\lHlllll |\\|\llllllzlllllllll l W10 M9140 ——_-—- _———-—I This is to certify that the thesis entitled ‘A STUDY OF UNIVERSITY FACULTY: A REZXAKIKATION CF T72 CONCEPT OF JOB Si‘rFISFACTION presented by laanfred Stommel has been accented towards fulfillment of :3 1e requirements for .” ' . 15%;? .degree in SOClOlOEfl .!ééih‘7119¢pz3 Major professor / , Date November 10, 1978 0-7 639 ' " ‘ M” M" ' I " 3.. ,2 ‘ f“ ‘ ' ' ' 1;. _‘,l\ 6 r. - I ‘ r 1 3 s .i ‘43 s." J: ‘\\ V 5 sapfil‘o’ 4 2002 A CASE STUDY OF UNIVERSITY FACULTY: A REEXAMINATION OF THE CONCEPT OF JOB SATISFACTION BY Manfred Stommel A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Sociology 1978 Ill. illrlul‘c .P.’ Ilia ABSTRACT A CASE STUDY OF UNIVERSITY FACULTY: A REEXAMINATION OF THE CONCEPT OF JOB SATISFACTION BY Manfred Stommel In this paper it is argued that differential levels of job satisfaction can be explained successfully only when taking into account expectations and values originating in the work-role itself. The study utilizes as data 347 questionnaires from a survey of MSU faculty administered in 1977. Variables like academic rank, sex, age, or size of college have no or only a slight impact on the level of job satisfaction. However, when comparing faculty occupying similar positions but differing in influence or past pro- motion patterns, differences in the level of job satis- faction can be observed. It is hypothesized that those differences can be adequately understood only with re- ference to implicit "standards of normaltiy". Since those standards shift at the same time as job characteristics change, comparability of levels of job satisfaction between different occupational positions seems greatly impaired. TABLE OF CONTENTS Page List of Tables iv THEORETICAL ISSUES PERTAINING TO THE CONCEPT OF JOB SATISFACTION . . . . . . . . . . . . . l PLAN OF STUDY, DATA . . . . . . . . . . . . . . . 8 JOB SATISFACTION AS DEPENDENT VARIABLE . . . . . . 10 THE IMPACT OF STRUCTURAL AND DEMOGRAPHIC VARIABLES ON JOB SATISFACTION . . . . . . . . 12 JOB SATISFACTION AND INFLUENCE . . . . . . . . . . 15 JOB SATISFACTION AND PROMOTION . . . . . . . . . . 21 THE IMPACT OF COLLEGE SIZE ON JOB SATISFACTION . . 25 THE IMPACT OF INFLUENCE PATTERNS ON JOB SATISFACTION . . . . . . . . . . . . . . . . 29 SUMMARY AND CONCLUSIONS . . . . . . . . . . . . . 3O APPENDICES . . . . . . . . . . . . . . . . . . . . 34 A. Comparison of Sample Data with Population Data . . . . . . . . . . . 34 B. Faculty Questionnaire 36 C. Correlation Coefficients (Spearman's Rho) for Personal-Influence-Items of Question 6: "Compared to other faculty in your department, how much influence do you feel you have on the following departmental matters" 41 D. Correlation Coefficients (Spearman's Rho) for Faculty-Influence-Items of Question 5: "In general, how much influence do university faculty have over the following matters?Tr 42 ii Page E. Construction of Personal and Faculty Influence Indices 43 REFERENCES . . . . . . . . . . . . . . . . . . . 45 List of References . . . . . . . . . . . . 45 General References . . . . . . . . . . . . 46 iii LIST OF TABLES Table Page 1 Frequence distribution of reSponses to Question 7 (Appendix B): "How would you describe your personal satisfaction as a faculty member and the satisfaction of other faculty members you know?" . . . . 10 2 Influence indices crosstabulated with satisfaction items Kendall's Tau B . . l8 3 Comparison of faculty groups scoring differently on influence indices with respect to their level of job satis- faction . . . . . . . . . . . . . . . . 20 4 Job satisfaction crosstabulated with age within each academic rank . . . . . 22 5 Relationship between academic rank, job satisfaction and college size . . . . . 27 iv THEORETICAL ISSUES PERTAINING TO THE CONCEPT OF JOB SATISFACTION The concept of job satisfaction has been of con- siderable importance for American social scientists deal- ing with problems of work and industrial relations. Although definitions of "job satisfaction" vary somewhat from study to study, there seems to be overall agreement that it refers to "affective orientations on the part of individuals toward work roles which they are pre- sently occupying." (Vroom, 1964: p. 99) In trying to account for variations in the degree of job satisfaction, several types of explanation have been proposed. The first type of explanation might be called a "materialistic" one, in that it attempts to link job satisfaction directly to "material" conditions. The underlying assumption, here, is that job satisfaction is a direct "reflection" of structural, i.e. "objective" con- ditions. For instance, job characteristics like pay, opportunity for promotion, job security, fringe benefits, working conditions, etc. are considered to be major factors contributing to satisfaction or dissatisfaction. (Herzberg et al., 1959; Schwartz et al., 1963, p. 45-53) Other researchers have linked job satisfaction to l occupational categories (cf, review in Vroom, 1964, pp. 129-132). Generally, the results have been that higher occupational status is usually accompanied by higher job satisfaction (a rule with quite a few exceptions). A second type of explanation has been proposed to account for degrees of job satisfaction: it links the level of job satisfaction to personality characteristics. This view asserts that differences in job satisfaction are due to the fact that ultimately each individual re- acts somewhat differently to the same or, at least, similar work conditions. Since it is believed to be possible to categorize people into personality-types, job satisfaction, in this view, is seen as a consequence of differential adaptation patterns of different personality- types. (Vroom, 1964, pp. 160) Carried to an extreme, this approach sees dissatisfied individuals as incapable of adjustment to particular work conditions because of peculiar traits in their personality. A third approach to the study of job satisfaction tries to combine both of the foregoing views. Here, job satisfaction is seen both as determined by "objective" characteristics of the job as well as by individual motives and evaluations of various features connected to the job. Kalleberg, for instance, argues that job satisfaction may be best explained by perceived job char— acteristics and work values. (Kalleberg, 1971, pp. 124-163) This approach seems to be more promising, because it takes into account the differential importance attached to various job features. In addition, the importance of a particular work feature may vary for different individ- uals. It seems, for instance, to be obvious that for a person who is financially independent the amount of pay is probably of less importance than other features of the job. For this individual, then, job satisfaction may only be to a slight extent a function of the amount of pay. Consequently, Kalleberg proposes that one way to understand the meanings, individuals attribute to their work activities is "to specify the range of gratifications that are available from work in an industrial society and to assess the degree, to which particular individuals value each of these dimensions." (Kalleberg, 1971, p. 127) This could be done by gathering information regard- ing the importance workers attribute to various job char— acteristics. Subsequently, the rewards provided by various job characteristics are weighted by the importance individuals attach to them. This weighted combination of rewards is said to be a good predictor of job satisfaction. One problem of this procedure seems to be the implicit assumption that work values are to be considered as prOperties of individuals. At least, Kalleberg does not ask the question whether there are systematic and predictable variations of work values among the workers. It will be argued, here, not only that there are system- atic variations in work values, but also that these varia- tions are related to the job characteristics themselves. In trying to clarify this problem, the thesis that job satisfaction ultimately can only be explained in psychological terms will be considered. Stanley E. Seashore has made the strongest case for explaining job satisfaction in purely psychological terms and his assump- tions are at the bottom of much of the research being done in this area. It is worthwhile quoting Seashore somewhat at length, because he has formulated one of the clearest explications in the literature: A worker's experience of job satisfac- tion, and the generation of behaviors consequent on job satisfaction, is un- deniably a psychological phenomenon, therefore, In all cases an individual phenomenon. This view is valid even though there may be, and indeed are, constancies arising from the realities of the external world and from human uniformities in psychological pro- cesses. It follows that a conception of the research issues must rest fundamentally upon bio-psychological propositions. (Seashore, 1975, p. 113) Although it may be true that job satisfaction, as experienced by individual workers, is a psychological phenomenon, it by no means follows that it is therefore an individual phenomenon. The topic of most of the re- search on job satisfaction is not to explain why individ- ual X or Y is satisfied or dissatisfied. But rather the question is what are the factors contributing to differential rates or degrees of job satisfaction. Dif- ferential rates of job satisfaction, however, are social phenomena and can better be explained within a socio- logical framework. Not discounting the possibility that different personality-types could have an impact on job satisfaction (although I am not sure what "personality- type" means), with respect to rates of job satisfaction between different subgroups only two possibilities seem to exist; (1) either, different personality-types are randomly distributed throughout the social categories in question (and'in that case they could not account for differential rates), or, (2) personality-types are not randomly distributed - and would, indeed, account for, at least, some of the variation in rates - (but in that case, the sociological question would still be left, what accounts for the recruitment patterns of different per- sonality-types into different social categories.) In short, I do not think that a "conception of research issues" relating to job satisfaction can "rest fundamentally upon bio-psychological propositions." Besides questioning the individual basis of job satisfaction (in following a Durkheimian line of reason- ing), a further problem needs to be addressed. According to Seashore, "job satisfaction and dissatisfaction are of interest in the long run not so much as intrinsic ends to be maximized or minimized by society, but because they affect behavior and thus have important social implica- tions." (Seashore, 1975, p. 113) It is this latter assertion, which often serves as justification for re- search on job satisfaction. But evidence, so far, is in- conclusive as to whether or not different levels of job satisfaction have any important behavioral consequences - except, perhaps, rates of absenteeism. (Brayfield and Crockett, 1970, pp. 72-82) The aim of this research is to show that "job satisfaction" is a rather misleading concept. Job satis- faction will mainly be treated as a dependent variable in the hope that the analysis will reveal the peculiar interplay between "objective" structural variables re- lated to the job and "subjective" reactions to them. Level of job satisfaction is seen as a product of both kinds of variables and it is precisely this interplay which is believed to be resPonsible for the relatively low variation of rates of job satisfaction between other- wise completely different jobs. In other words, the reason why in most research on job satisfaction at least two thirds of all reSpondents report high job satisfaction is seen in the fact that, no matter what kind of work (barring "extremes"), people will always find reasons for being satisfied. Or to put it more sociologically: the standards, as a part of the work role, for job satisfaction vary from one job to another. This is to be the core criticism of Kalleberg's approach. By failing to tie what he calls "work values" systematically to the work roles, he proceeds as if they are randomly distributed among individuals. Job satisfac- tion, then, is a relative phenomenon, i.e. relative to the standards provided for by the job. Assuming this to be correct, meaningful comparison of differential rates of job satisfaction can only be made given the existence of a relatively fixed set of standards. In a study relating job satisfaction to occupa- tional status Zurin, Veroff, and Feld found only relatively small differences between reported satisfaction and job level. When lumping together those respondents said to ‘be very satisfied or satisfied into one group, its size ranged from 65% for unskilled workers to 83% for pro- fessionals respectively.r (Cited according to Vroom, 1964, p. 131) Rather than looking at the differences between these two (extreme!) groups, it seems to be remarkable that there is so little variation. Comparing skilled workers with managers and prOprietors almost yielded no difference at all. The percentages are 76% and 80% respectively. This not only suggests that there are dif- ferent standards at work for each occupational level, but also that a comparison of different occupations with. respect to levels of job satisfaction might be meaning- less: By measuring job satisfaction for different occupations, one does not measure the same thing. This is the final theoretical reason why job satisfaction is believed to be a variable of only limited usefulness. It probably does not have any explanatory power beyond the respective scope under which it is being studied. PLAN OF STUDY, DATA The following study investigates the relationship between job satisfaction and several other variables believed to have an impact on the variation of rates of job satisfaction. The data analyzed here were collected from a survey of faculty members at Michigan State University conducted in 1977 by the Social Science Research Bureau. Five colleges were selected that were believed to be reasonably representative of the University as a whole (see Appendix A). The analysis draws on the responses of a random sample of 347 faculty members from the five colleges. As can be seen by taking a look at the questionnaire of the 1977 survey (Appendix B), all responses, except for some demographic variables, were measured either at a nominal or ordinal level. This placed considerable restrictions on the types of statistical techniques which could be used. In order to allow for crosstabulations involving more than two variables, most four point ordinal scales had to be collapsed into dichotomous variables. Even then, introduction of more than one control variable often lead to very small cell-frequencies. Therefore, at certain points analysis could not always be followed through to the point of conclusive evidence. But still, the data allowed for enough manipulation to bring out some of the complexities. In most cases, the correlation-coefficient used is Kendall's Tau B to indicate the level of agreement between rank-orderings of two variables. If not explicitly stated otherwise, the coefficients in the following refer to Kendall's Tau B. MAJOR HYPOTHESIS SERVING AS GUIDELINE FOR THE ANALYSIS The following hypothesis is intended to limit the scope of the study. All further elaboration will be pre- sented in the subsequent analysis. HYPOTHESIS: Differences in levels of job satisfaction are due to perceived differences in levels of rewards, influence, prestige, etc., in so far as these differences contradict notions of social equity. ELABORATION: This hypothesis is meant to imply: When studying job satisfaction, standards of judgement as to what constitutes an apprOpriate level of reward must be taken into account. This, in turn, implies the notion of reference group (Merton, 1968). 10 If, for instance, a person or group of persons exercise more power than those con- sidered to be peers, job satisfaction is expected to be higher (controlling for countervailing forces). JOB SATISFACTION AS DEPENDENT VARIABLE Two questions pertaining to job satisfaction were asked in the 1977 survey. Table 1 shows questions and frequency-distribution of responses. TABLE 1 Frequency distribution of responses to Question 7 (Appendix B): "How would you describe your personal satisfaction as a faculty member and the satisfaction of other faculty members you know?" The satisfaction of other My satisfaction is: faculty seems to be: N of N of Cases % Cases % Very High 86 26% Very High 20 6% Fairly High 163 48% Fairly High 181 56% Fairly Low 71 21% Fairly Low 118 36% Very Low 17 5% Very Low 7 2% 337 100% 326 100% By taking the response to the question for personal satisfaction as a faculty member as an indicator of job satisfaction, two assumptions are made: First, it is 11 assumed that verbal reSponse to such a question actually corresponds to a hidden "psychological state" called satisfaction. Second, since job satisfaction is treated here as a summary variable, it is assumed that possible underlying dimensions referring to different aspects of the job are additive: The response to a question asking for general job satisfaction is imagined to be some kind of weighted average of all "specific" satisfactions. Whether or not this assumption is correct, cannot be examined using this particular piece of data. It is, for instance, entirely possible that for one group of faculty "satisfaction as a faculty member" pertains to the work role of being an academic professional, whereas others primarily think in terms of allegiance to the institution, i.e. MSU. For the purposes of this research, however, it is assumed that possible differences like that will not produce any systematic distortions. Looking at the reSponse rates in Table 1 two features should be mentioned: a) 74% of all faculty in the five colleges report being fairly or highly satisfied. This confirms the familiar pattern of a skewed distribution for satis- faction questions. b) When asked for the degree of satisfaction attributed to colleagues, faculty tend to shy away from making extreme judgements: 92% choose one of the two middle categories. 12 Both results are consistent with the literature (Vroom, 1964; Seashore, 1975) THE IMPACT OF‘STRUCTURAL AND DEMOGRAPHIC VARIABLES ON JOB SATISFACTION In trying to account for variation in job satis- faction, first, structural and demographic variables were introduced. More specifically, the following relation- ships were hypothesized to hold (or not to hold): 1) Female faculty should be less satisfied than male faculty. (Implicit is the assumption that female faculty do not have their own satisfaction standards as far as their work is concerned. Rather, compar- ing themselves to their male peers, they realize the relative deprivations.) 2) Faculty who obtained as their highest academic degree at the most a master's degree should be less satisfied than their colleagues with Ph.D. and/or professional degrees. 3) Those faculty working for more than 10 years at MSU should be somewhat more satisfied than their colleagues who joined MSU 1967 and later. Three reasons serve as a rationale for this hypothesis. a) Length of employment is seen as a crucial variable affecting one's ability to establish a privileged position: This effect is believed to be independent from promotion patterns. 4) 5) 6) 13 b) In addition, faculty members who were dissatisfied with their job are believed, in part, to have left MSU after some time. This would account for higher satisfaction among those employed for a longer time through self-selection. c) Those faculty members who stayed ten years or more at MSU are believed to have adjusted at least some of their expectations, which may have been at variance with their job when they joined the faculty at MSU; i.e., length of employment plays an important role in reducing cognitive dissonance. College size is believed to have a positive effect on job satisfaction. The two large colleges in the sample (which also contain the large departments) should provide more opportunities to faculty members contributing to their higher level of satisfaction. Academic rank is not believed to have strong impact on the level of job satisfaction. Although higher academic rank is usually associated with higher pay, higher prestige and status, as well as political in- fluence of various kinds, it is expected to have only modest impact on job satisfaction. The reason is that higher ranking faculty will expect those higher rewards as a normal characteristic of their job. All the hypothesized relationships between structural variables and job satisfaction are expected to be 14 rather low, since they will be mediated by variables indicating faculty's perception of their own position. The results, drawing on simple correlations be- tween job satisfaction and the mentioned variables, did not support all of the hypotheses. The only two signif- icant (at the .05 level) relationships found were those between satisfaction as a faculty member and size of college (Kendall's Tau B = +.ll) and respondents' highest academic degree (+.10). In other words, faculty in the two large colleges reported a higher level of satisfaction than their colleagues in the three small colleges, and faculty with a Ph.D. or professional degrees were more satisfied than those with only a master's degree. No relationship could be found between reported job satis- faction and sex (Phi .05 with females being slightly less satisfied than males), between job satisfaction and length of employment (Kendall's Tau B = +.02) and job satisfaction and academic rank (-.03). This indicates that faculty members who have been at Michigan State University for more than 10 years are not more satisfied than those joining the University after 1967. Associate and full professors do not show higher levels of satis- faction than assistant professors and instructors. In fact, the slightly negative correlation proved to be somewhat surprising, although no strong positive correla- tions were expected. This finding differs from Porter's 15 -study (Porter, 1962, pp. 375-384) who reports a positive relationship between level within the managerial hierarchy and degree of job satisfaction. He attributed this rela- tionship to greater opportunities for higher ranking management to satisfy needs for esteem, autonomy, and self-actualization. Contrary to Porter's finding, the data on MSU faculty suggest that there is no such positive correla- tion between occupational rank and job satisfaction, despite the opportunities associated with higher rank. However,~before any detailed analysis can be carried out, further variables must be introduced. JOB SATISFACTION AND INFLUENCE Since it is assumed that people act upon percep- tions of what they regard to be "real", rather than "reality" itself (i.e. reality as seen by the researcher), job satisfaction is expected to vary to a larger extent with variables measuring people's perception of their relative position. More specifically, two variables were selected for this study to indicate one's relative posi- tion: perceived personal influence (relative to other faculty members) and perceived influence of faculty as a group (mainly relative to administration). It is hypothesized that higher personal and faculty influence leads to higher job satisfaction. In l6 addition both influence variables should have an impact on the degree of job satisfaction faculty is willing to attribute to 95225 faculty members. In order to test this hypothesis, two indices were constructed to serve as indicators of perceived per- sonal and faculty influence. The index for perceived per- sonal influence utilizes responses to the eight items in Question 6 (see Appendix B). After cross-correlating all eight items with each other, (see Appendix C) a decision rule was adopted, namely, to exclude those items from the construction of the index which did not correlate with all other items averaging at least a strength of correla- tion of +.40 for Spearsman's Rho. A second decision rule was adOpted, namely, to exclude all items of Question 6 which did not correlate with certain key variables like academic ranks, support for collective bargaining, etc. averaging a strength of at least :.20 for Kendall's Tau B. Using both decision rules, six items were included in the index: b) Hiring of new faculty, c) Development of faculty personnel policies, d) Selection of a depart- ment chairperson, 3) Department budget allocations, f) Faculty load determinations, h) Selection of an academic dean. Two items were excluded from the index: a) Cur- riculum, and 9) Criteria for graduate student admission. The exclusion of these two items is also meaningful in theoretical terms. Both items measure personal influence 17 of faculty members with respect to student policies. Not having great influence over these matters compared to other faculty is to a certain degree, a matter of choice and, therefore, not considered to be a "deficiency". All six remaining items from "hiring of new faculty" to "selection of an academic dean" are referring to policy matters that are of great political importance for faculty members. The responses to all six items were summed up yielding a scale ranging from the value 6 to value 18 (see Appendix E). Dividing the scale into two approximately equally large groups resulted into 48% of faculty perceiving low personal influence and 52% of faculty perceiving high personal influence over those policy matters when comparing themselves to other faculty. A corresponding procedure was adopted in constructing the index of faculty in- fluence utilizing the parallel items in Question 5 (see Appendix B and Appendix D). This faculty influence index is made up of the same items as the personal influence index. We are now in a position to test the hypothesis that higher personal and faculty influence is associated with higher job satisfaction. Crosstabulating the indices for perceived personal and faculty influence with personal job satisfaction as well as satisfaction attributed to other faculty yields the following results: 18 TABLE 2 Influence Indices Crosstabulated with Satisfaction Items Kendall's Tau B Personal INFLUENCE ————— + .22* ————— Personal Satisfaction Personal INFLUENCE _____.+ .17* ————— Satisfaction Attributed to Other Faculty Faculty INFLUENCE + .25* ————— Personal Satisfaction Faculty INFLUENCE —————-+ .30* —-——— Satisfaction Attributed to Other Faculty * Significant at .05 level As can be seen in Table 2 both the personal in- fluence index and the faculty influence index are better predictors of job satisfaction than mere structural vari- ables. In addition, however, faculty influence proved not only to be a better predictor for satisfaction attributed to other faculty (as was implicitly expected), but also turned out to be a slightly better predictor for personal job satisfaction. (This result seems contrary to implicit assumptions.) Concerning the comparatively lower correlation coefficient between personal influence and satisfaction, the following interpretation is advanced. Since the personal influence index measures per- ceived personal influence relative to other faculty, a high score on this index may nonetheless be associated with a low score on the faculty influence index. Indeed, when correlating both indices, the strength of the 19 association seems to be rather weak, a mere Kendall's Tau B = +.19. This suggests that the two indices are relatively independent, i.e. they measure different dimensions. Consequently, it seems as though having higher personal influence than other faculty is not con- sidered to be of great importance as long as faculty as a whole have no influence over the policy items under- lying the construction of the index. When correlating personal influence with satis- faction and introducing faculty influence as control variable, the strength of the association between personal influence and faculty influence declines to a low Tau B = +.15 for the group who does not perceive high faculty influence. In addition, perceived personal in- fluence and perceived faculty influence are not only relatively independent, their effects on job satisfaction are also cumulative. If one compares those faculty mem- bers who score low on both personal and faculty influence indices with their counterpart who score high on both indices the following distribution results: 20 TABLE 3 Comparison of Faculty Groups Scoring Differently on In- fluence Indices with Respect to their Level of Job Satis- faction Faculty Perceiving: a) Low Personal Influence b) C) d) and Low Low and Faculty Influence Personal Influence H'gh Faculty Influence Hi h Personal Influence ang Low Faculty Influence High Personal Influence and H'gh Faculty Influence Personal Satisfaction: Low (37) 46% (16) 29% (17) 25% (15) 12% High (43) 54% (40) 71% (51) 75% (106) 88% N = 100% (80) (56) (68) (121) Table 3 clearly suggests the discriminating power the in- fluence indices have on job satisfaction when their effect is combined. Exactly the same strength of relationship emerges when comparing the same faculty subgroups with respect to satisfaction attributed to other faculty. In this case Kendall's Tau B is also +.38, i.e. faculty with high personal influence who also believe that faculty as a whole have high influence in various university policy matters tend to attribute much higher satisfaction to their colleagues than faculty who see themselves less powerful than their peers and believe in overall low faculty influence. 21 JOB SATISFACTION AND PROMOTION It has already been shown that academic rank does not correlate with job satisfaction. While usually higher rank is associated with more pay, more prestige and more influence, expectations as to what is considered to be an adequate level of rewards also increase with higher rank. The net effect is that rank and level of job satis- faction seem to be independent. The same line of reason- ing may be employed to explain why age does not have an effect on job satisfaction (Kendall's Tau B for age and satisfaction is -.02). Given that age and academic rank vary to a high degree with each other (+.63), we might hypothesize that expectations have been built up as to what is "normal" at a certain academic rank. Since the overall principle of this study is to show that differential rates of job satisfaction can be explained with reference to social standards of normality, it can be expected that while there is no overall relationship between age and job satisfaction, such a relationship will appear when look- ing at different age categories within each academic rank. To put it more clearly, it is hypothesized that deviations from the average level of satisfaction will occur among those faculty whose career did not follow the "normal" promotion patterns. Since no direct data on promotion were available, the following procedure was adopted. 22 Faculty of all four ranks (instructors, assistant, associate and full professor) were grouped into age cate- gories comprising five years each. All age categories below and above the one which contains the average age for each rank were collapsed, yielding the following tables. TABLE 4 Job Satisfaction Crosstabulated with Age within Each Academic Rank a) Instructors (Average Age: 34) Personal Satisfaction: Age: Low High N - 29 20% 80% 10 30 - 34 25% 75% 4 35 - 33% 67% 12 27% 73% 26 Kendall's Tau B: -.12 b) Assistant Professors (Average Age: 36) Personal Satisfaction: Age: Low High N - 34 19% 81% 48 35 — 39 24% 76% 25 40 - 35% 65% 20 24% 76% 93 Kendall's Tau B: -.14 23 TABLE 4 (continued) c) Associate Professors (Average Age: 42) Personal Satisfaction: Age: Low High N - 39 28% 72% 42 40 - 44 33% 67% 12 45 - 43% 57% 21 33% 67% 75 Kendall's Tau B: -.13 d) Professors (Average Age: 52) Personal Satisfaction: Age: Low High N - 49 26% 74% 58 50 - 54 22% 78% 27 55 - 21% 79% 57 23% 77% 142 Kendall's Tau B: +.06 In a way, the tables do support the assumption about the impact of "normality-standards" on differential levels of job satisfaction. If one looks at the second row of each table and compares it to the overall percent- age distribution of levels of job satisfaction within each academic rank (bottom row), one immediately discovers a striking similarity. Deviations from the mean percent- age distribution of the level of job satisfaction occur 24 only in the age groups lower and higher than the one con- taining the average age (neglecting, for a moment, the one exception of full professors). The following interpretation has been given: The five—year age groups in the middle row of each table are not simply statistical categories, but roughly correspond to a social "ideal": i.e. the "normar'instructor should be roughly between 35 and 40 years old, etc. If one assumes that those standards of what is to be considered the "normal" age for each academic rank exist, one can easily explain, why, for instance, the group of assistant professors older than 40 displays a lower level of satis- faction than their younger peers. They know that they should already be associate professor. For the three lower academic ranks the same line of reasoning can be applied to those faculty being younger than their average peer. They report a higher level of satisfaction because they are already instructor or assistant professor, etc. at an age when it is normally not expected. Professors seem to be an exception from the rule. But, being at the top of the faculty hierarchy, there is no reason for the older professors to be more dissatisfied than their younger peers, since there is no higher faculty rank left they could be promoted to. The lower rate of satisfaction of the younger professors seems to have dif- ferent reasons. As will be shown in more detail later, 25 the fact that younger professors exert so much more in- fluence than their lower ranking colleagues of the same age, is being resented by the latter. Young full pro- fessors may be powerful, but they are certainly not liked. So far, most of the analysis has been concentrated on zero-order correlations between the reported degree of job satisfaction and several other variables. In most cases, the correlations found were only of modest strength. Job satisfaction varies with perceived faculty influence (+.25), perceived personal influence (+.22), college size (+.1l), and respondents' highest academic degree (+.10), but no relationship could be found when correlating with age (-.02), length of employment (+.02), sex (-.05), and academic rank (-.03). In trying to account for more of the differences in levels of job satisfaction, a more detailed analysis is necessary. In the following, some of the more complex relationships will be elaborated. THE IMPACT OF COLLEGE SIZE ON JOB SATISFACTION Size of organizations has become a major focus of sociological analysis in recent years. Especially Peter M. Blau (1973) examined the impact of size of univer- sities and their departments on structural variables like differentiation but also in relation to power and in- stitutional innovation. 26 However, since the following analysis is based on a sample of two large colleges (more than 200 faculty members each) and three small ones (less than 90 faculty members each), all propositions about the impact of size have to be considered as tentative. Too many variables like differential funding or teaching load exist for which no control could be introduced. Nevertheless, a variety of specifications can be made. As has already been stated, faculty in the two large colleges tend to report higher levels of job satis- faction than faculty in the three small colleges. However, when controlling for sex, age, length of employment, and academic rank, the following pattern emerges. If one looks only at female faculty, faculty younger than 45 years, faculty who joined MSU after 1967, or instructors and assistant professors, no relationship can be observed between college size and level of job satisfaction. (Correlations between size and satisfaction within those subgroup rank from -.01 for lower ranking faculty to +.05 for faculty who joined MSU after 1967.) On the other hand, the strength of the correlation between size and job satisfaction increases to an average of +.20, when looking only at male faculty, faculty older than 45 years, faculty who joined MSU before 1967, and/or associate and full pro- fessors. In short, college size seems to have an impact on job satisfaction ggly among male senior faculty. Why should that be the case? 27 The original hypothesis designed to answer this question, had been that large colleges open up opportunities for higher ranking faculty which are not available in small colleges. TABLE 5 Relationship Between Academic Rank, Job Satisfaction and College Size College Size Small Large Job Satisfaction Job Satisfaction Academic Rank: Low High N=100% Low High N=100% Instructors & Assistant Professor 24% 76% 54 25% 75% 65 Associate & Full Professors 39% 61% 64 22% 78% 153 32% 68% 118 22% 78% 218 Kendall's Tau B: Kendall's Tau B: -.16 -.03 Although the data in Table 5 do not refute this hypothesis, they rather suggest a different solution. Only one subgroup differs significantly from all three others. The differences observed between small and large colleges among higher ranking faculty are entirely due to the fact that higher ranking faculty in small colleges report lower personal satisfaction than anybody else. It seems as though small colleges do not provide an 28 environment to satisfy demands of associate and full pro— fessors. Higher ranking faculty in small colleges have still another reason to be dissatisfied. The amount of personal influence accruing to them by virtue of their rank is much smaller in the three small colleges compared to the two large ones. When breaking down the relation- ship between academic rank and personal influence (an overall Tau B of +.38) by college size, one finds in small colleges a correlation of Tau B = +.27, whereas in large colleges, the strength of the relationship grows to +.42. Assuming that perceptions of personal influence reflect, by and large, actual influence patterns, it seems plausible that personal influence should be more strongly associated with academic rank in larger colleges than in smaller ones. As size of an organization increases, it becomes increas- ingly necessary to base one's interactions on social dis- tinctions rather than personal acquaintance with colleagues. Under those conditions the "automatic" link between rank and influence should be strengthened. Conversely, since in small colleges personal acquaintances and friendship should be relatively more important, the automatic link between rank and influence is weakened. Dissatisfaction arises out of this situation, since promotion often can- not be translated into social gains, although those gains were expected in the first place. 29 THE IMPACT OF INFLUENCE PATTERNS ON JOB SATISFACTION It has already been shown that perceived personal and faculty influence correlates more strongly with re- ported job satisfaction than any of the other variables taken alone. In addition, both influence variables proved to be decisive factors in explaining relationships between structural variables and job satisfaction. What has not yet been explored is the question, "What is the relative importance of each influence index in predicting satis- faction?" Somewhat surprisingly, it was found that the zero-order correlation between perceived faculty influence and job satisfaction was greater than the one between personal influence and job satisfaction. While it is hard to account for the absolute strengths of these zero-order correlations, it is possible to relate variations in the strength of the correlations between different subgroups to variations in the rate of perceived influence between the same subgroups. For example, perceived faculty influence has a much stronger impact on job satisfaction in small colleges (Tau B = +.34) compared to its impact in large colleges (+.l7). At the same time, only 46% of the faculty in small colleges (N = 118) believed that faculty as a group exerts strong influence. This compares to 65% of the faculty in large colleges (N = 211). One can explain now, the stronger impact of perceived faculty influence on job 3O satisfaction in terms of scarcity of a valued good. Faculty influence is lower in small colleges. Those who believe in its presence are all the more satisfied. Another example relates to perceived personal in- fluence. Personal influence is lower among female faculty than among male. However its correlation with job satis- faction is higher among female faculty (+.25 compared to +.20). Personal influence is a "scarce good" for female faculty. Being female and having it therefore increases one's satisfaction all the more. SUMMARY AND CONCLUSIONS It has been shown that variations in the rates of job satisfaction among different groups can be explained with reference to both structural variables as well as variables measuring one's relative position within the structure. The latter type of variables, using percep- tions of faculty and personal influence, proved to be consistently better predictors of job satisfaction than mere structural or demographic variables. This is not surprising assuming that people act upon perceptions of what they regard to be "real" rather than "reality" it- self. The findings show that only two structural vari- ables (size of college and reSpondent's academic degree) were somewhat positively related to job satisfaction. Size, academic rank and length of employment did not have any overall effects on job satisfaction. However, 31 stronger correlations could be discovered when relating job satisfaction to personal influence as well as promo- tion patterns. The latter two variables measure the relative position of the individual respondent within the structure. The impact of these variables on job satis- faction is explained in terms of cognitive dissonance ex- perienced by those faculty members who differ most from their peers. In the case of faculty's influence as a group, a similar line of reasoning is developed: The perceived influence of faculty as a whole implicitly measures faculty's position relative to other groups in the university, mainly the administrators. One problem, implicit in much of the analysis, again deserves attention. In the beginning it was said that different levels of job satisfaction are social phenomena rather than individual phenomena. When again looking at the very low rate of satis- faction among those faculty who reported low personal in- fluence despite their high academic rank, it will be re- called that low satisfaction was explained in terms of unfulfilled expectations. Those expecations are derived from social standards indicating what is to be considered an adequate level of personal influence for a particular faculty rank. If it was not for the existence of varying standards from one faculty rank to another, it would be hard to explain why perceiving low personal influence leads to lower rates of satisfaction among professors, 32 but not among instructors. The latter obviously accept their low influence as being "normal." This is precisely why Kalleberg's approach has been criticized. While recoqnizing the importance of "work values", he does not relate those values to expectations being part of the work role itself. The importance of those role-related expectations, the existence of which is hoped to have been demonstrated, contradicts Seashore's assertion that research on job satis- faction can ultimately rest on "bio-psychological proposi- tions." One major problem has not been addressed in this *paper. Does job satisfaction, taken as an independent variable, have any explanatory power of its own? Seashore, for instance, asserts that individuals in a state of dis- satisfaction would have to search for some outlet of the accumulated tension (Seashore, 1975, p. 117). But what kind of outlet? Resignation, destruction, or quest for change? Which kind of outlet is chosen depends to a large degree on structural arrangements limiting the range of possible actions. But, since structural variables in any case have to be taken into account in order to explain patterns of behavior, the question might be asked whether or not it is more useful also to seek the causes of behavioral patterns in those structural arrangements. 33 For example, in the context of this research job satisfaction proved to be negatively correlated with support for a collective bargaining unit representing faculty. But instead of relating the level of support for collective bargaining to tensions generated by dis- satisfaction, more is explained by relating it to the level of perceived faculty influence, one of the major variables contributing to differences in the degree of job satisfaction. By leaving the concept of job satisfaction completely out of the analysis no information has been lost. APPENDICES APPENDIX A Comparison of Sample Data with POpulation Data a) Comparison of faculty respondents, populations of five colleges, and university population with regard to academic rank Faculty Five College University Rank: Respondents Population Population E. E. E. E. E. .E Instructor 26 7.5% 57 8.1% 214 8.2% Assistant Professor 97 28.0% 168 24.0% 587 22.4% Associate Professor 79 22.8% 171 24.4% 632 .24.1% Professor 144 41.6% 304 43.4% 1190 45.4% Total 346* 100.0% 700 100.0% 2623 100.0% 0 x I O O -1 case 18 mlSSIng b) Comparison of faculty respondents, pulations of five colleges, and university population with regard to sex Faculty Five College University Spy: Respondents Population Population E. E. E. E. E. .E Male 267 77.6% 552 78.9% 2186 83.3% Female 77 22.4% 148 21.1% 437 16.7% (Total _ 344* 100.0% 700 100.0% 2623 100.0% 34 35 c) Comparison of faculty respondents with faculty populations in five colleges Faculpy College Colleges: Respondents Populations E E E .92 College A. 127 36.6% 265 37.8% College B 37 10.7% 73 10.4% College C 31 8.9% 53 7.6% College D 49 14.1% 95 13.6% College B 103 29.7% 214 30.6% Total 347 100.0% 700 100.0% APPENDIX B - FACULTY QUESTIONNAIRE FACULTY-ADMINISTRATOR RELATIONSHIPS CUESTIONNAIRE-l977 l. The following alphabetical list of university policies require decisions by some combination of Administrators, Faculty. and Students. Please indicate with an 'X' which group or groups should be involved in some way in that policy decisxon. Mark as many as are applicaSIe. In the next column, indicate which one group should have the most influence. Then, in the far right column, please indiEEEe the group Ehif”' actuaIIy has the most influence. For example, if for decision 3, faculty and students shou be nvolved in the decision, students Epould have the most influence, but faculty actuafly have the most influence, you would’respond: Should Have Actuallv Has Administrators Faculty Students Most Ianuence Most Influence A F S (A, F. or 5) _JA, F, or S) a. __5_ x S F Who should be involved in decision? CHECK AS MANY AS APPLY MARK ONLY ONE ' MARK ONLY ONE Should Have égtuallv Has University Administrators Faculty Students MOSt Influence Most Influence Policy Matters A F S (A, P, or S) .A, P, or S}._ ApPOinting a depu::munt ifairBerSOn - . - .1___“.I.L l I .l l III I l Appointing an acadeaic dean I 11 I l I 1_ I l I I ApPOinting a provost l ) [41 j I AL I I I Creating new educational pppgrams (41 I I I I I I I I ___. Determining faculty gglaries JJ 1 l I l I 4 1 _I 1 1 Determining if a temporary I faculty member should be rehired I I I I I I I I I I I untermining tenure for EgSUIEX members I I I I I I I I I I __ Determining undergraduate university admissions policy j j l I I I I I I I 1 Determining internal university budget allocations I I I I I I I I I I Determining university tuition and fees I I I I I I I I I I Disciplining a student for cheatingyon an exam I I I I I I I I I I Hiring new faculty members I I I I I I I I I I GO TO NEXT PAGE 36 37 2. now much priority should each of the following be given when considering faculty salaries? CHECK ONE ALTERNATIVE FOR EACH LINE A high Some Little pp priority priority priority priority at all Academic advisement of students (yL 1 l I ] 4L_I- Jab counseling and career guidance of students I I (II _J J I I rersonal values and ethical standards 1 I I I l J I L Popularity with students 1 1 l 11 I.1 I L applications I l I I I l I I Research activities IAI I.) I I, III_ Service activities in the gommunity, 1,1 I_I I I I I Service activities in the university (e.g,, committees) I j .111, 1.1 III Teaching effectiveness 1 1 4L 1 I I I.) 3. To what extent do you favor the establishment of a collective bargaining unit for H.S.U. faculty members? CHECK ONE To a great To some To a slight Do not favor extent extent extent at all __L_L__ _L_1_._ ____l_L__ ____l__L___. 4. What impact do you feel the establishment of a collective bargaining unit would have on the following changes? CHECK ONE ON EACH LINE A great Some A slight No impact impact impact impact at all ggualizing faculty salaries across units I I I I I I I I Improving the overall economic status of the faculty I I I I I I I I Reducing the merit basis of salary increases I I I I I I I I Acquiring additional funds from the - legislature I I I I I I I I Giving faculty greater involvement in decision-making I I I I I I I I Providing greater 19b security» I I I I I I I I GO TO NEXT PAGE 38 S. In general, how much influence do university faculty have over the following matters? CHECK ONE ON EACH LINE A great A moderate A slight Np deal of amount of amount of influence influence influence influence _gp_gll Curriculum ‘ I I l I I I I I"__ Hiring of new faculty I I I I I I I I Development of faculty personnel policies (reappointment! promotion, tenure) I I 1 I I I I I Selecrion of a department chairperson I I I I I I l I Department budget allocations I 1 [.1 L I I I Faculty load determinations I I I I I I I I gplveria for graduate student admission [ 1 I 1, If) I l Selection of an academic dean LA] * j I I I I I 6. Compared to other faculty in your department, how much influence do you feel you have on the following departmental matters? CHECK ONE ON EACH LINE . More than About the Less than other faculty same other faculty Curriculum I I .I I I I yiring of new faculty 11 I I I I I Development of faculty personnel policies igeappgintment, promotiopy_tenure) I I I I I I selection of a department chaigperson I I l I I I Department budget allocations If) I I I I {Acuity load determinations I I I_I I I I ~:iteria for graduate student admission [ 1 _1 1 1_1_ Selection of an academic dean 1 1 1 1 L71_ 7. How would you describe your personal satisfaction as a faculty member and the satisfaction of other faculty members you know? CHECK ONE IN EACH COLUMN My satisfaction The satisfaction of other is: faculty seems to be: Very high . . . . . . . __~__ Fairly high . . . . . . . . . _____. _____ Fairly low. . . . . . . . . _____. .____ Very low. . . . . . . . . . _____ .____. 8. How much emphasis should be placed on each of the following areas within your particular department? CHECK ONE ON EACH LINE A great Some A slight 33 emphasis emphasis emphasis emphasis at all Instruction of undergraduate majors I l I I I I I I ____ Instruction of ggaduate students 1 l I I I.) I I Advisement of graduate majors L11 I I. I I I I ,.___ Advisement of undergraduate majors I I I I_ I I I I GO TO NEXT PAGE 39 A great Some A slight yg emphasis emphasis emphasis emphasis at all Basic research I I, I I I I I I ipplied research I I _I I I I I I Instruction of undergraduate non-majors I I I I I l I I antinuing/life-long education I 1 I I III I I Implementation of a strong affirmative action program I I I I, III I I 9. To what extent do you agree with each of the following statements about current university issues? CHECK ONE ON EACH LINE Agree to a Agree to Agree to a Do not agree ggeat extent some extent slight extent at all Procedures for reappointment of faculty are generally fair to the facultyimembers involved I l I I, I l I I In making student admissions decisions, academic aptitude should be given the greatest weight I I I I Giving college credit for remedial courses is justified I 1 I 1 Eliminating academic programs or depart- ments is a legitimate means of budget reallocation I I flLHI_ lhe percentage of graduate students at HSU should be increased considerably above its present 20s level I I I 1 The current grievance procedures for faculty are adequate I 1 I ) ;uu-year community colleges would probably better serve the needs of 295‘ disadvantaged students I 1 I j The university is dealing with its current budgetary problems in the most reasonable wayipossible I I I I University rules are often ignored by facuIty I I I J Life-long education is important enough to compete with other university programs {or resources I l 1,1 There should be greater university coor- dination of programs. even if it means loss of unit autonomy wa Hoo. um ucmowuacmwm mum mucoHUHHHwoo HH< .mmm n2 can» mmoa co comma we ucwwoflmwooo oz ammo owEoomom am no :Owuoeaom .: om. GOwamflEom pompous wumoomwm Ham efiwmuwuo .m we. mm. cOwumcflEHmumo eaoa aaasoou .u om. mm. mm. ncofiumooHHm uwmesn usefiuwmomo .o. me. on. as. as. commonsense unmeuwmmmo m «o noduoeamm .o mm. on. mm. as. me. moaoaaoa Hmccomwmm huaoomw no ucoEQOHo>mo .o em. mm. mm. we. we. mm. huHSOMM 30: mo mcwuw: .n ma. mm. mm. mm. em. mm. «m. Esaoowwwoo .e ammo Gmemeom meow» com mowoaaom owEwo ucmo newumcae ImooHHm uuwauwmno Hmccomwmm Imus «0 spam mums numuwo uomosn acmEuwmo >u~sumw measomm co«u locum new omoH acme Imp m mo mo acme 3m: mo 59H: nomamm .2 ewuouauo .m muasomm .u nuummmo .m cowuomamm .o noon>oo .o madman .n newwwso .m :mmnouume Hmucmsuummwo mcaonHOM on» so m>mns pox Homu so» on codenamed 5055 30c .uomEuHmmoo moo» mo mEouHamocmoncHl Andamuem HON Roam m .cmswmmmmv muoowowwmoou sawueamuwou ca huasumw umsuo on omwmmfioo: .0 nodummso U ”Hazmmmd H1 APPENDIX E Construction of Personal and Faculty Influence Indices a) Personal Influence Index: Six items of Question 6 (see Appendix B) are used for the construction of the Personal Influence Items: b) Hiring of new faculty, 0) Development of faculty personnel policies, d) Selection of a department chairperson, e) Department budget allocations, f) Faculty load determinations, g) Selection of an academic dean. The response categories for these six items are coded in the following way: Less than About the Mbre than other faculty same other faculty Code: 1 2 3 Summing the scores for each individual respondent respectively on all six items yields a scale whose lowest possible value is 6 (respondent scores 6 x l) and whose highest possible value is 18 (respondent scores 6 x 3). The actual distribution of scores representing those individuals in the sample who responded to all six items (stus) is broken down into two categories: ues of the respondents have a score of 11 and/or lower, 52% of the respondents have a score of 12 and/or higher. Values 6 through 11 correspond to "low personal influence", values 11 through 18 correspond to "high personal influence". H3 nu. The cutting point chosen is that which comes nearest to dividing the scale into equally large groups, i.e., for which the categories "low personal influence" and "high personal influence" are chosen by approximately 50% of the respondents each. b) Faculty Influence Index: The faculty influence index uses the parallel six items in Question 5. The response categories fbr these six items are coded in the following way: No A.slight A moderate .A great influence amount of amount of deal of at all influence influence influence Code: I 2 3 u Summing the scores for eadh respondent respectively on all six items yields a scale whose lowest possible value is 6 (6 x 1) and.whose highest possible value is 2% (6 x H). 53% of the respondents have a score of 15 and/or less, ”7% of the respondents have a score of 16 and/or higher. Values 6 through 15 correspond to "low faculty influence", values 16 through 2H cor— respond to "high faculty influence". REFERENCES LIST OF REFERENCES Brayfield, Arthur H. & Walter H. Crockett. 1955. "Employee Attitudes and Employee Performance".in Victor H. Vroom and E.L. Deci (1970) Management and Motivation pp. 72-82. Harmondsworth, England: Penguin Books. Herzberg, Frederick, B. Mausner & B.B. Snyderman. 1959. The Motivation to Work. New York: Wiley & Sons. Kalleberg, Arne L. 1977. "Work Values and Job Rewards: A Theory of Job Satisfaction." American Sociological Review 42: 124-143. Porter, Lyman W. 1962. "Job Attitudes in Management: I. Perceived Deficiencies in need Fulfillment as a Function of Job Level." Journal of Applied Psychology 46: 375-384. Schwartz, M.M., E. Jenusaitis & H. Stark. 1963. "Motiva- tional Factors among Supervisors in the Utility Industry." Personnel Psychology 16: 45-53. Seashore, Stanley E. "Defining and Measuring the Quality of Working Life% in Louis E. Davis and Albert B. Cherns (eds.) The Quality of Workigg Life Volume I (1970) pp. 105—118. New York: The Free Press. Vroom, Victor H. 1964. Work and Motivation. New York: Wiley & Sons. 45 GENERAL REFERENCES Adams, J. Stacey. 1963. 'Toward an understanding of inequity'. Journal of Abnormal Social Psychology 67: 422-436. Blau, Peter M. 1973. The Organization of Academic Work. New York: Wiley & Sons. Centers, Richard. 1948. 'Motivational Aspects of Occupational Stratifications'. Journal of Social Psychology 28: 187-217. Davis, Louis E. & Albert B. Cherns (eds.). 1975. The Quality of Working_Life. New York: Free Press. Dubin, Robert. 1958. The World of Work. Englewood Cliffs, New Jersey: Prentice Hall. Festinger, Leon. 1957. A Theory of Cognitive Dissonance. Stanford, Ca1if.: Stanford University Press. Gurin, Gerald, Joseph Veroff & Sheila Feld. 1960. Americans View their Mental Health: A Nation Wide Interview Survey. New York: Basic Books. Herzberg, Frederick, R. Peterson & D. Capwell. 1957. Job Attitudes: Research and Opinion. Pittsburgh: Psychological Service of Pittsburgh. Herzberg, Frederick. 1966. Work and the Nature of Man. New York: The World Publishing Company. Homans, George C. 1974. Social Behavior: Its Elementary Forms. New York: Harcourt Brace JovanoVIch. Ladd, Everett C. & Seymour M. Lipset. 1973. The Divided Academy. New York: McGraw-Hill. Lawler, Edward E. & Lyman W. Porter. 1967. 'The Effect of Performance on Job Satisfactionfi in B.M. Bass and S.D. Deep (1972) Studies in Organizational Psychology pp. 151-162. Boston: Allyn and Bacon. 46 47 Likert, Rensis. 1961. New Patterns of Management. New York: McGraw-Hill. Madge, John. 1962. The Origins of Scientific Sociolggy. New York: Free Press. March, James G. & Herbert A. Simon. 1958. Organizations. New York: Wiley & Sons. Maslow, A.H. 1954. Motivation and Personality. New York: Harper. Merton, Robert K. & Alice S. Rossi. 1968. "Contributions to the Theory of Reference Group Behavior".in Robert K. Merton Social Theory and Social Structure pp. 279- 334. New York: Free Press. Morse, Nancy C. 1953. Satisfactions in the White-Collar Job. Ann Arbor: University of Michigan. Morse, Nancy C. & Robert S. Weiss. 1970. 'The Function and Meaning of Work and the Job'.in Victor H. Vroom and E.L. Deci: Management and Motivation. Harmonds- worth: Penguin Books. Siegel, Lawrence S. & Irvins M. Lane. 1974. Psychology in Industrial Organizations. Homewood, Illinois: RiChard D. Irwin. MICHIGAN STATE UNIV. LIBRARIES IllWWWI"WW“IIW"I?VIHWIHIHHWH 31293103449140