F. I. DIVISION OF RESEARCH Graduate School of Business Administration The University of Michigan April 1980 JOB CHANGE AND THE MIDDLE SEASONS OF A MAN'S LIFE Working Paper No. 214 Raymond E. Hill Edwin E. Miller The University of Michigan FOR DISCUSSION PURPOSES ONLY Accepted for publication in the Academy of Management Journal, Spring of 1981.

JOB CHANGE AND THE MIDDLE SEASONS OF A MAN'S LIFE Raymond E. Hill Associate Professor of Organizational Behavior and Industrial Relations Edwin E. Miller Professor of Industrial Relations Graduate School of Business Administration The University of Michigan Ann Arbor, Mi. 48109 Abstract This study examines the job change decision from the standpoint of Daniel Levinson's recent theory of adult male development. The findings suggest that developmental tasks faced by adults in mid-life are related to various criteria in making the decision to change jobs. The establishment period of adulthood as well as the mid-life transition were particularly strong influences on the job change decision. These results are discussed from the standpoint of adult development and human resource management..s r. _

In recent years, there has been increased interest and research in the area of adult development, particularly as it relates to the world of work and occupational experience (Levinson, Darrow, Klein, Levinson & McKee, 1978; Gould, 1972; Shein, 1978; Vailant, 1977; Van Maanen, 1977). This research reveals adult development as more orderly and lawful than previously suspected and characterizes adult life as a series of developmental periods which have considerable impact on the manner in which people relate to major life decisions. The particular life decision of interest in this study is job change, an area which has also stirred increased interest in recent years. Organizations are concerned because turnover is often costly and disruptive; individuals are concerned because job change is often stressful (Brett, in press). Schein states in his recent book on careers: What is, from the point of view of the organization, simply a process of "turnover" may be, from the point of view of the individual, a major transition crisis involving a search for one's occupational niche (1978, p. 172). Note: The authors wish to thank Professor Jeanne Brett of Northwestern University for her kind and useful comments on an earlier draft of this paper, and to R. Julian Cattaneo for providing excellent computing skills. -' A.~CF —.

-2 -The present investigation examined the relation between the importance of various job change decision criteria and adult life stage. The study was focused, however, around two stages in mid-life which tend to occur primarily between the ages of 35 and 50. The expectation was that the job change criteria would vary in importance depending on the life stage of the decision maker in these two "middle seasons." Before detailing the hypotheses and findings, however, the adult development model (which is treated as the independent variable) will be briefly described. It is based on Daniel Levinson's (1978) recent discoveries as articulated in his study of forty men, The Seasons of Man's Life. His subjects were all between the ages of 35 and 45 and were taken from four different occupations. The foundation of Levinson's theory is that adults are periodically faced with new, but predictable development tasks over their entire life course and that working through these new challenges is the essence of adulthood. The adult life course is characterized by a continual series of structure building periods followed by structure changing periods in which attempts are made to work some of the "flaws" out of the prior stable phase. What then, is the nature of these developmental tasks? Levinson's basic model is briefly summarized below, starting with late adolescence. The age categories are approximate and not rigid. Also the summary below does not do justice

-3 - to either the complexity or richness of the basic framework. I. Early Adult Transitional Period (Age 18-22) Leaving the family of origin is the major task of this period. The young man feels half in and half out of the family. Much energy is expended during this period to reduce dependence on familial support and authority. Peer support becomes critical to this task, but parental control is often replaced by peer control. Marriage, the military or college often function as major transitional mechanisms here, and symbolize the first step into the adult world. II. Getting into the Adult World: Structure Building Period (Age 22-28) Getting into the adult world on a more secure basis is the prime task of this period. It is usually characterized by exploration and provisional commitments to adult roles, responsibilities and relationships. The person's energy is primarily absorbed around forming an initial occupational structure as well as forming more mature friendships and relationships. Mentors are often critical during this period in helping the young man learn the proper emotional tone of adult life as well as providing insights around more technical issues. The culmination of this period usually produces an initial life structure involving marriage and occupation. Equally important is the formation and clearer recognition of one's Dream which serves as a guiding force and provides idealized images of the man's future life structure. III. Age 30 Transitional Period (Age 28-32) In the late twenties and early thirties, one often begins to experience some "flaws" in the initial life structure and begins a transformation process aimed at re-working the initial structure to make up for "what's missing." Marriage and career shifts are common during this period. While it is a period of internal instability, the occupational and personal attachments formed during the twenties may be questioned, examined and then essentially reconfirmed with no apparent external change. A common pattern is for the young man to qualitatively shift the type of work he does even

-4 - though remaining in the same occupation, and this often involves the use of prior repressed abilities. He often seems to be "someone else" after the age thirty transition in certain respects, or more accurately "someone more." IV. Settling Down: Structure Building Period (Age 33-40) With a re-worked version of his initial life structure, the man is now ready to seek order, stability, security and control. He wants to make deeper commitments, and invest more of his self into work, family and other valued activities, There is an inclination to "sink roots" and strive for upward achievement; making The Dream a reality slowly begins to be a urgent and earnest matter. This settling down period often starts with a variety of illusions, however, the biggest of which is that realization of The Dream will result in unmitigated happiness. Thus toward the end of settling down from approximately 35 to 40, a sub-period emerges which represents a time of peaking, and is referred to as Becoming One's Own Man (BOOM). The usual feeling here is that regardless of what one has accomplished to date, he is not sufficiently his own man. Simultaneously, there is often a strong push to achieve the "career high" and finally "make it" in the occupational world. The man is pining for that crucial promotion (or at least one which seems "crucial"). This key promotion or other recognition which one feels will serve as the final liberating symbol represents an important affirmation by society. V. Age 40 Transitional Period (age 38-42) This is a major structure changing period and represents the turning point between young and middle adulthood. The transition occurs whether one "succeeds" or "fails" in the search for affirmation by society. It is only the shape of this transition that varies. The disparity between The Dream and reality usually triggers considerable soul searching since The Dream has been a guiding force since the twenties. In addition, even considerable success to date may seem hollow or bittersweet. There is an increase in one's awareness of his own mortality, and a sense of aging. A major developmental task for mid-life is the changing relationship to the self. Self acceptance, regardless of one's

-5 -plight becomes crucial to a fulfilling life in middle and late adulthood. The tyranny of The Dream is usually given up to a significant degree. Often the man becomes more existential in a sense and is more concerned with happiness in the "here and now." The imagery of "The latter," which symbolizes achievement and advancement fades in importance. VI. Beginning of Middle Adulthood: Structure Building Period (Age mid-40s) The result of this prior turning point is often a man who is more "individual" or more pronounced in his own individuality and specialness. He is less dominated by the need to win, to be right, to achieve by external criteria, and more concerned about implementing his own particular competence and enjoying his work and life. The key to this turning point lies in accepting one's self, in giving up the tyranny of external evaluations, in seeing reality clearly while being guided more by one's own internal lights. While the man is capable of greater separateness, he is also capable of greater attachment and concern for other people. Thus whereas apprenticeship roles are common prior to age forty, mentoring roles acquire increased importance to the man after forty. The developmental process during this period is often referred to as mid-life individuation. Levinson did not study people beyond the age of 45 (beginning of middle adulthood) and thus does not have sound evidence for describing development on into later adulthood; however, he is quite certain developmental opportunities continue throughout the entire life course. Of particular interest in the present sutdy was the idea of a "career peak" in the late thirties, followed by a mid-life period of lessened competitive pressure. The underlyingmodel has been suggested by previous research. Vaillant (1977, p. 202) for instance, notes that the iaturational pattern of "career consolidation" followed;\

-6 -by a period of "generativity" has been confirmed by the major American studies of adult development. Bernice Neugarten (1964) characterizes the fundamental change at mid-life as a process of turning inward toward the self, of heightened introspection, wherein enjoyment of life and work per se increases in importance. Hall, in a snythesis of the literature on work career stages with that of life stages suggests that:..the models resemble general biological growth and decay curves; early period of exploration and trial, then growth, a stable period in the middle (maintenance), and a stage of decline and withdrawal from the work environment (1976, p. 55). The hypothesis concerning the relationship between job change decisions and life stage revolve around the career peak (Levinson's settling down period) and the following maintenance stage (Levinson's mid-life transition) and will be operationalized below. Sample The data for the study came from a convenience sample of approximately 600 males who had changed jobs recently. The entire readership list of a national business magazine was scanned for persons who had sent in change of address cards within a three-month period ending in late 1976. This created a group of approximately 4,000 persons who were then sent questionnaires in early 1977 asking whether they had changed jobs and locale. Thus our assumption is that most persons had changed jobs within the past three to six months at the time

-7 - of the survey. However, it is possible that the job change occurred earlier for some respondents since job change may precede actual geographic relocation. The questionnaire asked a variety of questions concerning their attitudes and opinions regarding the job change decision. Of the 4,000 persons polled, approximately 690 responded, but only about 600 were usable from the standpoint of complete data. Thus the effective response rate is approximately fifteen percent. While this fifteen percent is admittedly low, it may be an underestimate, since many persons who sent in a change of address notice may not have necessarily changed jobs. Whether the sample is particularly representative of the population of managerial job changers is open to question at any rate; and it should also be noted that job change may be more frequent in some age groups than others (i.e., age thirty transition versus establishment). Thus we do not know whether the sample is equally representative for all age categories. Nevertheless, the fit between the data and the underlying theoretical model was of striking interest to the authors, and provided the impetus to pursue the study in spite of obvious sampling limitations. Table 1 shows a breakdown of the sample by the level of their current position, and educational background. The respondents were asked to give their ages in categories, roughly corresponding as it turned out, to Levinson's developmental periods. The Variables In addition to other questions, the participants were asked to indicate on a five-point Likert-type scale the importance or amount of influence the following factors had on their decision to change jobs (where 1 = least important and 5 = most important). These factors represent a subset of the ten most relevant dimensions from a larger set of fifteen.

Table 1 Level of Current Position and Educational Background Level Vice Presidential or higher Divisional or subsidiary management Middle Management, administrative Middle Management, engineering or manufacturing Middle Management, marketing or sales Non-managerial Education Ph.D. or LLB degree Masters degree Some graduate work Bachelor's degree Some college No college Percent 25% 14% 21% 20% 9% 11% 100% Percent 7% 29% 23% 27% 11% 3% 100% N 148 83 124 118 53 65 591 N 41 171 136 160 65 18 591

-9 - 1 Opportunity for greater responsibility 2 Special qualifications for job 3 Opportunity to become more visible 4 Experience for future assignments 5 Proven performance in this area of work 6 Particular geographic location of the job 7 Good background for top management 8 Increase promotion potential 9 Relative state and local taxes new and old locale 10 Relative other cost of living new and old locale These particular decision criteria were the result of research done in a variety of prior settings. Bassett and Myers' (1968) work on the selection decision process represents the basic contribution to the construction of the instrument. They concerned themselves with isolating various behaviors of personnel decision makers, and used the data derived from an interview study to develop the questionnaire. In an extension of this selection process research to the international setting, Miller (1972) studied the decision behavior of managers selecting Americans to go abroad. More recently, Miller and Cheng (1976) concentrated their attention on the circumstances which influence Americans to accept overseas asdiignments. The questions used in the current study are a modified version of the instrument used by Miller and Cheng in their investigation on the overseas job change. The age categories from the questionnaire were as follows: (1) below 30, (2) 30-34, (3) 35-39, (4) 40-49, (5) 50 and over. Since Levinson's age periods were not rigid with respect to their boundaries, it was felt the above categories could capture the essential dynamics of the various periods with

-10 - category (1) representing entry into the adult world, category (2) as the age-30 transition, and category (3) as the settling down and BOOM period. Category 4 combines the transition and entry into middle adulthood, and category 5 represents being firmly established in middle adulthood. Hypotheses The prior expectation was that life stage would be related to the importance attached to the various job change decision criteria. In particular, the following hypothetical relations were of interest. 1. Career peak hypothesis: The career oriented variables which reflect increasing decision making power and responsibility will peak in importance during age category (3) (35-39 yrs, the settling down and BOOM period). These variables include: opportunity for greater responsibility opportunity to become more visible experience for future assignments good background for top management increase promotion potential 2. Mid-life individuation hypothesis: The non-career oriented variables which reflect immediate job and life concerns will be at their maximum during age category (4) (40-49 yrs, the mid-life transition and maintenance). These variables include: special qualifications for job

-11 - proven performance in this area of work particular geographic location relative state and local taxes. relative cost of living It should be noted that this is not a complete test of Levinson's framework as it relates to job change, but rather it is only an examination of two periods (settling down and mid-life transition). However, the influence of these two periods can only be known by comparing all five periods, since both hypotheses involve relative maximas on the decision criteria. Furthermore, age category (4) does not fit Levinson's model as well as the others. The model focuses on ages 40-45 whereas our data involves ages 40-49. It was felt, however, that the influence of the mid-life transition could be captured by the 40-49 age bracket for two reasons. First, Levinson does not view the age categories as completely rigid in terms of their boundaries, and second, the mid-life transition is described as the single most potent developmental period in terms of its influence on adult personality. A last important qualifier on these hypotheses is that they pertain only to the job change decision and as such reflect only a narrow slice of the work, personal and family issues articulated by Levinson's model. In other words, the job change decision is merely one derivative of Levinson's framework, and the present study certainly cannot be.viewed as a comprehensive empirical test of its total validity..-' '- --

-12 - Analytical Procedures Average importance ratings on the decision criteria for each hypothesis were calculated by age category, and the mean vectors were compared using multivariate, one factor analysis of variance (MANOVA) where age categories represented levels of the factor. MANOVA was selected for comparing the means since there were significant intercorrelations among the ten variables. The intercorrelations of the dependent variables are presented in Table 2. Following the two MANOVAs (one for each hypothesis) all ten variables were grouped together and subjected to multiple discriminant analysis in an effort to interpret the underlying multivariate data structure. While this follow up procedure to MANOVA has its critics, it is established enough by other researchers to be useful in the present context (Spector, 1977; Borgen & Seling, 1978; Cooley & Lohnes, 1971). RESULTS Multivariate Analysis of Variance Tables 3 and 4 show the MANOVA analysis for each hypothesis and includes the means for decision criteria by age category, sample sizes, degrees of freedom and the approximated F statistic. This analysis indicated the observed differences in the mean vectors across age periods were statistically significant beyond

i't^~~~~~, TABLE 2 Intercorreltaions of Dependent Variables 1 2 3 4 5 6 7 8 9 10 1. Greater responsibility 2. Special qualifications.31 3. More visibility.32.16 4. Experience for future assignments.32.14.28 5. Proven performance.21.44.08.13 6. Geographic location -.16.02 -.14 -.18.01 7. Background for top mg'mt.23.10.21.30.13 -.02 8. Promotion potential.38.09.30.38.10 -.05.44 9. State and local taxes -.13.08 -.07 -.09.11.28.01.04 10. Cost of living -.08.11 -.05 -.06.07.33.02.02.83 r >.08, p <.05 r >.11, p<.01 I!

-14 - TABLE 3 MANOVA On Career Peak Variables by Developmental Age Period Age Categories 1. Opportunity for greater responsibility 2. Opportunity to become more visible 3. Experience for future assignments 4. Background for top management 5. Increase promotion potential Sample Size (N) (1) (2) (3) (4) (5) Midlife Entry Age Settling Transition Establishment into 30 Down and Entry in Adult Transi- and to Middle Middle World tion BOOM Adulthood Adulthood (30 and under) (31-34) (35-39) (40-49) (50 and over) 4.10 4.23 4.44 4.23 4.11 3.46 3.60 3.59 3.44 3.48 4.16 4.19 4.21 4.04 3.72 2.88 2.95 3.02 2.97 2.75 3.89 3.97 3.99 3.71 3.18 107 142 109 135 70 df = 20,1834 F = 2.78 p<. 0001

-15 - TABLE 4 MANOVA on Mid-Life Individuation Variables by Developmental Age Period I Age Categories 1. Special qualifications 2. Proven performance in this area of work 3. Geographic location 4. State and local taxes 5. Cost of living Sample Size (N) (1) (2) (3) (4) (5) Midlife Entry Age Settling Transition Establishment into 30 Down and Entry in Adult Transi- and to Middle Middle World tion BOOM Adulthood Adulthood (30 and under) (31-34) (35-39) (40-49) (50 and over) 3.48 3.51 3.62 3.76 4.01 3.27 3.43 3.60 3.73 3.97 3.23 3.17 3.03 3.20 3.01 2.29 2.15 2.07 2.54 2.36 2.43 2.30 2.21 2.71 2.51 112 146 116 143 74... df = 20, 1931 F = 2.93 p<.0001

-16 - the.0001 level in both cases. Pairwise comparisons afi! strata for the career peak hypothesis indicated signifi banti differences (p<.01) between each stratum and stratum ((3),. For the mid-life individuation hypothesis, significant difr — ferences were observed between (1) vs. (4), (1) vs.. ((3),7 (2) vs. (4), (2) vs. (5), (3) vs. (4), and (3) vs. ((5)., Discriminant Function Analysis Table 5 shows the overall results of the discriniiTranrt analysis. Function 1 accounted for 78.1 percent of te dSi-iscriminable variance, and was significant at the.001 IgiTeL. Functions 2,3 and 4 were not statistically significant,, Eand wzid, not be continued in the analysis. Table 6 shows the stLUtT.ie matrix, which is the correlations between the dependerft ar-iahllks and the linear compound represented by the discriminamtt TxnL-imn. The standardized discriminant weights were not shown sdnL-ce intercorrelations among the dependent Tarisables xvrmprcroi.ss the usefulness of the standardized weights for interprEtittii'no purposes. In interpreting the structure matrix, the 3aJtterm,, sign and magnitude of the correlations will be consii4-ed~T.s While the mathematical bases of discriminant and fartr canlayses are somewhat different, they present analogous iteLrp; -'tiranal issues. (In fact, the logic for constructing the tr.uTfiK e matrix in the first place derives from factor analytiac:Piraictices.) Inspection of function 1 in Table 6 suggests the miimoal differentiation among the groups on job change criteria revolves

-17 - / Table 5 Discriminant Analysis Results Signifigance of Discrimination Discriminant Eigen Percent Discriminant 2 Sign Function Value Variance Functions X df Level 1.19 78.1 1,2,3,4 126.5 40.001 2.03 12.1 2,3,4 29.3 27 NS 3.02 8.8 3,4 13.3 16 NS 4.002 1.0 4 1.1 7 NS

-18 - TABLE 6 Discriminant Structure Matrix Discriminant Function Variable 1 1. Opportunity for greater responsibility -.09 2. Opportunity to become more visible -.09 3. Experience for future assignments -.41 4. Background for top management -.08 5. Increase promotion potential -. 58 6. Special qualifications.54 7. Proven performance in this area of work.52 8. Geographic location -.05 9. State and local taxes.25 10. Cost of living.24

-19 - around positive loadings on special qualifications (.54), proven performance (.52), state and local taxes (.25) and cost of living (.24), but negative loadings on promotion potential (-.58), and experience for future assignments (-.41). In fact, all the career peak variables have negative loadings whereas all the individuation variables except geographic location have positive loadings. Function 1 could tentatively be labeled Implementation of Special Competence versus Promotion Potential. This structure is further clarified by illustrating the group centroids in discriminant space. Table 7 shows the group centroids on Function 1. The group centroids are clearly ordered in accordance with increasing age, with groups 4 and 5 (midlife transition and establishment) becoming distinctly different. This suggests that implementation of special competence and a lessening of promotion potential become important differentiators of job change after 40 years of age. Groups 1, 2 and 3 tend to be clustered together on the negative side of the axis (concern for promotion potential) whereas groups 4 and 5 move decidedly toward the positive side which represents implementation of special competence. In this regard, it should be noted that group 5 makes the most dramatic movement of all groups away from promotion potential and toward special competencies as the bases on which job change is predicated. The distance between group 4 and group 5 is slightly larger, for instance, than the distance between group 1 and group 4. The discriminant analysis thus lends strong intuitive (and statistical) support to the mid-life individuation hypothesis.

-20 - TABLE 7 Centroids for Five Groups on One Discriminant Function Discriminant Function (Implementation of Special Competence vs. Promotion Potential Group 1 (1) Under 30 yrs. Entry into Adult World -.49 (2) 31-34 yrs.-Age 30 Transition -.47 (3) 35-39 yrs.-Establishment and BOOM -. 33 (4) 40-49 yrs.-Mid-Life Transition and Entry to Middle Adulthood.15 (5) 50 and Over yrs.-Establishment in Middle Adulthood.81. ~ ~ ~ ~ ~ ~ ~ ~ -

-21 - DISCUSSION AND CONCLUSIONS Looking back to hypothesis 1, there is clear support for the career peak phenomenon in job change decisions. All five of the hypothesized variables peaked for age category 3 (establishment and BOOM), as indicated in Table 3. Hypothesis 1 is also partially supported by examination of group 3's centroid position on the discriminant function (Table 7); having a smaller value on function 1 indicates concern for promotion potential. However, groups 1 and 2 are lower than group 3, and thus the discriminant vector does not separate the groups in strict accordance with the career peak phenomenon. Generally, however, the data peak as expected, and provide, empirical validation of Levinson's (1978) adult development framework, as it pertains to occupational experience during the establishment period. In addition, the career peak data are corroborative of what is considered to be a generally agreed upon model of adult personality change. Neugarten (1964) summarizes this by describing research in which men prior to forty view the world as achievement-demanding and are content to devote themselves to those demands. After forty years of age, an increase in self-awareness and preoccupation with one's emotions emerges. In commenting on the change from the thirties to forties she notes: The basic-nature of the age change appeared to be a movement from an active, combative, outer-world orientation to the beginnings of an adaptive, conforming, and more innerworld orientation (p. 196).

-22 - The mid-life individuation hypothesis also has substantial support when both the MANOVA and discriminant results are taken together. The economic advantage variables of localized taxes and cost of living peak for group 4 in accord with the expectation. Geographic location as a decision criterion had a dual peak, once in category 4, and again in age group 1. It is of interest that "special qualifications for the job" and "proven performance in this area of work" peaked not in group 4, but in group 5 (those fifty years of age and older). This suggests that the process of mid-life individuation as it is expressed around the job change decision continues on far beyond the age 40 transition. Since Levinson did not study men beyond the age of 45, our hypotheses were not guided by empirically grounded theory for group 5. The data presented here perhaps suggest further elaborations of the theory. That groups 4 and 5 are more oriented toward implementing their special competence and less promotion oriented is strongly indicated by discriminant function 1 which clearly separates groups 4 and 5 from the others and is statistically significant. Mid-life individuation appears analogous to the principle of increasing differentiation in general systems theory (Katz & Kahn, 1978). If human personality is viewed in systemic terms, the tendency to implement increasingly personalized competencies through work would seem to be a corrolary of this generalized differentiation process.

-23 - Hall (1979a) also provides useful insights on the underlying mid-career dynamics from a socialization perspective by suggesting that: Later career development is more a process of individuation than socalization. The person becomes more of his or her own socializing agent. The person pursues more a "Protean" or self-directed career. (p. 1) In a related paper he further notes how the 20-year followup of the AT&T management progress study found an increasing tendency for people to "be themselves" (Hall, 1979b, p. 3). Mid-career change was intepreted by Hall as a balance restoring mechanism wherein the person attempts to get back to their original "dream" after being nudged away from it through the adaptational demands of early career experiences. In earlier career stages, the individual may feel more need to comply with organizational socialization demands, perhaps at the expense of their own sense of identity, Thus, mid-life individuation could be conceptualized as homeostatic processes which protect the individual's integrity (health?, see Brett, in press) particularly if human personality is again viewed from a systems theory per/eictive. The major limitations of the present research are two: first, the study has a cross-sectional design, which creates attributional problems. Maybe these particular samples just happen to use different decision criteria because they were subjected to different socio-historical influences which were related to age, but not caused by it. The middle-aged cohorts undoubtedly have more financial resources at their command, and can perhaps literally "afford to be themselves" (individuation). If the younger groupings had the same economic resources would

-24 - they behave more like the older age grouping? Mid-life is also a time when health problems rise sharply. Would the younger cohorts be more existential in their outlook given a higher incidence of health related problems? Furthermore, it is quite possible that it is not age per se, but rather work and career experience itself which create mid-life changes. Johnny Miller is a talented professional golfer who rose to the pinnacle of the highly competitive golfing world in 1974 at the age of 27. Then, shortly thereafter, at the age of 29 he went into a mysterious slump which has plagued him for some three years now. At age 30 when asked to reflet on his downward slide, he related how he had been playing golf under high pressue competitive conditions for twenty years! Most people do not fully launch a managerial career until the late twenties, let alone reach the zenith of their profession. What appears to be predictable mid-life change may be as much a function of the structure of managerial occupations. The wear and tear of competition, and the need for some respite simply may not occur with any significance until around age 40 when the typical person has had ten or fifteen years of experience. (On the other hand adult development theorists might intepret Johnny Miller's slump as a derivative of the age 30 transition.) What is needed is a panel study in which the same respondents are studied at different points in time, but even then, age as a causative factor cannot be completely untangled from the confounding effects of economic and socio-historical forces. Nevertheless, we tend to believe the age related differences in

-25 - decision criteria are part and parcel of the developmental tasks suggested by adult personality theorists, the latest of whom is Levinson (1978). The underlying theoretical framework has a variety of longitudinal precursor studies, which creates more confidence around this attributional issue. The secondresearch design limitation is that the data are post-decision impressions regarding why people made particular choices, and as such are subject to potential distortion. The question remains, however: why would people systematically distort the decision criteria in ways which conform to adult development theory? Our conclusion is that there is essential validity to the underlying model and post decision processes cannot alter the fundamental influence of the particular developmental period. The positive aspect of this particular research design is of course its efficiency and low cost relative to the sample size. It would be extremely difficult to find and follow this many job changers on a longitudinal basis; yet longitudinal designs must be a part of future research in this area. Finally, some note of the practical implications of the present data is in order. It appears that the job change decision is influenced by the developmental tasks the person is working on as an adult, and is therefore likely to have high personal significance for the individual. It would seem that simply understanding the process of adult development generally would be of value to personnel and human resource specialists if organizations are to become fuller agents of human development.

-26 - For one thing, it might make organizations more sensitive to the unique meanings job changes can hold for the individual, and thus provide insights which would result in retaining valued persons; or it could also lead to fuller counseling dialogues between human resources specialists and employees regarding how to better adapt to the occupational world generally, a function which may become more prevalent in an ever mobile society. Manfred Kets de Vries (1978) in a psychoanalytic exploration of mid-career issues, argues strongly for the importance of an expanded counseling function aimed directly at mid-career problems. Schein also articulates this expanded role for human resource management which the authors believe merits a long concluding quote: Any human resource planning and development system must attempt to match the needs of the organization with those of the individual. If such a system is to work, much more effort must be devoted to fully understanding the needs and characteristics of the individual. Those needs derive not only from the individual's working life, but also from the interaction within the total "life space" of issues of work, family and self-development. One of the weaknesses of traditional employee and management development systems has been the tendency to assume that employees can be conceived of as leaving family and self at home when they come to work and that, therefore, the organization need worry only about creating opportunities for work-oriented development activities. As the study of adult development progresses, it is becoming more and more clear that work, family and self-concerns interact strongly within people throughout their lives. This interaction simply cannot any longer be ignored (1978, p. 17). x.,

-27 -References 1. Bassett, G.A. & Meyer, H.H., Selecting Managers: A Study of the Personnel Decision Process, New York: Personnel and Industrial Relations, General Electric Co., 1968. 2. Borgen, F.H. & Seling, M.J. "Uses of Discriminant Analysis Following Manova: Multivariate Statistics for Multivariate Purposes." Journal of Applied Psychology, 1978, 63. 689-697. 3. Brett, J.M. "The Effect of Job Transfer on Employees and Their Families" in Cooper, C.L. and Payne, R. (eds.) Current Concerns in Occupational Stress, Chichester, England, Wiley, In Press. 4. Cooley, W.W. & Lohnes, P.R. Multivariate Data Analysis. New York, N.Y.: Wiley, 1971. 5. Gould, R.L. "The Phases of Adult Life: A Study of Developmental Psychology," American Journal of Psychiatry, 1972, 129, 521-531. 6. Hall, D.T. Careers in Organizations. Pacific Palisades, Ca.: Goodyear, 1976. 7. Hall, D.T. "Socialization Processes in Later Years: Adult Development in Organizations" Paper presented at the 39th annual Academy of Management meetings, Atlanta, GA, 1979 a. 8. Hall, D.T. "Midcareer Change: There's Less Than Meets the Eye" Paper presented at the 39th annual Academy of Management meetings, Atlanta, GA, 1979 b. 9. Katz, D. & Kahn, R.L. The Social Psychology of Organizations. New York, Wiley, 1978. 10. Kets de Vries, M.F.R., "The Midcareer Conundrum". Organizational Dynamics, Autumn, 1978, 45-62. 11. Levinson, D.J., Darrow, C., Klein, E., Levinson, M. and McKee, B. The Seasons of a Man's Life. New York, Knopf, 1978. 12. Miller, E.L., "The International Selection Decision: A Study of Some Determinants of Managerial Behavior in the Selection Decision Process," Journal of International Business Studies, 1972,3, 49-65. 13. Miller, E.L. & Cheng, J.L. "Circumstances that Influenced the Decision to Accept an Overseas Assignment," Proceedings of the Academy of Management, 1976, 336-339. 14. Neugarten, B.L. "A Developmental View of Adult Personality" in Birren, J.E. (ed.) Relations of Development and Aging. Springfield, Ill., Thomas, 1964, 176-208.. \~- 15. Nunnally, J.C. Psychometric Methods. New York, McGraw Hill, 1967. 16. Schein, E. Career Dynamics: Matching Individual and Organizational Needs. Reading, Pa.: Addison-Wesley, 1978.

-28 -17. Spector, P.E. "What to do with Significant Multivariate Effects in Multivariate Analysis of Variance." Journal of Applied Psychology, 1977, 62, 158-163. 18. Vaillant, G.E. Adaption to Life. Boston, Ma.: Little, Brown, 1977.