.S. Bureau of the Census. U. S. Census Population1960, Vol. I, Characteristics of the Population Part A, Number of Inhabitants. U. S. Government Printing Office, Washington, D. C., 1961.

120 Item Column Variable Code (8) 12 Type of onset of visual loss (9) 14 Age of onset of visual loss Congenital Progressive Adventitious-Traumatic Congenital thru 2 yrs 2 yrs thru 5 yrs 5 yrs thru 15 yrs 16 yrs thry 20 yrs 21 yrs thry 25 yrs 25 yrs and older 1 2 3 1 2 3 4 5 6 (10) 15,16 Completed years of edu- Grade 6 cation before visual " 7 loss, i.e., education as " 8 a sighted person " 9 " 10 17,18 Completed years of edu- " 11 cation as a visually " 12 impaired person 1 yr college 2 yrs college 3 yrs college 4 or more yrs B.A. M.A. or M.S. Professional degree Ph.D.. or Ed.D. Ex. A subject who has a B.A. and lost his vision in the 11th grade would be coded as such: Before After 1 0 0 7 (11) 19 Length of time in voca- Up to 1 mo tional training (tech- 1 mo+ thru 3 mo nical or business 3 mo+ thru 6 mo school) as sighted 6 mo+ thru 1 yr person 1 yr+ thry 18 mo 18 mo+ thru 2 yrs 20 Length of time in voca- 2 yrs+ thru 3 yrs tional training (tech- over 3 yrs nical or business school) as a visually impaired person 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 1 2 3 4 5 6 7 8

121 Item Column,., Variable Code Technical school is defined as any institution preparing persons for employment in a skilled trade (electronics, office machines repair, building trades, etc. ) Business school is defined as an institution preparing one for employment in office type skills (bookeeping, dictaphone typing, etc.) (12) -- Employment history To be coded later Instruction: Begin with current or last employer and work backward to first employer. Record job title, duties, length of time on each job, income, hours employed per week, whether each employment segment was performed as a sighted or visually impaired person. (13) 21-25 Annual income for 1969 if employed Record annual income in actual number of dollars. (14) 26 Hours employed per week as indicated in item 11. Under 10 hrs per wk 10+ thru 20 hrs 21+ thru 30 hrs 31+ thru 40 hrs 41+ thru 50 hrs 51+ thru 60 hrs over 60 hrs per wk 1 2 5 4 5 6 7 (15) 27-28 Duration of employment Code actual no. of yrs 29 (16) 30-31 Code fraction of a yr in tenths Duration of unemployment Code actual no. of yrs 32 Code fraction of a yr in tenths (17) 33 Functional vision (Subject' s estimate) Check subject's estimate of vision as he reports it. To be coded later.

122 Item Column Variable 34 Functional vision (Primary reading methods used) Code (17) cont'd Ordinary print without magnification Ordinary print with magnification Large print Braille Talking book, tapes Used most often Used frequently Used infrequently Not used Used most often Used frequently Used infrequently Not used Used most often Used frequently Used infrequently Not used Used most often Used frequently Used infrequently Not used Used most often Used frequently Used infrequently Not used 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 5 0 Sighted reader Used most often Used frequently Used infrequently Not used 1 2 3 0 Instruction: Code Column 34 according to the following combinations of usable vision. Ordinary print and ordinary pen and paper and/or typewriter.................................. 1 Ordinary print with magnification and ordinary pen and paper and/or typewriter.......................2 Large print and ordinary pen and paper and/or typewriter........ 0.................

123 Item Column Variable Code Large print (with or without magnification) and special pen or typewriter............................4 Prefers braille or auditory media for reading, but indicates can write print with ordinary pen, special pen, or large typewriter........................ 5 Braille, slate and stylus, talking books, tapes, sighted reader.............................6 Functional Vision (Writing methods used by subject) (16) cont 'd Ordinary pen and paper Special large pen Regular typewriter Slate and stylus Often Seldom Never 1 2 5 Often Seldom Never 1 2 3 Often Seldom Never 1 2 3 Often Seldom Never 1 2 (17) 35 Degree of independence in travel Most independent Very independent Moderately independent 1 2 3 4 5 6 7 Moderately dependent Very dependent Most dependent Instructions: Of sections A, B, C, and D (relative to the extent of travel to distant places, to work, in city of residence, or in neighborhood) code (in Column 35) the number which best represents the subject's ability to travel independently.

124 3. Classes, Scales, and Scale Abbreviations of the CPI* CLASS I. MEASURES OF POISE, ASCENDANCY, AND SELF-ASSURANCE 1. Do Dominance 2. Cs Capacity for Status 5. Sy Sociability 4. Sp Social Presence 5. Sa Self-acceptance 6. Wb Sense of Well-being CLASS II. MEASURES OF 7. Re 8. So 9. Sc 10. To 11. Gi 12. Cm SOCIALIZATION, MATURITY, AND RESPONSIBILITY Responsibility Socialization Self-control Tolerance Good impression Communality CLASS III. MEASURES OF EFFICIENCY ACHIEVEMENT POTENTIAL AND INTELLECTUAL 13. Ac 14. Ai 15. Ie Achievement via conformance Achievement via independence Intellectual efficiency CLASS IV. MEASURES OF INTELLECTUAL AND INTEREST MODES 16. 17. 18. Py Fx Fe Psychological-mindedne ss Flexibility Femininity *Reproduced by special permission from The California Psychological Inventory by Harrison G. Gough, Ph.D. Copyright 1956. Published by Consulting Psychologists Press, Inc.

APPENDIX D ADDITIONAL DATA ANALYSES 125

126 TABLE D.1 FORMULA FOR ESTIMATING THE MAHALANOBIS DISTANCE BETWEEN TWO POPULATIONS 2 P P D = E Z wij(x.l - x.2)(xj1 - x.2) P 12 1 1 2 where p in the symbol D indicates the number of characters used (Rao, 1956, pp. 246-247); and N wij (N + N 2) [ Z (x.lc - xil.)(xlc - x.j.) 1 2 O 1 N2 + Z (x.2c -.2.)(x.2c - x.2.)] 1 1 0 0 c=l with the first subscript denoting the variate, the second denoting the group, and the third indicating the individual (Anderson, 1966, p. 169).

127 Y \Y I J I\^ ) / Figure D.1. Projection of "new" scores onto discriminant function 1. The discriminant function transforms the individual test scores to a single discriminant score. That score is the individual's location along line y1.

TABLE D.2 WITHIN GROUPS CORRELATION MATRIX Cs.57 Sy.75.67 Sp.59.67.69 Sa.74.58.72.70 Wb.35.35.39.40.26 Re.40.48.30.25.22.50 So.01.12.02 -.07 -.05.45.46 Sc -.11.15 -.08 -.12 -.25.54.41.53 To.32.62.39.56.28.62.63.29.47 Gi.23.33.25.05.04.48.29.37.70.37 Cm.19.18.50.22.20.34.25.09 -.03.10 -.20 Ac.52.51.54.28.43.50.58.40.38.44.45.26 Ai.26.53.26.50.27.44.50.11.38.78.31 -.09.35 Ie.51.71.63.66.48.57.60.21.28.76.27.27.58.68 Py.42.55.38.46.40.33.533.05.30.55.37.07.38.58.53 Fx.10.29.11.50.20 -.03.09 -.22 -.05.38 -.10.21 -.07.52.24.38 Fe -.39 -.31 -.45 -.42 -.39 -.31 -.02.20.11 -.22 -.03 -.28 -.15 -.15 -.41 -.27 -.05 Ek.45.62.45.64.38.58.52.12.29.77.26.2L.33.71.70.53.41 -.27 Do Cs Sy Sp Sa Wb Re So Sc To Gi Cm Ac Ai Ie Py Fx Fe H-i O

129 TABLE D.3 EMPLOYMENT KEY ITEMS Item Scale(s) of Direction Chi Item Square No. ItCPI Origin Scored uare ~,,Ratio 13. I am very slow in making up my mind. 41. For most questions there is just one right answer, once a person is able to get all the facts. Fe(F)* Ai(F) False False 7.21 7.14 47. Women should not be allowed to drink in cocktail bars. 79. I am afraid of deep water. 90. As long as a person votes every four years, he has done his duty as a citizen. 128. It takes a lot of argument to convince most people of the truth. 154. I like tall women. 182. I would rather go without something than ask for a favor. Sa(F), Sp(F), Cs(F), Cs(F) Re(F) Cs(F) Cs(T) Sa(F), So(F) False False False False True False 17.8 7.2 7.5 9.6 14.3 8.9 184. I have had more than my share of things to worry about. 210. I very much like hunting. 225. People pretend to care more about one another than they really do. Ie(F), To(F), So(F) Do(F), Fe(F), Re(F) Ai(F), Sy(F) False False False 8.6 7.4 6.7 *(T) (F) = true = false

130 TABLE D.3 (Concluded) Item Scale(s) of Direction Chi No.Item CPI Origin Scored Suare L,__,,'X_....Ratio 237. The future is too uncertain for a person to make serious plans. Ai(F), To(F) False 14.5 261. We ought to let Europe get out of its own mess; it made its bed, let it lie in it. Re(F) False 6.6 300. Police cars should be especially marked so that you can always Sc(F), Sa(T), see them coming. Re(F) False 12.1 311. I cannot do anything well. Cm(F) False 8.9 338. I never worry about my looks. So(F) False 8.6 372. I have reason for feeling jealous of one or more members of my family Wb(F) False 7.3 377. Most of the arguments or quarrels I get into are over matters of principle. Fx(F) False 7.8 383. I usually have to stop and think before I act even in trifling matters. Do(F) False 8.3 425. I have often felt guilty because I have pretended to feel more sorry about something than I really was. Wb(F) False 7.4 452. I dislike to have to talk in front of a group of people. Do(F) False 12.6

131 TABLE D. 4 COMPARISON OF SCALE MEANS ACROSS STANDARDIZATION AND CROSS-VALIDATION SAMPLES Standardization Sample Cross-Validation Samples College CPI Scales Employed Unemployed Working Not Working eg Students N=51 N=50 N=ll N=15 N=17 Dominance 29.35 25.74 27.18 23.00 29.00 Capacity for Status 18.82 15.60 19.09 15.06 17.88 Sociability 24.76 22.78 26.00 21.93 25.76 Social Presence 33.50 30.22 33.72 31.33 36.23 Self-acceptance 20.49 18.76 20.90 18.60 22.29 Well-being 36.76 33.44 37.36 34.53 35.64 Responsibility 30.35 28.06 28.63 24.53 29.58 Socialization 35.88 33.38 35.36 33.26 36.00 Self-control 30.56 28.34 29.72 27.93 25.70 Tolerance 19.62 16.42 19.18 17.80 20.47 Good Impression 20.60 19.18 19.63 18.46 16.88 Communality 26.56 24.98 25.09 25.20 25.88 Achieve.-Conform. 27.33 24.48 26.27 23.00 25.76 Achieve.-Independ 16.60 13.26 15.00 13.80 18.47 Intellect. Effic. 35.72 32.48 35.63 34. 0 37.35 Psych. Minded 10.98 9.22 10.00 10.46 10.35 Flexibility 6.54 5.30 6. 6.13 9.47 Femininity 18.17 18.44 16.72 16.46 16.05 Employ Key 14.84 9.16 13.36 10.73 13.88

132 TABLE D.5 COMPARISON OF SCALE STANDARD DEVIATIONS ACROSS STANDARDIZATION AND CROSS-VALIDATION SAMPLES Standardization Sample Cross-Validation Samples College CPI Scales Employed Unemplo'red Working Not Working Studets Students N=51 N=50 N=11 N=15 N=17 Dominance 6.o5 6.46 5.67 5.89 7.54 Capacity for Status 4.38 4.26 4.03 4.69 3.58 Sociability 5.12 5.55 3.66 6.61 3.89 Social Presence 6.10 5.96 4.02 6.61 6.49 Self Acceptance 3.80 4.17 3.78 3.64 4.17 Well-being 4.94 6.02 5.73 5.16 3.58 Responsibility 4.91 5.03 3.72 7.06 4.34 Socialization 6.21 5.67 6.15 6.004.55 Self-Control 6.52 7.19 8.77 8.16 5.57 Tolerance 6.01 5.53 4.33 5.82 3.84 Good Impression 5.32 5.82 6.40 7.46 4.15 Communality 1.61 2.72 3.72 2.36 1.93 Achiev.-Conform 4.79 4.66 4.31 7.32 4.67 Achieve.-Indep. 4.09 4.87 3.74 3.02 4.12 Intell. Effic 6.03 6.66 3.38 7.07 3.69 Psych. Minded 2.95 2.42 2.52 2.58 2.59 Flexibility 3.78 3.64 3.91 2.26 4.04 Femininity 3.99 3.75 3.34 4.32 3.74 Employ. Key 2.58 3.57 3.44 2.753 2.99

PROFILE SHEET FOR THE t bug # nwwy/oivea a: MALE Notes: Do Cs Sy Sp Sa Wb Re So Sc To Gi Cm Ac Ai Ie Py Fx Fe t I MALE NORMS I w I 0 l) -. -. U dV) 30 — 30 -30 - - 3 - 30 -40 -50 -a a - z:o - 0 - 2 70o — -- 45 0 --- ---— 70o _____ _ -40 -.45 - ~4 4 - -530 15 — 35 - -25 - 45 - 5 - 0 —____ -30 40 - 45 -:5050 " - 308 ______0 -0:0 60 - 20 - - - 10 ~-35 - - - - - - o -2 - -40 35 - ~-25~ ~- - -25 -:~ ~ -10 30 - 1 ": * J io -4 ' 5 --- 40 _ ~ 4 —0- 0-0- -0 - -0 25 — - 40 -! -20 60 ------ 0 -1: — 20 - - -_ _ — 10 - - — 20 ---- _ o _~ -._15 1- - -0 - - - "5- - - 10 -. —,-v - - ' 10 - i— 0 0 - --- 5c -2li -. I? — 17 -- 1 --- lo -7 -C--- - 7 - y - 6 - I Employed Group (N=ll) Unemployed Group (N-15) H Do Cs Sy Sp Sa Wb Re So Sc To Gi Cm Ac Ai le Py Fx Fe Figure D.2. Comparison of mean scale scores of working and nonworking Cross-Validation subjects.

PROFILE SHEET FOR THE &4IWua::40c1yma nllr: MALE Notes: Do Cs Sy Sp Sa Wb Re So Sc To Gi Cm -. -- - 90 - -35 N=17 -55 — -45 90 -30 - _ __ __- -3 -o40 -30: - 50 -50 30 - - - - - -35 - -5 - -'- 50 -35:-25~~~~~~ -. ~ ~ — O--- - - -3 - -2 - -5: - - 25 T60:. -: - -- - - -:35 -25 - 35 -2a5=45 -30 -25 _ i40 _4530 6030 — 40 60 -1 - - ~~ - -20 - o 0 o o -2 -0 AA 5-35 -- - -20 - -/\ -4 - -40 -35 -:30 - -2 --- -20.41- -40 50 _: ____ ^ -:20 - 1 _2- 350_ _ 4 0 _- - - -- 1 - -2 - 30 1 -15 40 - - ~-o - — 20 - -2 - 35 — 15 - - 15 5 - - -S - — 25- -- - 30 — 20 - - 0- - - -10 -- - - -2- - - 0 -5 - -1520 30 — ---- 3< 20 -2: —. 0 - t55 -0 - -' -- -70 60 a QL a 50 C- O -40 - 20 -0 0 - H 2 -,5 o - - - -: -20 - - 20 10 - to-o Ia- - - - - -0 =2 I - — 100 - - 15 - 10- _10_ _-;2 0 1 - 5: 0 0 0 -I _, _ -1 -: - _ - - - -20. _ -_ _ -_ -2 _0 n -_ _ - 0 _ _ _ —; -_ - 510 - 1_ 0 rr -0 Do Cs Sy Sp Sa Wb Re So Sc To Gi Cm Ac Ai lI Py Fx Fe Figure D.3. Mean scale scores of college students in the CrossValidation sample.

135 TABLE D.6 CROSS-VALIDATION SUBJECTS' ACTUAL GROUP MEMBERSHIP, PREDICTED CLASSIFICATION, SQUARED MAHALANOBIS DISTANCES, AND ASSOCIATED PROBABILITY OF PREDICTED GROUP MEMBERSHIP FOR THE TWO-VARIABLE EQUATIONS Group 1 WORKING CASE 1 2 3 4 5 6 7 8 9 10 11 EMPLOY EMPLOY UNPLOY EMPLOY UNPLOY EMPLOY EMPLOY EMPLOY EMPLOY EMPLOY EMPLOY o.848( 1.019( 16.196( 1.529( 4.o68( o. 425( 0.222( o.928( 1.947( 2.216( 1.013( 0.674) 0.992) 0.002) 0.507) 0.135) 0.752) 0.929) 0.988) 0.687) 0.984) 0.810) 2.305( 10.612( 3. 744( 1.582( 0.355( 2.643( 5.366( 9.760( 3.523( o. 445( 3.914( 0.326) 0.008) 0.998) 0.495) 0.865) 0.248) 0.071) 0.012) 0.313) o. o6) 0.190) Group 2 NOT WORKING CASE 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 EMPLOYED UNEMPLOYED UNPLOY UNPLOY UNPLOY UNPLOY UNPLOY EMPLOY EMPLOY UNPLOY UNPLOY UNPLOY UNPLOY UNPLOY EMPLOY UNPLOY UNPLOY 7.560( 2.427( 3.562( 6.838( 10.199( 1.013( o.680( 6.532( 2.127( 3.218( 7.471( 9.110( 1.265( 3.262( 3.741( 0.266) o. 403) o.243) o.048) 0.018) 0.810) o.905) o. 68) 0.324) 0.186) 0.036) 0.018) 0.759) 0.189) 0.183) 5.527( 1.640( 1.292( 0. 88o( 2.196( 3.914 5.182( 1.296( o.658( 0.266( 0.911( 1.o67( 3.562( 0.350( o.750( o.754) 0.597) 0.757) 0.952) 0.982) 0.190) o.095) 0.932) 0.676) 0.814) 0.964) 0.982) 0.241) 0.811) 0.817)

136 TABLE D.7 CROSS-VALIDATION SUBJECTS ACTUAL GROUP MEMBERSHIP, PREDICTED CLASSIFICATION, SQUARED MAHALANOBIS DISTANCES, AND ASSOCIATED PROBABILITY OF PREDICTED GROUP MEMBERSHIP FOR THE FIVE-VARIABLE EQUATIONS Group 1 WORKING CASE 1 2 3 4 5 6 7 8 9 10 11 EMPLOY EMPLOY UNPLOY UNPLOY UNPLOY EMPLOY EMPLOY EMPLOY EMPLOY EMPLOY EMPLOY 1.105( 5.318( 23.154( 5.584( 6.048( 1.935( 1.433( 3.210( 2.938( 4. 494( 5. 737( o. 774) 0.998) o.000ooo) 0.253) 0.082) 0.700) 0.972) 0.971) 0.877) 0.997) 0. 734) 3.565( 16.103( 6.683( 3.200( 1.211( 3.626( 8.542( 10.222( 6.865( 16. 061( 7.764( 0.226) 0.002) 1.000) 0.767) 0.918) 0.300) 0.028) 0.029) 0.123) 0.003) 0.266) Group 2 NOT WORKING CASE 1 2 5 4 5 6 7 8 9 10 11 12 13 14 15 UNPLOY UNPLOY UNPLOY UNPLOY UNPLOY EMPLOY EMPLOY UNPLOY UNPLOY UNPLOY UNPLOY UNPLOY EMPLOY UNPLOY UNPLOY 12.129( 15.856( 6.928( 18.369( 11. 492( 6.242( 5.782( 26. 433( 9.150( 4.844( 9.736( 12.586( 4.099( 10. 924( 5.285( 0.247) 0.036) 0.158) 0.008) 0.035) 0.981) 0.676) 0.004) o. 409) 0.108) 0.018) o.oo6) 0.812) 0.032) 0.154) 9.802( 9.306( 3.579( 8. 711( 4.877( 1l4. 105( 7.255( 15.358( 8. 418(. 618( 1.712( 2.246( 7.021( 4.(97( 1.883( 0.753) 0.964) 0.842) 0.992) 0.965) 0.019) 0.324) 0.996) 0.591) 0.892) 0.982) 0.994) 0.188) 0.968) 0.846)

PROFILE SHEET FOR THE 4n aymaI lye/M1iu:a( Aw'nij MALE Notes: Do Cs Sy Sp Sa Wb Re So Sc To Gi Cm Ac Ai le Py Fx Fe E I MALE NORMS -35 90 -45 -55 _ D R,,..s _;.......g. _ S s -40 -20 -20 -30 70 - _ -_30 _ ~-30 -40 - -_! — 5aI _ ^30 — 40 - - - -50;- _ — - -50 - -25 _-35 _2_. _ 5 ___ 45 -4 - -5 -30 - - - -2535 -25 30 35 _25 _ - -; _. -:: 40 - - 45 - 30 -40 011 0 -0 -0r O tA 50 - 40 -.-35 -25 -30 -40,:- - 40 Z35 — 30 -25 - 20 L-10 — 20 -35 -20 _;__ \_ _-25 4_ — 30~ ~ 3 15 30 - \ 1 -20 -70 60 CA aL~ -50 o (t, o o40 -40 H \I -— I 30 -20 -10 -0-:20 - -: -25-_ 30 _ _ _ _. _ - - - -25 - O -20 - 1030 -20 - - -:2:0 - - - -0 10 -20 -_ _ - _ 25 - -2 - 15 - 0 - 20 15 - - -530 -1525 - - - - - _ -~~ _:5 _5 - -_ - o -0 i5- -1 - 02 - 0 - 20 - 0- - - -20 10 10 -; — 1 _ - - 10 - - - -. - 5 _25 _7 — 257 10 15 - 0- 5 5 5 i 5 5 -5 10 0 5 --20. 0 -: - 10 10 0 0O 5 30 20 10 0 Do Cs Sy Sp Sa Wb Re So Sc To Gi Cm Ac Ai le Py Fx Fe Figure D.4. Grand mean scale scores across all groups.

BIBLIOGRAPHY Adorno, T. W., Frenkel-Brunswik, E., Levinson, D. J., and Sanford, R. N. The Authoritarian Personality. New York: Harper, 1950. Aiken, L. R., Jr. The relationship of dress to selected measures of personality in undergraduate women. Journal of Social Psychology, 1963, 59, 119-128.. Personality correlates of attitude toward mathematics. Journal of Educational Research, 1963, 56, 476-480. Anastasi, A. Psychological Testing. New York: MacMillan, 1968. Anderson, H. E. Regression, discriminant analysis, and a standard notation for basic statistics. In Catell, R. B. (Ed.), Handbook of Multivariate Experimental Psychology. Chicago: Rand McNally, 1966. Anderson, T. W. Introduction to Multivariate Statistical Analysis. New York: Wiley, 1958. Bass, B. M. and Berg, I. A. (Eds.). Objective Approaches to Personality Assessment. Princeton, N.J.: Van Nostrand, 1959. Bauman, Mary K. Emotional Factors Inventory. Philadelphia: Personnel Research Center, 1950. _. Adjustment to Blindness: A Study as Reported By the Committee to Study Adjustment to Blindness. Harrisburg: Pennsylvania State Council for the Blind, 1954. Characteristics of Blind and Visually Handicapped People in Professional, Sales and Managerial Work. Pennsylvania: Office for the Blind, 1963.. Group differences disclosed by inventory items. International Journal for the Education of the Blind, 1964, 13, 101-106. Bauman, Mary K. and Hayes, S. P. A Manual for the Psychological Examination of the Adult Blind. New York: The Psychological Corporation, 1951. Bauman, M. K., Platt, H. and Strauss, S. A measure of personality for blind adolescents. International Journal for the Education of the 138

139 BIBLIOGRAPHY (Continued) Blind, 1963, 13, 7-12. Bauman, M. K. and Yoder, N. M. Adjustment to Blindess —Reviewed. Springfield, Ill.: Charles C. Thomas, 1966. Bennett, L. A. and Rudoff, A. Evaluation of modified administration of the California Psychological Inventory. Journal of Clinical Psychology, 1957, 303-304. Blank, H. Robert. Psychoanalysis and blindness. The Psychoanalytic Quarterly. 1957, 26, 1-24.. Blindspots in the professional worker about blindness. New Outlook for the Blind, 1958a, 52, 173-175.. Countertransference problems in the professional worker. New Outlook for the Blind, 1958b, 52, 185-188. _. Dreams of the blind. Psychoanalytic Quarterly, 1958c, 27, 158-161. _. Psychiatric problems associated with congenital blindness due to retrolental fibroplasia. New Outlook for the Blind, 1959, 53, 237-245. Bonk, E. Counseling implications of the MMPI for blind people in selected occupations. Dis. Abstr., 1955, 15, 2095. Bowman, P. H. and Matthews, C. V. Motivations of Youth for Leaving School. Quincy, Ill.: University of Chicago, Quincy Youth Development Project, 1960. Burlingham, D. Some notes on the development of blind children. Psychoanalytic Study of the Child, 1961, 16, 121-144.. Some problems of ego development in blind children. Psychoanalytic Study of the Child, 1965, 20, 194-208. Canter, F. M. Simulation on the California Psychological Inventory and the adjustment of the simulator. Journal of Consulting Psychology, 1963, 27, 253-256. Carson, G. L. and Parker, C. A. Leadership and profiles on the MMPI and CPI. Journal of College Student Personnel, 1966, 7, 14-18.

BIBLIOGRAPHY (Continued) Carroll, T. J. Blindness. Boston: Little, Brown, 1961. Cholden, L. S. The role of the psychiatrist in the adjustment center for the blind. New Outlook for the Blind, 1952, 46, 225-228.. A Psychiatrist Works with Blindness. New York: American Foundation for the Blind, 1958. Cole, N. J. and Taboroff, L. H. The psychological problems of the congenitally blind child. American Journal of Orthopsychiatry, 1955, 25, 627-659. Cooley, W. W. and Lohnes, P. R. Multivariate Procedures for the Behavioral Sciences. New York: Wiley, 1962. Cowen, E. L., Underberg, R. P. and Verillo, R. T. The development and testing of an attitude to blindness scale. Journal of Social Psychology, 1958, 48, 297-304. Cowen, E., Underberg, R., Verillo, R., and Benham, F. Adjustment to Visual Disability in Adolescence. New York: American Foundation for the Blind, 1961. Crites, J. 0., Bechtoldt, H. P., Goodstein, L. D., and Heilbrun, A. B., Jr. A factor analysis of the California Psychological Inventory. Journal of Applied Psychology, 1961, 45, 408-414. Cronbach, L. J. Essentials of Psychological Testing. New York: Harper and Brothers, 1960. Cross, O. Braille edition of the Minnesota Multiphasic Personality Inventory for use with the blind. Journal of Applied Psychology, 1947, 31, 189-198. Cutsforth, T. D. The Blind in School and Society. New York: American Foundation for the Blind, 1951.. Personality and Social adjustment among the blind. In P. A. Zahl (Ed.), Blindness. New York: Hafner, 1962, 174-187. Dean, S. Adjustment testing and personality factors of the blind. Journal of Consulting Psychology, 1957, 21, 171-177. Dembo, T. Leviton, G. L., and Wright. B. A. Adjustment to misfortune

141 BIBLIOGRAPHY (Continued) a problem of social-psychological rehabilitation. In M. H. Goldberg and J. R. Swinton (Eds.), Blindness Research: The Expanding Frontiers. University Park, Pa.: Pennsylvania State University Press, 1969, 254-318. Dicken, C. F. Simulated patterns on the Edwards Personal Preference Schedule. Journal of Applied Psychology, 1959, 43, 372-378.. Simulated patterns on the California Psychological Inventory. Journal of Counseling Psychology, 1960, 7, 24-31.. Convergent and discriminant validity of the California Psychological Inventory. Educational and Psychological Measurement, 1963a, 23, 449-459.. Good impression, social desirability, and acquiescence as suppressor variables. Educational and Psychological Measurement, 1963b, 23, 699-719. Dishart, M. Testing the blind for rehabilitation using a psychological profile. New Outlook for the Blind, 1959, 53, 1-14. _. A study of the effectiveness of psychological services for the blind. New Outlook for the Blind, 1960, 10 (Reprint). Dixon, W. J. (Ed). BMD Biomedical Computer Programs. Berkeley and Los Angeles, California: University of California Press, 1968. Domino, G. Differential prediction of academic achievement in conforming and independent settings. Journal of Educational Psychology, 1968, 59, 256-260. Edwards, A. The relationship between judged desirability of a trait and the probability that the trait will be endorsed. Journal of Applied Psychology, 1953, 37, 90-93.. The Social Desirability Variable in Personality Assessment and Research. New York: Dryden Press, 1957. Elonen, Anna S., and Cain, Albert C. Diagnostic evaluation and treatment of deviant blind children. American Journal of Orthopsychiatry, 1964, 34, 625-633. Fink, M. B. Cross-Validation of an underachievement scale. California

142 BIBLIOGRAPHY (Continued) Journal of Educational Research, 1963, 14, 146-152. Finney, Ben C. and Van Dalsem, Elizabeth. Group counseling for gifted high school students. Journal of Counseling Psychology, 1969, 16, 87-94. Fitting, E. A. Evaluation of Adjustment to Blindness. New York: American Foundation for the Blind, Res. Ser. 2, 1954. Fraiberg, S. and Freedman, D. A. Studies in the ego development of the congenitally blind child. Psychoanalytic Study of the Child, 1964, 19, 113-169. Fraiberg, S., Siegel, B. L., and Gibson, R. The role of sound in the search behavior of a blind infant. The Psychoanalytic Study of the Child, 1966, 21, 327-357. Freedman, S. Reactions to blindness. New Outlook for the Blind, 1965, 59, 344-346.. Personality growth. New Outlook for the Blind, 1966, 60, 173 -176. Freud, S. Psychogenic visual disturbance according to psychoanalytic conceptions. Arztliche Standeszeitung, Vienna, 1910. (Republished in Collected Papers, Vol. 2, New York: Basic Books, 1959.) Fries, M. E. and Woolf, P. J. Some hypotheses on the role of congenital activity type in personality development. The Psychoanalytic Study of the Child, 1953, 8, 48-62. Gill, L. J. and Spilka, B. Some nonintellectual correlates of academic achievement among Mexican-American secondary school students. Journal of Educational Psychology, 1962, 553, 144-149. Goldberg, M. H. (Ed.). Blindness Research: The Expanding Frontiers. University Park, Pa.: Pennsylvania State University Press, 1969. Goodstein, L. D., Crites, J. 0., Heilbrun, A. B.,Jr., and Rempel, P. P. The use of the California Psychological Inventory in a University Counseling Service. Journal of Counseling Psychology, 1961, 8, 147-153. Goodstein, L. D., and Schrader, W. J. An empirically-derived

143 BIBLIOGRAPHY (Continued) managerial key for the California Psychological Inventory. Journal of Applied Psychology, 1963, 47, 42-45. Gough, H. G. Manual for the California Psychological Inventory. Palo Alto, Calif.: Consulting Psychologists' Press, 1957. (Revised edition, 1964.). The adjective check list as a personality assessment research technique. Psychological Reports, 1960a, 6, 107-122.. Theory and measurement of socialization. Journal of Consulting Psychology, 1960b, 24, 23-30.. Clinical versus statistical prediction in psychology. In L. Postman (Ed.), Psychology in the Making. New York: Alfred Knopf, 1962, 526-584.. Academic achievement in high school as predicted from the California Psychological Inventory. Journal of Educational Psychology, 1964a, 55, 174-180.. Achievement in the first course in psychology as predicted from the California Psychological Inventory. Journal of Psychology, 1964b, 57, 419-430.. A cross-cultural study of achievement motivation. Journal of Applied Psychology, 1964c, 48, 191-196.. Conceptual analysis of psychological test scores and other diagnostic variables. Journal of Abnormal Psychology, 1965a, 70, 294-302. Cross-cultural validation of a measure of asocial behavior. Psychological Reports, 1965b, 17, 379-387. _ Appraisal of social maturity by means of the CPI. Journal of Abnormal Psychology, 1966a, 71, 189-195. ___. A cross-cultural analysis of the CPI Femininity scale. Journal of Consulting Psychology, 1966b, 30, 136-141.. Graduation from high school as predicted from the California Psychological Inventory. Psychology in the Schools, 1966c, 3, 208-216.

144 BIBLIOGRAPHY (Continued) Gough, H. G. College attendance among high-aptitude students as predicted from the California Psychological Inventory. Journal of Counseling Psychology, 1968a, 15, 260-278.. Cross-cultural approaches to the study of delinquency. Paper presented at a symposium on delinquency, Annual meetings of the American Psychological Association. San Francisco, September, 1968b.. An interpreter's syllabus for the California Psychological Inventory. In P. McReynolds (Ed.), Advances in Psychological Assessment, Vol. I, Pal Alto, Calif.: Science and Behavioral Books, 1968c. ___. A leadership index on the CPI. Journal of Counseling Psychology, 1969, 16, 283-289. Gough, H. G., Chun, K., and Chung, Y. Validation of the CPI Femininity scale in Korea. Psychological Reports, 1968, 22, 155-160. Gough, H. G., DeVos, G., and Mizushima, K. Japanese validation of the CPI social maturity index. Psychological Reports, 1968, 22, 143 -146. Gough, H. G., Durflinger, G. W., and Hill, R. E., Jr. Predicting performance in student teaching from the California Psychological Inventory. Journal of Educational Psychology, 1968, 59, 119-127. Gough, H. G. and Fink, M. B. Scholastic achievement among students of, average ability, as predicted from the California Psychological Inventory. Psychology in the Schools, 1964, 1, 375-380. Gough, H. G. and Hall, W. B. Prediction of performance in medical school from California Psychological Inventory. Journal of Applied Psychology, 1964, 48, 218-226. Gough, H. G. and Heilbrun, A. B., Jr. The Adjective Check List Manual. Palo Alto, Calif.: Consulting Psychologists' Press, 1965. Gough, H. G. and Sandhu, H. S. Validation of the CPI socialization scale in India. Journal of Abnormal Psychology, 1964, 68, 544 -547. Gough, H. G., Wenk, E. A., and Rozynko, V. Parole outcome as predicted

145 BIBLIOGRAPHY (Continued) from the CPI, the MMPI, and a base expectancy table. Journal of Abnormal Psychology, 1965, 70, 6, 432-441. Graham, M. D. Toward a functional definition of blindness. New Outlook for the Blind, 1959, 53, 285-288.Greenberg, H., Allison, L., Tewell, M., and Rich, C. The personality of junior high and high school students attending a residential school for the blind. Journal of Educational Psychology, 1957, 48, 406-410. Greenberg, H. and Jordan, S. Differential effects of total blindness and partial sight on several personality traits. Exceptional Children, 1957, 24, 125-124. Hall, C. S. and Lindzey, G. Theories of Personality. New York: John Wiley and Sons, 1957. Hase, H. D. and Goldberg, L. R. Comparative validity of different strategies of constructing personality inventory scales. Psychological Bulletin, 1967, 67, 231-247. Hathaway, S. R. Problems of personality assessment. Proc. XIV International Congress of Applied Psychology, Munksgaard, Copenhagen, 1961.. Personality inventories. In B. B. Wolman (Ed.). Handbook of Clinical Psychology. New York: McGraw Hill, 1965. Hathaway, S. R. and McKinley, J. C. Manual for the Minnesota Multiphasic Personality Inventory. New York: Psychol. Corp., 1943. Heilbrun, A. B., Jr., Daniel, J. L., Goodstein, L. D., Stephenson, R. R., and Crites, J. 0. The validity of two-scale pattern interpretation on the California Psychological Inventory. Journal of Applied Psychology, 1962, 46, 409-416. Hill, R. E., Jr. Dichotomous prediction of student teaching excellence employing selected CPI scales. Journal of Educational Research, 1960, 53, 9, 349-351. Holland, J. L. The prediction of college grades from the California Psychological Inventory and the Scholastic Aptitude Test. Journal of Educational Psychology, 1959, 50, 1355-142.

146 BIBLIOGRAPHY (Continued) Jackson, D. Stylistic reponse determinants in the California Psychological Inventory. Educational and Psychological Measurement, 1960, 20, 339-346. Jackson, D. and Messick, S. Content and style in personality assessment. Psychological Bulletin, 1958, 55, 243-252. Jackson, D. and Messick, S. Acquiescence and Desirability as Response Determinants on the MMPI. Princeton, N.J.: Educational Testing Service, 1960. Johnson, M. C. Some hypotheses about the null. Session 72, Statistical Method and Theory, American Educational Research Association, 1967. Paper presented at Annual Meeting of the New York, February Johnson, R. T. and Frandsen, A. N. The California Psychological Inventory profile of student leaders. Personnel and Guidance Journal, 1962, 41, 343-345. Keimowitz, R. I. and Ansbacher, H. L. Personality and achievement in mathematics. Journal of Individual Psychology, 1960, 16, 84-87. Kelly, E. Lowell. California Psychological Inventory. In Buros, 0. K. (Ed.). The Sixth Mental Measurements Yearbook. Highland Park, N.J.: Gryphon Press, 1965, pp. 168-170. Krause, E. A. Dependency and the blind: family vs. therapeutic work setting. New Outlook for the Blind, 1962, 56, 553-357. Lebo, D. and Bruce, R. S. Projective methods recommended for use with the blind. Journal of Psychology, 1960, 50, 15-38. Lessinger, L. M. and Martinson, R. A. The use of California Psychological Inventory with gifted pupils. Personnel and Guidance Journal, 1961, 39, 572-575 -Levanthal, A. M. An anxiety scale for the CPI. Journal of Clinical Psychology, 1966, 22, 459-461.. Additional technical data on the CPI Anxiety scale. Counseling Psychology, 1968, 15, 479-480. Journal of Lichtenstein, E. and Bryan, J. H. CPI correlates of the need for

147 BIBLIOGRAPHY (Continued) approval. Journal of Clinical Psychology, 1966, 22, 453-455. Liddle, G. The California Psychological Inventory and certain social and personal factors. Journal of Educational Psychology, 1958, 49, 144-149. Lowenfeld, B. Psychological problems of children with impaired vision. In W. M. Cruickshank (Ed.). Psychology of Exceptional Children and Youth. Englewood Cliffs, N.J.: Prentice-Hall, 1955, 1963, 226-310. MacFarland, D. C. Social isolation of the blind. Journal of Rehabilitation, 1966, 32, 32, 49. MacKinnon, D. W. The personality correlates of creativity: A study of American architects. In G. H. Nielsen (Ed.). Proc. XIV International Congress of Applied Psychology. Copenhagen, 1961. Vol. II, 11-39. Copenhagen: Munksgaard, 1962. Maxfield, K. E. and Perry, J. D. The intelligence status of some vocational rehabilitation clients. New Outlook for the Blind, 1961, 55, 19-20. Maxwell, M. J. An analysis of the California Psychological Inventory and the American council on education psychological tests as predictions of success in different college curricula. American Psychologist, 1960, 15, 425 (Abstract). Megargee, E. I., Bogart, P., and Anderson, B. J. Predicting leadership in a simulated industrial task. Journal of Applied Psychology, 1966, 50, 292-295. Mitchell, J. V., Jr. and Pierce-Jones, J. A factor analysis of Gough's California Psychological Inventory. Journal of Consulting Psychology, 1960, 24, 455-456. Mizushima, K. and DeVos, G. An application of the California Psychological Inventory in a study of Japanese delinquency. Journal of Social Psychology, 1967, 71, 45-51. Nagera, H. and Colonna, A. Aspects of the contribution of sight to ego and drive development. Psychoanalytic Study of the Child, 1965, 20, 267-287 -

148 BIBLIOGRAPHY (Continued) Nichols, R. C. and Schnell, R. R. Factor scales for the California Psychological Inventory. Journal of Consulting Psychology, 1963, 27, 228-235. Norfleet, M. A. Personality characteristics of achieving and underachieving high ability senior women. Personnel and Guidance Journal, 1968, 46, 976-980. Nunnally, Jum C. Psychometric Theory, New York: McGraw-Hill, 1967. Peterson, D. R., Quay, H. C., and Anderson, A. C. Extending the construct validity of a socialization scale. Journal of Consulting Psychology, 1959, 23, 182. Pierce, J. V. Personality and achievement among able high school boys. Journal of Individual Psycholoy, 1961, 17, 102-107. Potter, G. S. A method for using the Minnesota Multiphasic Personality Inventory with the blind. In: W. Donahue and D. Dabelstein (Eds.). Psychological Diagnosis and Counseling with the Adult Blind. New York: American Foundation for the Blind, 1947. Pumroy, D. K. Relationship between the social desirability scale and the California Psychological Inventory. Psychological Reports, 1962, 10, 795-796. Query, W. T. CPI factors and success of seminary students. Psychological Reports, 1966, 18, 665-666. Rao, C. R. Advanced Statistical Methods in Biometric Research. New York: John Wiley and Sons, 1952. Raskin, N. J. Visual disability. In J. F. Garrett and E. S. Levine (Eds.). Psychological Practices with the Physically Disabled. New York: Columbia University Press, 1962, 341-375. Rawls, D. J. and Rawls, J. R. Personality characteristics and personal history data of successful and less successful executives. Psychological Reports, 1968, 23, 5, 1032-1034. Rosenberg, L. A., McHenry, T. B., Rosenberg, A. M., and Nichols, R. C. The prediction of academic achievement with the California Psychological Inventory. Journal of Applied Psychology, 1962, 46(6), 385-388.

149 BIBLIOGRAPHY (Continued) Rusalem, H. The environmental supports of public attitudes toward the blind. Outlook for the Blind, 1950, 44, 277-288. Sandler, A. Aspects of passivity and ego development in the blind infant. Psychoanalytic Study of the Child, 1963, 18, 343-368. Scholl, G. T. A vocational adjustment follow-up study of groups of the visually handicapped. Unpublished mimeo material prepared for conference, Ann Arbor, Mich. December 1969. Scholl, G. T., Bauman, M. K., and Crissey, M. S. A Study of the Vocational Success of Groups of the Visually Handicapped. Ann Arbor, Mich.: The University of Michigan, 1969. Schaie, K. W. The effect of age on a scale of social responsibility. Journal of Social Psychology, 1959, 50, 221-224. Schwartz, M. L., Dennerell, R. D., and Lin, Y. Neuropsychological and psychosocial predictors of employability in epilepsy. Journal of Clinical Psychology, 24, 2, 174-177. Scott, R. A. The Making of Blind Men. Hartford, Conn.: Connecticut Printers, 1969. _. The Making of Blind Men. New York: Russell Sage Foundation, 1969. Snider, J. G. Academic achievement and underachievement in a Canadian high school as predicted from the California Psychological Inventory. Psychology in the Schools, 1966, 3, 370-372. Southern, M. L. and Plant, W. T. Personality characteristics of very bright adults. Journal of Social Psychology, 1968, 75, 119-126. Strong, E. K., Jr. Strong Vocational Interest Blanks. Palo Alto, Calif.: Consulting Psychologists' Press, 1959. Springob, H. K. and Struening, E. L. A factor analysis of the California Psychological Inventory on a high school population. Journal of Counseling Psychology, 1964, 11, 173-179. Stein, K. B., Gough, H. G., and Sarbin, T. R. The dimensionality of the CPI Socialization scale and an empirically derived typology among delinquent and nondelinquent boys. Multivariate Behavioral

150 BIBLIOGRAPHY (Concluded) Research, 1966, 1, 197-208. Underberg, R. P., Verillo, R. T., Benham, F. G., and Cowen, E. L. Factors relating to adjustment to visual disability in adolescence. New Outlook for the Blind, 1961, 55, 253-259. Vingoe, F. J. Note on the validity of the California Psychological Inventory. Journal of Consulting and Clinical Psychology, 1968, 32, 6, 725-727 -Wachs, T. D. Personality testing of the handicapped: a review. Journal of Projective Techniques and Personality Assessment, 1966, 30, 339-3555 Watley, D. J. Career progress: a longitudinal study of gifted students. Journal of Counseling Psychology, 1969, 16, 100-108. Wright, B. A. Physical Disability-A Psychological Approach. New York: Harper and Row, 1960. Wright, G. N. and Trotter, A. B. Rehabilitation Research. Madison: The University of Wisconsin Press, 1968.

UNIVERSITY OF MICHIGAN 3 9015 02223 2303