Division of Research December 1984 Graduate School of Business Administration The University of Michigan AN EMPIRICAL INVESTIGATION OF THE MULTIPLE CONSTITUENCIES MODEL OF ORGANIZATIONAL EFFECTIVENESS Working Paper No. 406 Kim S. Cameron FOR DISCUSSION PURPOSES ONLY None of this material is to be quoted or reproduced without the expressed permission of the Division of Research. The research reported here was supported by a contract from the National Institute of Education (#400-83-0009). The comments of David Whetten, John Rohrbaugh, Ellen Chaffee, and Anne Tsui were especially helpful during the preparation of this manuscript.

e I~~ ~~ ~~~ ~~ ~~ ~~ ~~ ~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I I I i I r I

AN EMPIRICAL INVESTIGATION OF THE MULTIPLE CONSTITUENCIES MODEL OF ORGANIZATIONAL EFFECTIVENESS Abstract Increasing attention has been paid recently to the multiple constituencies model of organizational effectiveness, which defines an organization as effective If it minimally satisfies the demands of its strategic constituencies. This study represents the first Investigation of the basic propositions of this approach to effectiveness. It focuses on the extent to which effective organizations satisfy important constituencies, and If they do, which constituencies are most satisfied. Data was received from 1240 Individuals In 29 organizations In the United States. The multiple constituencies model received partial support from the findings, but an Important caveat must be acknowledged. The model may have validity for some circumstances but not others, and It may be applicable to some types of organizations but not others. A universalistic or uniqulvocal acceptance of this model should not be advocated.

AN EMPIRICAL INVESTIGATION OF THE MULTIPLE CONSTITUENCIES MODEL OF ORGANIZATIONAL EFFECTIVENESS Kim S. Cameron Graduate School of Business Administration The University of Michigan Ann Arbor, Michigan 48109 (303) 763-1179 September 1984 The research reported here was supported by a contract from the National Institute of Education (#400-83-0009). The comments of David Whetten, John Rohrbaugh, Ellen Chaffee, and Anne Tsui were especially helpful during the preparation of this manuscript.

AN EMPIRICAL INVESTIGATION OF THE MULTIPLE CONSTITUENCIES MODEL OF ORGANIZATIONAL EFFECTIVENESS Most recent theoretical writing on organizational effectiveness has emphasized the Importance of the different viewpoints represented by an organization's multiple constituencies. This paper reports the first attempt to empirically investigate the multiple constituencies approach to organizational effectiveness. A brief description of the multiple constituencies model is presented below followed by a description of a study to assess its fundamental assumptions. The Multiple Constituencies Approach to Organizational Effectiveness Many authors writing on organizational effectiveness have emphasized the subjective nature of effectiveness judgments (Cameron & Whetten, 1983; Keeley, 1984; Scott, 1981; Zammuto, 1982). That Is, judgments of effectiveness are argued to be products of personal values and preferences, therefore the criteria used by individuals to make those Judgments are ultimately subjective. For example, organizational goals, the most common source for criteria of effectiveness, are derived from negotiated tradeoffs of subjective preferences about what ought to be accomplished by organizations, to what extent it should be accomplished, and for whom goal attainment Is good (see Perrow, 1970; Pfeffer, 1977; Scott, 1981). By their very nature, the establishment and the evaluation of goals are subjective activities. Similarly, other sources of effectiveness criteria used In place of goals, such as survival, flexibility, resource acquisition, or efficiency are equally subjective. For example, survival is sometimes tauted as the ultimate objective criterion of effectiveness (Gibson, 1

Ivancevich, & Donnelly, 1979). However, survival as a criterion of effectiveness is as much a product of individual values and preferences as any other criterion. There Is no objective reason why survival should be a more Important ultimate end of organizations than the development of other attributes (e.g., flexibility). As Kaufman (1976) and Aldrich and McKelvey (1983) pointed out, It Is sometimes better both for the organization and for Its constituencies that it does not survive. Organizational demise Is preferred by some constituencies, while others prefer survival. (As an example, consider the controversy surrounding the closing of public schools In several cities, or the federal bailout of Chrysler as examples of survival belng a subjective criterion preferred by some but not by others.) This realization that the values and preferences of individuals must be tapped In order to assess effectiveness has led to the development of what has come to be called the multiple constituencies approach to effectiveness. This approach Is explicit In trying to account for the multiple and sometimes contradictory criteria used by different groups and Individuals when judging the effectiveness of organizations. Its underlying assumption is that organizations are effective to the extent to which they satisfy the preferences of constituencies (see, for example, Cameron & Whetten 1981; Connolly, Conlon, & Deutsch, 1980; Keeley, 1978; Miles, 1980; Pteffer & Salancik, 1978; Zammuto, 1982). Organizations that are most effective, according to this approach, are those that at least minimally satisfy relevant const i tuenc i es. There Is some controversy, however, about what it means to satisfy constituencies and which constituencies are the most important for the 2

organization to satisfy. A brief discussion of these two issues will help Introduce the research questions being pursued in this study. issues in the Multiple Constituency Approach Satisfaction has generally been equated with need fulfillment (Berger & Cummings, 1979; Costello & Lee, 1974, Cummings, 1982; Schmitt & Mellon, 1980). Individuals report a sense of satisfaction when they perceive that their needs have been met. Dlmarco and Norton (1974), LaFollette and Sims (1975), Lawler, Hall, and Oldham (1974), and others found that organization level effectiveness is associated with individual satisfaction. That is, when organizations perform in such a way as to fulfill individuals' needs, they tend to report high levels of satisfaction. In the multiple constituencies approach to effectiveness, constituencies are assumed to be satisfied when the organization does what constituencies want done, or when the performance of the organization is consistent with the preferences of the constituencies. As stated by Pfeffer and Salancik (1978): "Organizational effectiveness Is an external standard of how well an organization is meeting the demands of the various groups and organizations that are concerned with its activities (p. 11)." Determining the validity of this approach to effectiveness rests, then, on the ability to determine the match between an organization's performance and the preferences of constituencies. Nlsbet and Wilson (1977) and Slovic and Lichtenstein (1971) concluded, after extensive reviews of the empirical literature, that Individuals are not good at either identifying or explicating their preferences. Whereas constituencies actually hold preferences regarding organizational performance, it is very difficult for 3

researchers, and even for the individuals themselves, to determine what those preferences are and what Is their relative Importance. Just asking people for their preferences Is not likely to produce valid responses, as evidenced by a variety of empirical Investigations. Argyris and Schon (1978) suggested that Individuals hold two different kinds of theories or cognitive maps In their minds: theorles-in-action, which are used to guide behavior, and theories-in-use, which are theories that remain cerebral and are used to interpret the world. According to Argyris and Schon, theorles-in-action are seldom cognitively mapped, so that it is very difficult to Identify them merely by asking Individuals to describe them. Observable behavior is the primary source for identifying the theories-in-use held by other people. Similarly, the preferences that guide constituencies to conclude that an organization Is effective (I.e., theorles-in-action) are best analyzed by observing actual judgment behavior rather than asking for a list of preferences. Researchers of organizational effectiveness have not yet done that with multiple constituencies. The second issue in the multiple constituency approach is: which constituencies should the organization satisfy? This Issue is problematic because no organization can satisfy all constituencies. Not only are the preferences of different constituencies often contradictory, so that to satisfy one group may lead to the dissatisfaction of another (Dubin, 1976; Friedlander & Pickle, 1967; Whetten, 1978), but the relative Importance to the organization of various constituencies differs. For example, Pfeffer and Salanclk (1978) argue that judgments of organizational effectiveness should be 4

based on the satisfaction of the mast pawerful constituencies, or those controlling the critical resources needed by the organization. Keeley (1980) favors the least advantaged constituencies, or those most likely to be harmed by the organization. Zammuto (1982) favors the most heterogeneous mix of constituencies, and equates effectiveness with the ability to satisfy a broad mix of preferences. Van de Ven and Ferry (1980) argue for organizational members as the most important constituency, or those Inside the system who hold a stake In the organization's performance. Connolly, Conlon, and Deutsch (1980) suggest that no constituency should be favored above others, but that the satisfaction of any constituency Is sufficient for an organization to be effective. Although each of those authors differs In their perspectives on which constituencies should be satisfied for an organization to be effective, all argue that the preferences of constituencies should serve as the most appropriate criteria of effectiveness. It is just that whose preferences are most Important remains problematic. Some Investigators have Included Judgments from several constituencies In their studies of organizational effectiveness, but neither the authors listed above nor any other Investigators have conducted empirical assessments using the multiple constituencies approach. The purpose of this paper is to take a first step In Investigating empirically the assumptions upon which this approach Is based. Research Quest ons In order to investigate the multiple constituency model, the preferences of multiple constituencies had to be assessed to determine which preferences were most Important to which constituencies, and the 5

extent to which organizational performance was consistent with those preferences. Two major research questions guided this study. These questions center on the two major Issues surrounding the multiple constituencies approach. First, what preferences are held by the strategic constituencies of an organization, and how do the preferences of different constituencies differ from one another? Second, does the organization's effectiveness profile match one constituency's preference profile more closely than It matches others; if so, are they the most powerful constituencies, the least advantaged constituencies, or others? In addressing these two issues, this research aims to determine to what extent the multiple constituencies model has legitimacy and utility for assessing effectiveness in organizations. For example, the first Issue, to what extent do different constituencies actually hold different preferences of effectiveness for organizations, addresses an area of controversy in organizational theory. Several researchers have found that the judgments of effectiveness of different groups have low or negative correlations with one another (Dubin, 1976; Friedlander & Pickle, 1968; Whetten 1978). Others have found that individuals differ regarding their relative priorities when making personal judgments of effectiveness (Hltt & Mlddlemiss, 1978; Rohrbaugh, 1982). Other authors have suggested that constituency preferences reflect personal Interests and are therefore bound to conflict with other constituency preferences (Pfeffer & Salanclk, 1978; Scott, 1981). On the other hand, Baldridge (1981), Cameron (1978), Ramsay and Dodge (1983), Reimann (1982), and others have found that different constituencies hold highly similar perceptions and preferences regarding 6

organizational effectiveness. No research to date has assessed and compared the preferences of multiple constituencies with the actual performance of organizations at the organizational level of analysis, however, so the extent to which multiple constituency preferences differ has never been determined. The second Issue in the multiple constituency perspective is: Do organizations perform so as to satisfy the preferences of certain constituencies, and If so, which constituencies are they? This Issue focuses on the relative power of constituencies in Influencing organizational behavior. Miles and Cameron (1982), Miller and Frleson (1979), Miles and Snow (1978) and others have found evidence for inertia In the kinds of performance patterns displayed by organizations. This inertia has been given labels such as strategic competence, organizational saga, culture, structure, and performance stability. These forces of Inertia are relatively strong, and organizations have been found to persist In certain kinds of performance regardless of relatively severe pressures for change (see, for example, Miles and Cameron's account of the performance patterns consistently displayed by various tobacco firms in the face of extreme environmental pressures). On the other hand, Chandler (1977), Child (1972), Quinn and Cameron (1983), and others have found evidence that organizations respond to major constituency demands, and even change their definitions of effectiveness, because of those constituencies. These authors point to the dependency of organizations on their constituencies and the necessity of accounting for their preferences as major precursors of change. To date, however, no research has Investigated this relationship between constituency preferences for 7

effectiveness and organizational performance. It is not clear If organizations operate independent of, or In spite of, the preferences of strategic constituencies, or If they operate primarily in response to those demands. Furthermore, despite the different points of view mentioned above, no empirical research has addressed the Issue of which constituency is most appropriate in judging an organization's effectiveness, so these points of view all are based on authors' biases or suppositions. Which constituency's preferences are most closely allied to organizational performance Is still a matter of speculation. In the section below, the methods used for Investigating these Issues are described, followed by a report of the results of the study. The final section of the paper discusses the Implications of this Investigation for the multiple consituency approach to effectiveness and for future research on the effectiveness construct. METHODS OrganIzational Sample Twenty nine institutions of higher education In the northeast United States were selected for study. These Institutions are all medium size, liberal arts and comprehensive Institutions, and elite doctoral Institutions and proprietary schools were omitted from the sample. All are four-year schools, with seven offering only bachelors degrees, five offering masters degrees, and seventeen offering doctorates. Eleven of the schools are publicly supported and eighteen are private, Institutional age ranges from approximately 30 years to over 200 years. Faculties are unionized In nineteen of the schools with ten being non-unionized. Undergraduate student enrollments range 8

from just over 1000 to just over 10,000 with the average being 4200 students. Confidentiality was promised to each Institution, so names of schools are not included in this report. All data were collected from these Institutions during the 1980 academic year. Meyer and Rowan (1977), Scott (1981), and Thompson (1967) each pointed out that organizational effectiveness in this kind of organization Is best measured by social or opinion-type data. Because there is no ready-made, consensual standard available upon which to base judgments of effectiveness, perceptions of some constituency must be used. Scott (1981, p. 326) summarized this view as follows: When standards of desirability are themselves ambiguous, then the organization must resort to social tests. Social tests are those validated by consensus or by authority. Their validity depends on how many or on who endorses them. Institutionalized organizations...are likely to depend on social tests for their assessment of effectiveness. In this sample of Institutions, therefore, It was necessary to obtain data from a group of respondents associated with each Institut I on'rather than relying on some type of objective indicator, This Is because, as Scott pointed out, no objective Indicator exists in these organ I zations. Resapondent Samnple In each of the 29 Institutions, members of seven different strategic constituencies provided data. These seven groups are generally acknowledged to be members of the dominant coalition In Institutions of higher education (Cameron, 1978). They are (1) faculty department heads, (2) student affairs administrators, (3) financial 9

administrators, (4) academic administrators, (5) general administrators, (6) members of the board of trustees, and (7) representatives of major funders (e.g., senate finance committee members, foundation grant committee members, church budgetary committee members). Faculty department heads comprised the largest group In the sample with 40 percent, while major funders comprised only five percent of the sample. Fourteen percent of the sample were trustees, twelve percent were general administrators, ten percent were academic administrators, seven percent were financial administrators, and twelve percent were student affairs administrators. Approximately 75 Individuals were contacted at each Institution, and response rates for Institutions ranged from 35 percent to 68 percent. In all, 1240 Individuals participated In the study representing an overall response rate of approximately 60 percent of those contacted. Instruments Each respondent was requested to complete two different Instruments. These questionnaires were mailed to each individual along with a stamped return envelope. The first Instrument asked individuals to provide descriptive Information, not evaluative judgments, regarding the extent to which their institution possessed certain characteristics. These characteristics were found In previous research to be Indicative of Institutional effectiveness (Cameron, 1978, 1981, 1983). That -Is, correlations between the effectiveness scores produced by this Instrument and factors such as financial health, objective measures of performance, and objective institutional characteristics are statistically significant, Indicating evidence for the external validity of these Items. Respondents were not instructed that they 10

were rating effectiveness. They were only told that they were to describe the characteristics of their institution. Three examples of the 57 Items are presented below to Illustrate the descriptive nature of this first questionnaire. "How many faculty members and administrators at this college would you say serve in the community In government, on boards or committees, as consultants, or In other capacities?" "Estimate what percent of the graduates from this Institution go on to obtain degrees In graduate or professional school." "Approximately what proportion of the undergraduate courses offered at this college are designed to be career-oriented or occupation-related as opposed to liberal education, personal development, etc.?" This questionnaire assesses nine separate dimensions of organizational effectiveness, and these dimensions are summarized In Table 1. The result of this questionnaire is an effectiveness profile for each Institution at the organization level of analysis (I.e., each institution receives a score on each of the nine dimensions indicating the extent to which the Institution Is effective on each dimension). This profile can be thought of as the overall Institutional performance rating from the viewpoint of members of the dominant coalition. TABLE 1 ABOUT HERE The second Instrument was designed using the "policy capturing" or "social judgment analysis" methodology (Hammond, McClelland, & Mumpower, 1980; Hammond, KI ltz, & Cook, 1978; Hammond, Rohrbaugh, Mumpower, & Adelman, 1977; Hitt & Middlemiss, 1979; Huber, 1974; Rohrbaugh, 1979; Rohrbaugh, 1982; Slovic and Lichtenstein, 1971). This 11

Table 1 Nine Dimensions of Organizational Effectiveness in Institutions of Higher Education DIMENSION DEFINITION 1. Student Educational Satisfaction 2. Student Academic Development The extent to which students are satisfied with their educational experiences at the institution. The extent of the academic growth, attainment, and progress of students at the institution. 3. Student Career Development The extent of occupational preparedness of the students, and the emphasis on career development provided by the institution. 4. Student Personal Development The extent of student development in nonacademic, noncareer oriented areas, and the emphasis on personal development provided by the school. 5. Faculty and Administrator Employment Satisfaction 6. Professional Development and Quality of the Faculty 7. System Openness and Community Interaction 8. Ability to Acquire Resources 9. Organizational. Health The extent of satisfaction of faculty members and administrators with their employment at the institution. The extent of professional attainment and development of the faculty, and the emphasis on development provided by the institution. The extent of interaction with, adaptation to, and services provided for the external environment by the institution. The ability of the institution to acquire needed resources such as high quality students and faculty, financial supports, etc. The extent to which the internal processes and practices in the institution are smooth functioning, viable, and benevolent.

methodology accepts the assumption that Individuals have a difficult time identifying their own preferences and explicitly ordering them in order to map their judgments. Individuals called upon to Identify their preferences and to use them In judging organizational effectiveness often fall or are inaccurate in their attempts (Hoffman, 1960, Slovic, 1969). Constituency preference profiles, like cognitive maps (Bougon, Weick, & Binkhorst, 1977), therefore, are best identified by analyzing the actual judgments of effectiveness made by the constituency. To obtain these judgments, the second questionnaire was designed as follows. Individuals were presented with descriptions of the performance of 25 hypothetical institutions and asked to make overall evaluative judgments as to how effective each organization was (see Rohrbaugh, 1979, 1982). Each Institution had been assigned a particular score on each of the nine dimensions of effectiveness Identified In Table 1 (low, medium, high levels were presented on the dimensions). These dimensions were defined and Illustrated at the beginning of the questionnaire. Respondents were instructed that the dimensions had been found In previous research to be indicative of organizational effectiveness In higher education, and their task was to judge the overall effectiveness of each of the 25 Institutions presented on the questionnaire. Judgments were made on a scale of 1 (low) to 7 (high) according to the Institution's level of performance on a combination of the nine dimensions. Preference profiles, or weightings of the relative Importance of the different dimensions of effectiveness for each person, result from an analysis of these judgments. Figure 1 illustrates this procedure. 12

FIGURE 1 ABOUT HERE The difference between this questionnaire and the first one completed by the respondents Is that this questionnaire asks for evaluations and preferences; the first one asked for objective descriptions. In this questionnaire, individuals' responses were based on what they liked or valued; in the first questionnaire they did not express likes or values. The judgment analysis procedure uncovered attributes of Individuals (i.e., personal preference profiles); the first questionnaire uncovered attributes of organizations (I.e., Institutional effectiveness profiles). In summary, respondents produced two kinds of data from these two questionnaires. First, they responded to 57 Items that described characteristics of their own institution. These characteristics group together Into nine dimensions Indicating organizational effectiveness. Second, they evaluated the overall effectiveness of a set of (hypothetical) Institutions based on a diagram of their performance. This judgment task produced profiles of constituencies' own preferences (i.e., which criteria of effectiveness they most prefer). It is Important to note that ratings (from the first questionnaire) are not consistently related to actual judgments (from the second questionnaire). Sometimes ratings are positively correlated with actual judgments, and sometimes the correlations are negative (Slovic & Lichtenstein, 1971; Nlsbet & Wilson, 1977). Individuals are sometimes especially critical in their evaluations of criteria that they highly value, and sometimes they are especially lenient and forgiving. Some individuals may tend to rate effectiveness especially 13

Figure 1 An Example of the Judgment Task Given to Respondents on the Second Questionnaire INSTRUCTIONS: In this questionnaire, we are asking that you make judgments about the overall effectiveness of 25 colleges and universities.... We have presented the "performance profiles" of these institutions based on their scores on the nine dimensions of organizational effectiveness [defined above]. In order to judge the effectiveness of each institution, you will need to pay particular attention to the scores of each dimension. Some dimensions of performance may be more important to you than others in determining your overall evaluation, so please note the different scores on each dinension before making your judgment.... Please rate each school on a scale of '1' (ineffective) to '7' (highly effective). Highly effective Somewhat effective Ineffective (1) (7) (6) (5) (4) (3) (2) Student Educational Satisfaction Student Academic Devlotpmnt Student Career evelopmnt Student Persoal Oevelopmnt Faculty & Adlnistrator Sat1sfaction Pr tssionl Developmnt i Qual ty of Faculty Systm Opemness C Commnity Interaction Ability to Acquire Isources Organiational iHalth — mm _ -__. momi m mu Iw MTISG OF FECTnVESS, S OL 23: Stdent Educational Satisfction Student Academc Developnt Student Career Developeint Student Personal Development faculty i Administrator Satisfaction P 0ft 1 sonal Developwnt & Qulity of Faculty Systm Openness I Cao nity Interaction Ability to Acqui Resources Organiational Healt -mm m mu mml M ---e sMTIIG OW UFM IVESS, SCM0L 24:

low on criteria they value, whereas other Individuals may tend to rate effectiveness especially high. Because of this divergence, there is no reason to expect the data from the two questionnaires to be related In any a priori way. That is, there Is no reason to assume that respondents will either rate an institution to be highly effective on the criteria they prefer (a positive correlation), or that they will rate Institutions low on the criteria they prefer (a negative correlation). One section of the first questionnaire also asked for Information regarding respondents' perceptions of the relative power and importance of several constituencies as they affected the performance of the Institution. These ratings were used In analyses to be explained later. Ratings of relative power were received In a Llkert format for students, faculty, top adminstrators, alumni, trustees, major funders, and legislators. Anaivses Several different analyses were required In order to address the two major research questions. -On the first questionnaire, psychometric and analysis of variance tests helped determine the reliability and discriminant validity of the organizational effectiveness profiles for the 29 schools. On the second questionnaire, multiple regression and reliability analyses helped Identify preference profiles for the seven constituency groups. These analyses In combination addressed the first major research question. The second research question was addressed by comparing the organizational effectiveness profiles (first questionnaire) with the preference profiles (second questionnaire) by 14

means of correlation analysis. More detail regarding these analyses is reported in the results section. RESULTS Questonnai re 1 It was expected that nine separate dimensions of effectiveness would emerge from this questionnaire (Cameron, 1978, 1981, 1982), so tests were conducted to see If those dimensions did In fact occur. The results confirm that the same dimensions were present in this study as have been identified in previous studies. Internal consistency reliabilities for these dimensions ranged from.72 to.92 with a mean reliability coefficient of.82. Factor analysis of the 57 questionnaire items also resulted In the dimensions loading on their own factors. Average Intercorrelation among the nine dimensions was.42 indicating that, whereas the dimensions are conceptually distinct, certain of the dimensions do vary together in ratings of effectiveness (see Cameron [1981J for an analysis of the Interdimensional coveriance). These results Indicated that the nine dimensions of organizational effectiveness have adequate Internal consistency reliability and discriminant validity to be used as the basis for the Institutional performance profiles. An examination of the mean scores of each of the 29 Institutions across the nine dimensions showed that each school had a unique-profile of effectiveness scores, and no school scored high (or low) on all the dimensions of effectiveness. 1Results of the factor analyses using both orthogonal and oblique rotations are not described here In detail In order to conserve space, but they are available from the author. 15

It was also important to make certain that constituencies responded to characteristics of the institution and not merely as a result of their occupying a certain job or position. To make that determination, the ratings of each of the seven constituencies were analyzed by means of MANOVA and ANOVA. Significant differences were found among these groups In their ratings of the nine dimensions. Table 2 presents the results of these analyses for both Institutional and constituency membership effects. Significant effects were found on all dimensions for the Institution factor, indicating that the Institution being judged made a difference In the ratings obtained. Percent of variance accounted for by Institution ranged from.16 to.49 (mean R2 =.32) Indicating that this was a powerful factor In determining the ratings that were given. Of the total variance accounted for by the two factors in the ANOVA models (I.e., the Institution factor and the group membership factor), the institution factor accounted for the large majority of the explained variance. TABLE 2 ABOUT HERE Significant effects also were found on six of the nine dimensions for the constituency membership factor. The percent of variance accounted for by group membership, however, ranged from.00 to.04 percent (with an average of.015 percent), indicating that it was not a very important factor In the ratings given to the Institutions, What was more Important for the testing of the multiple constituencies model Is that the average differences among the groups of constituencies were not large enough to affect the rank orderings of the dimensions. That 16

Table 2 The Effects of Institution and Constituency Group Membership on Ratings of Organizational Effectiveness MANOVA Institution Group Interaction d.f. F d.f. F d.f. F 252,6396 6.91* 45,3730 2.47* 1188,6690 1.06 Dimension of Total Institutiqn Group Interactin Effectiveness*** R F R F R2 F R.. _~~~~~ 1 2 3 4 5 6 7 8 9.52.56.45.38.41.48.31.58.38 14.63* 18.57* 8.57* 7.31* 7.51* 11.59* 4.73* 19.46* 6.17*.42.45.30.24.28.37.16.49.21 5.13* 6.81* 0.91 0.61 4.15* 0.45 3.53** 6.74* 7.89*.02.02.00.00.02.00.02.02.04 1.01 1.07 1.23 1.16 1.09 1.15 1.01 1.09 1.12.08.09.15.14.11.11.13.07.13 * P<.001 ** p<.01 *** Numbers of dimensions of effectiveness are the same as those in Table 1.

Is, when rank orderings of the dimensions of effectiveness were produced for each of the seven constituencies (based on the constituencies' ratings resulting from the first questionnaire), all constituency ratings held the same rank order. This suggests that agreement among constituencies was sufficient to use a rank ordering method In comparing institutional effectiveness profiles with constituency preference profiles. The MANOVA results confirmed that, across all nine dimensions combined, both institution and group membership effects are significant In Influencing ratings of effectiveness. No significant interaction effect occurred between these two factors on any single dimension (ANOVAS), or when considering all dimensions together (MANOVA), however, Indicating that institution and group memberships Influenced ratings of effectiveness Independently. To summarize, these two analytic procedures —psychometric tests of the nine dimensions of effectiveness, and MANOVA and ANOVAs of the different constituency ratings —revealed, first, that Institutions have unique profiles across the nine dimensions of effectiveness, and second, that different constituency groups rate effectiveness somewhat differently. However, the average differences among constituencies were not great enough to change the relative rank orderings of the nine effectiveness dimensions. These (rank ordered) effectiveness profiles for the institutions were next used in comparison with the preferences of the various constituency groups in investigating the multiple constituencies model of effectiveness. That model suggests that different constituencies hold different sets of preferences for the organization, and when 17

organizational performance matches those preferences, the organization Is said to be performing effectively because of its satisfied constituencies. If organizational performance differs from a constituency's preferences, ineffective performance Is a more likely conclusion of that constituency. The Investigation of this assumption is the aim of the analyses of the second questionnaire. Quest onna re 2 Analyses of the second questionnaire were designed to determine, first, If individuals within a constituency group all tended to hold similar preferences (so that they could be treated as a group), and second, If the different constituency groups held different or similar preference profiles. In order to answer research question 1, it had to be determined if the respondents could actually be categorized into the appropriate constituency groups and If the different constituencies' preference profiles could be Identified. Preference profiles were obtained for each of the respondents by means of the "judgment analysis" methodology. In simple terms, this methodology uses the following logic. When Individuals make judgments about effectiveness, certain criteria are considered to be more Important than others. Usually the relative Importance of the criteria are not consciously known by individuals. When respondents make a large number of Judgments about effectiveness, a researcher can make the criteria upon which those judgments are based more or less Important by giving them different weights. This variance In weights Is expected to produce different overall Judgments of effectiveness by the respondent. By regressing the overall judgments produced by the respondent on the weighted criteria of effectiveness, a regression 18

coefficient results for each criterion which shows its relative Importance to the respondent in making judgments about effectiveness. In this study, the judgment criteria were the nine dimensions of effectiveness presented at different levels of achievement (e.g., low, medium, high). Research uest I on 1 The first analysis of this questionnaire addressed the question: to what extent do Individuals within the various constituency groups hold similar preference profiles? That Is, can individuals be treated as members of constituencies, or must they be analyzed individually? Reliability analyses were conducted on the preference profiles of Individuals within each of these constituency groups to determine the homogeneity of each group. Judgments of effectiveness for the 25 separate cases served as data for the analyses. Table 3 Indicates that the internal consistency reliability coefficients are high within each group (i.e., preference profiles for Individuals within each group are similar). In addition, the MANOVA indicates that significant differences exist between groups in their preference profiles. TABLE 3 ABOUT HERE (Follow-up contrasts between groups are not possible with multiple dependent variables. However, in 18 of the 25 hypothetical schools being rated by the different constituencies, group membership had a significant effect on the judgments.) These results suggest that treating these constituency groups as separate entitles In the remainder of the analyses Is entirely appropriate. The groups hold 19

Table 3 Internal Consistency and Discriminant Validity of the Preference Profiles of Seven Different Constituency Groups CONSTITUENCY Alpha Value General Administrators Academic Administrators Financial Administrators Student Affairs Administrators Faculty Department Heads Board of Trustee Members Major Funders.82.86.78.85.87.91.81 MANOVA Source d.f. F Value Significance Constituency Group Membership 150,2350 2.18.0001

unique preference profiles compared to one another, yet the individuals within each group hold similar profiles. With distinct constituencies able to be Identified, the analysis next focused on the differences among those groups. Table 4 reports the results of the analyses, with the preference profile of each constituency presented in rank order form (I.e., the dimension most valued by the constituency is listed at the top; the least valued dimension Is listed at the bottom). An examination of the table suggests that differences do exist among the constituency groups in the extent to which they value the various criteria of effectiveness. TABLE 4 ABOUT HERE There are two ways to analyze the results In Table 4. One is to consider the rank order correlations among the groups to determine if groups had similar rank orderings. The second is to test for equality of regression equations using a procedure developed by Morrlson (1983). Rank order correlations among the groups range from.22 to 1.0 with the mean correlation coefficient being.57. (General administrators and trustees had Identical rankings, as did academic administrators and faculty department heads.) All constituencies held the four dimensions composed of student development criteria to be of higher Importance than the dimensions comprising more organizationally related criteria (e.g., Organizational Health, System Openness and Community Interaction). Which student-centered dimension was held to be most Important differed among the groups, however (for example, general administrators valued Student Educational Satisfaction most highly, whereas academic administrators valued Student Academic Development 20

TABLE 4 Preference Profiles for Seven Constituency Groups in the Form of Rank Orders of Nine Dimensions of Organizational Effectiveness General Academic General Administrators Dimension Beta* Academic Administrators Dimension Beta...... Financial Administrators Dimension Beta Student Affairs Administrators Dimension Beta 1** 2 3 6 4 5 8 9 7.497.475.373.371.343.276.216.125.117 2.568 1.472 6.394 3.384 4.289 5.278 8.198 9.097 7.079 F = 35.88** R2 =.89 1.530 2.522 3.424 5.341 4.307 6.285 8.200 7.121 9.093 F = 44.14** R2 =.91 2.536 3.472 9.409 4.405 6.301 1.271 5.166 8.104 7.016 F 35.79** R2 =.89 F2 30.31** R2=.87 Faculty Department Heads Board of Trustee Members Major Funders Dimension Beta Dimension Beta.... _... Dimension Beta 2.565 1.459 6.411 3.406 4.309 5.249 8.190 9.118 7.065 F = 37.65** R2 =.89 1.521 2.495 3.424 6.357 4.348 5.254 8.185 9.086 7.079 F = 35.29** R2 =.89 3.561 4.477 2.396 1.317 6.284 5.240 8.166 9.013 7.006 F = 14.06** R2 =.76 * Betas greater than.12 are significant at the.05 level. ** p <.001 *** Dimension numbers refer to the dimensions listed in Table 1.

most highly). In general, this analysis shows that there are five preference profiles held by these seven groups. These profiles differ In only a slight to moderate extent In their relative rank orderings. The high R2 statistics for each group (ranging from.79 to.91) indicates that the nine dimensions of effectiveness used as criteria for judgment account for the large majority of the variance In these groups' preferences regarding effectiveness. The second analysis of Table 4, an examination of the equality of regression equations, was conducted using three different tests. The first test examined the homogeneity of the regression planes, or the extent to which the vectors of coefficients were significantly different from one another. The resulting F Is 1.06, p>.10, indicating that the regression planes are not significantly different. The second test examined the equality of the Intercepts. That is, given differences In the Independent variables (scores on the nine effectiveness dimensions), did respondents give significantly different overall ratings of effectiveness? The F is 12.8, p<.000, Indicating a significant intercept difference among the groups. The third test, a combination of regression planes and Intercepts, called a test of homogeneity of relationships, resulted In an F of 2.87, p<.000, Indicating that the combination of regression planes and intercepts produces a significant difference among the groups as well. In sum, these tests suggest that some significant differences result when comparing the preference profiles of certain constituency groups. When considering the vector of regression weights In a linear combination, no significant differences arise. However, when considering differences In ratings of effectiveness produced by the 21

different constituencies, significant differences are exhibited. With reference to research question 1, It can be tentatively concluded that constituencies hold somewhat different preference profiles for Institutions of higher education. The rank orderings of preferences vary only slightly among several of the groups, but more among others. The combination of weightings assigned to the effectiveness dimensions do not differ significantly, but the level of the ratings of overall effectiveness given to institutions that hold similar profiles of achievement on those dimensions do differ significantly among the groups. The answer to research question 1, therefore, Is that constituencies hold Identifiable preferences for Institutional performance, and that those preferences differ somewhat from other groups. The differences arise from variation In evaluations of effectiveness given by the different groups when exposed to the same objective performance data. Research Question 2 Investigating the second major research question Involved comparing the effectiveness profiles of the seven constituency groups In each school. This was done after rank ordering the effectiveness dimensions on which the Institutions performed, highest to lowest, and rank ordering the dimensions that each constituency most preferred, highest to lowest. Rank order correlations were then computed for each of the 29 Institutions comparing their effectiveness profiles with the preference profiles of the seven constituencies within each institution. Table 5 presents the results of these correlations. 22

TABLE 5 ABOUT HERE Recall that on the second questionnaire, respondents judged the overall effectiveness of 25 hypothetical institutions. When all members of a particular constituency group (In an institution) failed to judge the effectiveness of all 25 cases, that constituency was dropped from the analysis for that institution. That is, the preference profiles of only those groups in each Institution that judged all 25 cases were Included in this part of the analysis. In order to conclude that an institution's performance profile matched a constituency's preference profile, some decision had to be made regarding the magnitude of the correlation coefficient to be accepted. A coefficient of.4 was used because its statistical significance was p<.l. (Because of a small number of cases [i.e., N=9], correlations of r<.6 were needed to reach the.05 level, and only nine of the 192 correlations in Table 5 reached that level.) Moreover, the.4 level appeared to be the dividing line between high and low correlations.2 Two major findings relative to the second major research question emerged from these analyses. First, some schools tended to satisfy multiple constituency preferences while other schools satisfied no constituencies. That is, correlations between the performance profiles of some schools and the preference profiles of several of their 2Selecting a correlation coefficient other than the one Indicating the traditional p<.05 level was Justified because of the low probability of observing large correlation coefficients In this data set. The small N coupled with the restriction of variance in rank order correlations served as the justification for selecting the observed dividing point (I.e., r=.4) between high and low correlations as the comparative standard. 23

Table 5 Rank Order Correlations Between Performance Profiles of 29 Institutions and the Preference Profiles of 6 Constituency Groups Within Each Institution School General Academic Financial St. Affairs Dept. Trustees Funders Number Admin. Admin. Admin. Admin. Head 1.000.383 -.450* -.233.200 2.350 -.033 -.017 -.067.383.151.233 3 -.603** -.046.226 -.209.000 4.450* -.333.033.400*.217 -.517*.350 5 -.050.483*.150.383 -.133 -.217.533* 6 -.083 -.533*.067 -.083 -.183 -.083 -.067 7.250 -.183.250.517*.233.483* 8 -.050.483*.150.383 -.133 -.217.533* 9.733**.667**.283.633**.567*'.017 10 -.084.167.300.250.283.300 11.417*.550*.650**.183 -.067 12.050.183.050 -.267.100 -.200 -.283 13.250.583*.417*.150.067.033.243 14.352.084 -.117.134.092.210.059 15 -.467* -.100 -.233.283 -.050.350.000 16 -.083.067 -.050.033.317.017 17.533*.317.117.000.133.233.267 18.150.383.517* -.267.100 19 -.017.033.267 -.033.217 -.200 20 -.267.167.100.017 -.250.383 21.017 -.033.250 -.017 -.100.100 22.033.467*.033.517*.267.017.417* 23.033.467*.033.517*.267.017.417* 24.634**.600**.317.100.550* 25.150.167 -.283.167.183 -.083.017 26 -.033 -.300.267 -.150.183 -.116 27.150.217.250.460* -.050.500* 28.150 -.033.600**.550*.567*.417*.500* 29.500*.517*.717** -.183.300 * p<.10 ** p<.05

constituency groups were high, but for other schools the correlations were low or negative for all groups. The multiple constituencies model assumes that organizations that satisfy the preferences of important Interest groups (I.e., whose effectiveness matches the preferences of constituencies) are, by definition, more effective than those that do not. To determine If that assumption has validity in this sample, an overall score across all nine dimensions of effectiveness was computed for each school. This represented a summary (mean) effectiveness rating for each institution. A comparison between the fourteen schools that satisfied no constituencies and the seven schools that satisfied multiple (three or more) constituencies revealed that schools with profiles consistent with preferences of multiple constituencies all were In the top half of the total sample In their overall effectiveness, while those schools that satisfied no constituencies all were In the lower half of the sample In overall effectiveness. A comparison of the mean effectiveness score of the schools with compatible profiles versus the schools without compatible profiles resulted In a significant difference at the p<.05 level. That Is, schools that satisfied multiple constituencies were more effective on the nine dimensions of organizational effectiveness than were schools that satisfied no constituencies. Moreover, In comparing the top ten schools in performance with the bottom ten schools, nine of the top ten schools satisfied multiple constituencies while eight of the schools in the bottom ten satisfied no constituencies. The implication is that highly effective schools satisfied multiple constituencies; ineffective schools satisfied no constituencies. 24

The second major finding refers to the extent to which the organization's effectiveness profile matches one constituency's preference profile more closely than others. Table 6 records the total number of high positive correlations between the seven constituency groups' preference profiles and their own institution's effectiveness profile. This table suggests that the groups with the most compatible profiles are academic and student affairs administrators, followed by general administrators (each of which is an internal group). The external constituency groups (I.e., trustees and major funders) and faculty department heads have less compatible profiles (i.e., are likely to be less satisfied) than other groups. This result holds even when dropping out those Institutions that had one or more constituencies missing from the analysis. TABLE 6 ABOUT HERE Given this difference In the extent to which different constituencies are likely to be satisfied by Institutional performance, the second part of the research question became particularly relevant: which of the constituencies tended to be most satisfied-the most powerful, the least advantaged, Internal groups, external groups, or others? On the first questionnaire (I.e., the questionnaire that produced descriptions of effective organizational characteristics), a rating was produced for the relative power of nine major constituencies In affecting Institutional performance. Unfortunately, those groups do not match exactly the seven constituency groups used on the second questionnaire, but there Is enough overlap to suggest an answer to this 25

Table 6 Total Number of High Positive Correlations (p<.10) Between a Constituency's Preference Profile and Their Institution's Effectiveness Profile Constituency Total Number of High Positive Correlations General Administrators 8 Academic Administrators 10 Financial Administrators 3 Student Affairs Administrators 9 Faculty Department Heads 4 Trustees 4 Major Funders 5

I part of the research question. The constituencies that were rated on the first questionnaire according to their relative power are listed in Table 7 along with their mean ratings (on a 1 to 7 scale) and rank orderings. (Note that Top Administrators is a category that comprises four of the constituencies used In producing preference profiles, namely general, academic, financial, and student affairs administrators.) TABLE 7 ABOUT HERE Table 7 results suggest that the constituencies rated as most powerful are a so the most like I y to have compatibl b e preferences with institutional effectiveness. In this case, Top Administrators are both the most powerful and the most likely to be satisfied because of their compatible profiles. Major Funders, on the other hand, are next most likely to be satisfied even though they are rated as the least powerful of all the nine constituencies. Faculty and Trustees are both less satisfied and less powerful than are Administrators, but they are more powerful than Major Funders, while being less satisfied. Table 7 also contains rankings produced In a study by Gross and Grambsch (1974) wherein the relative power of several constituencies was rated by 7200 faculty and administrators. Rankings of constituencies In their study are similar to those produced in this study, and similar conclusions seem merited. That Is, In terms of the research question, these data suggest that both the most powerful and the least powerful constituencies tend to have compatible profiles more than constituencies ranked in the middle in relative power. Major funders are not likely to be as satisfied as the top administrator 26

Table 7 Ratings of Relative Power and Influence of Different Constituencies Compared Satisfaction of Those Groups to the Preference Constituency on Mean Rating of Rank Order of Constituency on Rank Order of Pref- Gross & Grambsch (] Questionnaire 1 Relative Power Relative Power Questionnaire 2 erence Satisfaction Constituencies/Rank Top Administrators 6.02 1 Academic Administrators 1 Top Administrators Student Affairs Admin. 2 General Administrators 3 Faculty 5.39 2 Faculty Department 5 Faculty Heads Department Heads Trustees 5.07 3 Trustees 5 Trustees Students 4.54 4 Students Accreditation Teams 4.24 5 State Legislatures 3.65 6 State Legislatures Alumni 3.52 7 Alumni Federal Government 3.31 8 Federal Government Major Funders 3.20 9 Major Funders 4 Major Funders [974) tings 1 3 5 2 5 7 9 6 7 we >

groups, but they are likely to be satisfied more than other groups rated as being more powerful. DISCUSSION AND CONCLUSIONS This study has been the first attempt to investigate empirically the multiple constituencies model of organizational effectiveness. Up to now the model has been a product only of theoretical perspectives (e.g., Connolly, et al., 1980; Keeley, 1978; Pfeffer & Salancik, 1978; Zammuto, 1982) and not of empirical results. Because more and more authors are beginning to advocate the multiple constituencies model when they consider effectiveness, it Is Important that some Investigations be made of Its legitimacy and utility. Cameron and Whetten (1983) pointed out that the multiple constituencies model is not necessarily the best nor the most legitimate model of effectiveness available because no one model can possibly be appropriate in all circumstances, for all organizational types, for all purposes, on all levels of analysis, and so forth. A variety of models and definitions of organizational effectiveness are needed so that the most appropriate model can be found for different conditions. However, any model that Is to be useful to practitioners and researchers must be shown to be valid and based on reasonable assumptions. That has been the purpose of this Investigation. To test the assumptions of the multiple constituencies model, at least two conditions had to be present. First, organizations had to be studied with multiple constituencies that were Invested In and Informed about organizational performance. These constituencies had to be familiar enough with the functioning of their own organization to be 27

able to assess Its effectiveness. Institutions of higher education were selected for study because they have multiple constituency groups that meet those conditions. Second, two different assessments were required In order to compare constituency preferences with organizational functioning. An effectiveness profile had to be obtained for individual organizations, and, a preference profile had to be generated for different constituencies of those organizations. The effectiveness profile and the preference profiles had to be based on the same criteria so comparisons among profiles could be made. This second requirement was met by basing the comparisons of constituency preferences and organizational functioning on the nine dimensions of effectiveness found by Cameron (1978, 1983b) to be indicative of organizational effectiveness, and by measuring those criteria via instruments with no a priori relationship. Each of 29 institutions was found to hold unique effectiveness profiles based on these dimensions, and five somewhat unique preference profiles were found among seven major constituencies. The results of this Investigation are supportive of the multiple constituencies model, but they are not unequivocal. Unqualified support for the model was not obtained. For example, It was found that respondents' preference profiles were highly compatible with other members of their constituency, and the constituencies preference profiles did vary somewhat from one another. However, the rank order correlations among the different profiles were moderate to high, and one of the three tests of homogeneity of preference profiles found the preference wenightng not to be significantly different among groups. 28

Significant differences were present among the groups relative to the magnitude of their preference profiles, however. The more crucial results regarding the multiple constituency model, however, relate to the second major research question: do organizations satisfy one constituency more than others? Satisfying a constituency in the multiple constituencies model means performing consistently with the preferences of that constituency. Strong and numerous matches between constituency preference profiles and organizational effectiveness profiles did not result. It is true that the organizations satisfying the most constituencies were more effective on the average than organizations satisfying no constituencies (a finding consistent with the multiple constituencies perspective), but some conspicuous aberrations were found to this general pattern. For example, one school In this sample (school #21) was third highest In overall effectiveness (I.e., in Its overall mean when the nine dimensions were combined), and highest In the sample on two of the nine Individual dimensions of effectiveness. Yet, the correlations between its performance profile and the preference profiles of Its constituencies ranged from.25 to -.10, with a mean correlation of.04. It also must be kept In mind that almost half the organizations satisfied none of these major constituencies, yet they still survive. This finding Is contrary to the multiple constituencies perspective. Evidence also was found suggesting that the most powerful constituencies have the most compatible preference profiles relative to Institutional functioning. But the least powerful constItuencies also have more compatibility than do other more powerful groups. Because of 29

the relatively small number of high positive correlations between constituencies' preference profiles and organizations' effectiveness profiles, no strong conclusions can be drawn. In general, support was found for the multiple constituencies model of organizational effectiveness, but the support Is not overwhelming. There Is reason to believe that empirical support for the multiple constituencies model of effectiveness would be difficult to find. One assumption of the model Is that organizations should operate so as to satisfy the preferences of certain constituencies In order to obtain critical resources and perpetuate organizational survival. This perspective rests on the assumption that organizations survive (and function effectively) as a result of the support provided them by constituencies. "... organizations survive to the extent that they are effective. Their effectiveness derives from the management of demands, particularly the demands of Interest groups upon which the organizations depend for resources and support" (Pfeffer & Salanclk, 1978, p. 2). A contrary point of view, however, and one supported by some empirical evidence, Is that factors such as organizational inertia (MIIler & Friesen, 1980), negotiated tradeoffs and politics (Hannan & Freeman, 1978), a tendency to satlsfice In organizational action (March & Simon, 1958), organizational saga and cultures (Clark, 1970), or life cycle dynamics (Cameron & Whetten, 1983), all operate to influence organizatlonal functioning and survival as much as, or more than, constituency groups. This point of view suggests that organizations' effectiveness can be largely Independent of constituency Influences, so that high positive correlations between what constituencies want from 30

I an organization and what the organization provides may not occur. This argument Is similar to the continuing debate between the strategic choice theorists (Child, 1973; Miles & Cameron, 1982), who argue that managers and organizational participants can change and Improve organizational functioning by their actions, and the population ecologists (Hannan & Freeman, 1977; Aldrich, 1979) who argue that organizations operate and survive Independent of managerial or participant action or Intent. Evidence has been amassed to support both sides of that argument, and In truth, both sides are likely to be at least partially correct. Depending on the type of organization, the environment, the time frame, and other such factors, both managerial action and uncontrollable inertia Influence organizational effectiveness. What this points out in regard to this study is that equivocal results may be the best that can be hoped for. Some organizations are affected by constituency preferences sometimes but not others, and additional powerful Influences besides constituency preferences must be considered In accounting for organizational effectiveness. The overall conclusion that emerges from this discussion, then, Is that the multiple constituencies model may have validity in some circumstances or for some organizations and not others, and practitioners and researchers must be cautious not to advocate the unequivocal or universal acceptance of the model. Organizational linkages to constituency demands and preferences are not universally consistent. Informed choices regarding which model Is most appropriate In which circumstance, therefore, Is a critical prerequisite for good research 31

on organizational effectiveness (see Cameron, 1984, for a discussion of when various models may be most appropriate). 32

REFERENCES Argyris, C., and Schon, D. Organizational learning! A theory of action perspective. Reading, Mass.: Addison-Wesley, 1978. Bluedorn, A. C. Cutting the gordlan knot: A critique of the effectiveness tradition In organization research. Sociology and Social Research, 1980, 64, 477-496. Bougon, M.; Weick, K.; and Binkhorst, D. Cognition in organizations: An analysis of the Utrecht Jazz Orchestra. Administrative Science Ouarterly, 1977, 22, 606-639. Cameron, K. S. Measuring organizational effectiveness In institutions of higher education. Administrative Science Quarterly, 1978, 2., 604-632. Cameron, K. S. Critical questions in assessing organizational effectiveness. Organizational Dynamics, 1980, 2, 66-80. Cameron, K. S. Domains of organizational effectiveness In colleges and universities. Academy of Management Journal, 1981, 2A4 25-47. Cameron, K. S. The effectiveness of Ineffectiveness. In Barry Staw and Larry Cummings (Eds.), Research in Organizational Behavior. Vol. 6, Greenwich, Conn.: JAI Press, 1983. Cameron, K. S. A study of organizational effectiveness and Its predictors. Discussion Paper Series. National Center for Higher Education Management Systems, Boulder, Colorado, 1983b. 33

I Cameron, K. S., and Whetten, D. A. Organlzational effectiveness! A comparison of multiple models. New York: Academic Press, 1983. Cameron, K. S., and Whetten, D. A. Perceptions of organization effectiveness across organizational life cycles. Administrative Science Ouarterly, 1981, 2t. 525-544. Campbell, J. On the nature of organizational effectiveness. In P. S. Goodman and J. M. Pennings (Eds.), New Perspectives on Organizational Effectiveness, San Francisco: Jossey-Bass, 1977. Clark, B. R. The distinctive college. Chicago: Aldlne, 1970. Connolly, T., Conlon, E. M., and Deutsch, S. J. Organizational effectiveness: A multiple constituency approach. Academy of Management Review, 1980, J. 211-218. Dubin, R. Organizational effectiveness: Sane dilemmas of perspective. Organization and Administrative Sciences, 1976, 2, 7-14. Etzloni, A. A comparative analysis of complex organizations. Revised Edition. New York: Free Press, 1975. Friedlander, F., and Pickle, H. Components of effectiveness in small organizations. Administrative Science Quarterly, 1968, 1, 289-304. Goodman, P. S.; Atkin, R. S.; and Schoorman, D. On the demise of organizational effectiveness. In K. S. Cameron and D. A. Whetten (Eds.), Organizational effectiveness! A comparlson of multiple models. New York: Academic Press, 1983. 34

Gross, E., and Grambsch, P. V. Changes In university organlzationn 1964-1971. New York: McGraw-Hill, 1974. Hall, R. H. Effectiveness theory and organizational effectiveness. Journal of Applied Behavioral Science, 1980, 16, 536-545. Hammond, K. R.; Klitz, J. K.; and Cook, R. L. How systems analysts can provide more effective assistance to the policy maker. JournaL of Anplied Systems Analysis, 1978, 2, 111-136. Hammond, K. R.; McClelland, G. H.; and Mumpower, J. Human iudnment and decision making. Theoriesr methods. and procedures. New York: Praeger, 1980. Hammond, K. R.; Rohrbaugh, J.; Mumpower, J.; and Adelman, L. Social judgment theory: Applications in policy formation. In M. Kaplan and S. Schwartz (Eds.), Human Judgment and Declsion Processes! Applications In Problem Settins, New York: Academic Press, 1977. Hitt, Michael A., and Middlemiss, R. Dennis. A methodology to develop the criteria weightings for assessing subunit effectiveness In organizations. Academy of Management Journal, 1979, 22, 356-374. Hoffman, P. The paramorphic representation of clinical judgment. Psvchologlcal Bulletin, 1960, 42, 116-131. Katz, D., and Kahn, R. The social psychology of organizations. York: Wlley, 1978. New Keeley, M. A social Justice approach to organizational evaluation. Administrative Science Quarterly, 1978, 22, 272-292. 35

a Kirchoff, B. A. Organizational effectiveness measurement and policy research. Academy of Management Review, 1977, 1, 347-355. Miles, R. H. Macro organizational behavior. Santa Monica, Calif.: Goodyear, 1980. Miles, R. H., and Cameron, K. S. Coffin nails and corporate strategies. Englewood Cliffs, N. J.: Prentice-Hall, 1982. Miles, R., and Snow, C. Organizatlonal strategy. structure, and process. New York: McGraw-Hll, 1978. Miller, D., and Friesen, P. H. Momentum and revolution In organizational adaptation. Academy of Management Journal, 21, 591-614. 1980, Molnar, J. J., and Rogers, D. C. Organizational effectiveness: On empirical comparison of the goal and system resource approach. Sociolooical Quarterly, 1976, 1I, 401-413. Nisbet, R. E., and Wllson, T. Telling more than we can know: Verbal reports on mental processes. Psychological Review, 1977, 14, 231-259. Pennings, J. M., and Goodman, P. S. Toward a workable framework. In P. S. Goodman and J. M. Pennings (Eds.), New Perspectives on Organizational Effectiveness. San Francisco: Jossey-Bass, 1977. Pfeffer, J., and Salanclk, G. R. The external control of organizations. New York: Harper and Row, 1978. 36

I Price, J. L. The study of organizational effectiveness. Saciological Quarterly, 1972, JS2, 3-15. Quinn, R. E., and Cameron, K. S. Life cycles and shifting criteria of effectiveness: Some preliminary evidence. Management Science, 1982, 22. Quinn, R. E., and Rohrbaugh, J. A spatial model of effectiveness criteria: Towards a competing values approach to organizational effectiveness. Management Science, 1983. Ramsay, V. J., and Dodge, L. D. Some empirical Issues In research on academic departments: Homogeneity, aggregation, and level of analysis. Research In Higher Education, 1983, lj, 409-420. Rawls, J. A theory of iustice. Cambridge, MA: Harvard University Press, 1971. Rohrbaugh, J. Improving the quality of group judgment: Social judgment analysis and the Delphi technique. Organizational Behavior and Human Performance, 1979, 24, 73-92. Rohrbaugh, J. OperationalIzing the competing values approach. Puh Iic Productivity Review, 1981, 5, 141-159. Scott, W. R. Effectiveness of organizational effectiveness studies. In P. S. Goodman and J. M. Pennings (Eds.), New Perspectives on Oraanizational Effectiveness. San Francisco: Jossey-Bass, 1977. Seashore, S. E., and Yuchtman, E. Factorial analysis of organizational performance. Administrative Science Quarterly, 1967, 12, 377-395. 37

Slovic, P., and Llchtensteln, S. Comparison of Bayesian and regression approaches to the study of Information processing in judgment. Organlzational Behavior and Human Performance, 1971, 6, 649-744. Slovic, P. Analyzing the expert judge: A descriptive study of a stockbrokers decision processes. Journal of Applied Psvchology, 1969, 1, 255-263. Slovic, P.; FIschhoff, B.; and LIchtenstein, S. Behavioral decision theory. In M. R. Rosenzweig and L. W. Porter (Eds.), Annual Review of Psychology, Palo Alto: Annual Reviews, Inc., 1977. Steers, R. M. Organizational effecti Monica, Calif.: Goodyear, 1977. A haw iniral vl w. v n Il e _' Santa "' ldI t,...,,l,~ J& U~AlI ~ I,, l. l I cm w ~ Van de Ven, A. H., and Ferry, D. Measuring and assessing organizations. New York: Wiley, 1980. Whetten, D. A. Coping with incompatible expectations: An integrated view of role conflict. Administratlve Science Ouarterly, 1978, 21, 254-271. Yuchtman, E., and Seashore, S. E. A system resource approach to organizational effectiveness. American SnocloglIcal Review, 1967,.2, 891-903. Zammuto, R. F. Assessing organizational effectveness! Systems change. adaptation, and strategy. Albany, NY: SUNY Albany Press, 1982. 38