Personal experience informs us that some coworkers or friends seem dissatisfied across a variety of job circumstances, whereas other individuals appear satisfied regardless of past and present job histories. This observation is buttressed by the work of Staw and Ross (1985), who examined the consistency of cross-situational job satisfaction in a national random sample of over 5,000 men. Their data revealed a correlation of .33 between a one-item measure of job satisfaction that was taken in 1969 and 1971 for individuals who experienced both occupational and employer changes. On the basis of these and other data, Staw and Ross argued that dispositional influences importantly influence job attitudes and that perhaps too much attention had been given to specific environmental aspects of jobs as determinants of job attitudes. That is, stable individual differences might be as important a factor in determining job attitudes as that of the job or work environment itself. A recent study by Gerhart (1987) underscored the notion that both dispositional and situational factors are associated with job satisfaction. He found evidence for consistency in job satisfaction across employer and occupational changes using a youth cohort. Moreover, when jobs were coded for complexity and the impact of changes in job complexity on job satisfaction were examined, changes in job satisfaction were found to be significantly associated with changes in job complexity.

In their provocative article, Staw and Ross (1985) articulated several individual difference characteristics that might underlie the dispositional explanation of job attitudes. They suggested, for example, that:

Job attitudes may reflect a biologically based trait that predisposes individuals to see positive or negative content in their lives. … Differences in individual temperament, … ranging from clinical depression to a very positive disposition, could influence the information individuals input, recall, and interpret within various social situations, including work. (p. 471)

Staw, Bell, and Clausen (1986) conducted further investigation of the impact of dispositions on job attitudes. They found consistent evidence of affective dispositions influencing job attitudes over long periods of time. Although they could document the stability of dispositions over time, they asserted that it was impossible to determine if affective dispositions originated from genetic or social forces and how external factors might influence these affective states.

It seems reasonable that genetic factors might influence the manner in which individuals respond to their work contexts. Genetic influences on a wide range of individual differences have been well documented (Rowe, 1987). Examples include general intelligence (Bouchard & McGue, 1981; Teasdale & Owen, 1984), information processing (McGue & Bouchard, in press), personality dispositions (Bouchard, 1984; Goldsmith, 1983; Tellegen et al., 1988), psychological interests (Nichols, 1978), and attitudes (Martin et al., 1986). No study, however, has directly examined the degree to which genetic factors might influence job attitudes. We address this question in this study.

Job satisfaction is commonly viewed as multidimensional (Locke, 1976), raising the possibility that particular elements of job satisfaction vary with respect to the degree of influence of genetic factors. We suggest that job satisfaction dimensions or elements that explicitly represent extrinsic work environmental factors (e.g., the working conditions, supervision, etc.) are less likely to demonstrate genetic components than job satisfaction elements that may reflect more direct experiences of the job by individuals (i.e., the “intrinsic” aspects of job satisfaction, such as challenge or achievement). The distinction between these two generic types of job satisfaction elements has been described in earlier research (Muchinsky, 1983; Wernimont, 1966). Thus, we offer a preliminary prediction that genetic influence will be stronger for intrinsic job satisfaction indicators than for extrinsic job satisfaction variables.

This study involved the measurement of job attitudes of monozygotic twins who were reared apart (MZA) from an early age. The use of twins for determining the relative contribution of genetic and environmental factors in explaining the variance of any reliably measured variable is well known (Hay, 1985; Rowe, 1987; Willerman, 1979). Monozygotic twins reared together (MZT) share the same genetic structure and common family environment. Any similarity between such twins may, therefore, be determined by one or both of these two broad classes of influences. Dizygotic twins reared together (DZT) share one-half of their genes (on average) in common by descent. This method of estimating the contribution of genetic factors assumes equality of environmental variance for both types of twins. Given this assumption, twice the difference between the two types of twins estimates the magnitude of genetic influences. The monozygotic twin reared apart (MZA) design is much simpler. In the absence of selective placement (we assume random placement with respect to trait-relevant environments) the intraclass correlation between such pairs is a direct estimate of the genetic contribution (broad-sense heritability) to any measured variable (Bouchard, 1984, 1987; Plomin, DeFries, & McClearn, 1980). MZA twins are relatively rare. The study of such twins offers a unique opportunity to determine the magnitude of genetic influences on traits of interest.

Sample

The MZA twins in this study participated in the Minnesota Study of Twins Reared Apart between 1979 and 1987. Details of their recruitment were reported in Bouchard (1984, 1987) and the references cited therein. The zygosity diagnosis of all twins included in this study was based on analyses of eight blood group systems, four serum proteins, six red blood cell enzymes, fingerprint ridgecount, ponderal index, and cephalic index. Probability of misdiagnosis is less than 0.001 (Lykken, 1978). From 1983 to 1987, all twin pairs who participated in the study were administered a work history questionnaire containing job satisfaction items as part of the assessment battery. The same work history questionnaire was also mailed to twin pairs who had previously participated in the study but who had not completed the job satisfaction items. Thirty-four monozygotic twin pairs provided complete data. The mean age of these individuals was 41.88 years (SD = 12.03). This sample included 25 female twin pairs and 9 male twin pairs. The mean age of separation for the sample was .45 years (SD = .79), indicating that the twin pairs were separated quite early in their lives. The mean age of reunion was 31.71 years (SD = 15.77). Almost two-thirds of the participants were engaged in their major job at the time of testing, and 62% reported well-above-average success in their jobs.

Measures

We asked respondents to complete the short form of the Minnesota Job Satisfaction Questionnaire (MSQ) developed by Weiss, Dawis, England, and Lofquist (1967). They were asked to respond to questions concerning the job that they had held for the longest period of their lives or the job that they considered to be their “major job.” The category housewife was considered as a major job and respondents were asked to respond as such (Arvey & Gross, 1977). The number and types of jobs that respondents targeted as their major job were wide and diverse, ranging from research chemist to coal miner. Of the 68 MZ twins, 17 indicated that they were responding to housewife as their major work role.

The MSQ is a well-regarded measure of job satisfaction (Muchinsky, 1983) and has been used in numerous studies. The short form of the inventory includes 20 items that are relevant to a number of job facets and on which respondents indicate their degree of relative satisfaction using a 5-point scale (1 = very dissatisfied, 5 = very satisfied). The manual for the MSQ indicates that factor analytic procedures have produced an “intrinsic” satisfaction scale and an “extrinsic” satisfaction scale. The intrinsic satisfaction scale consists of those 12 items that reflect ability utilization, achievement, the chance to do things for other people in the job, and so forth. The extrinsic subscale consists of 6 items that concern the way company policies are administered, the quality of working conditions, and so forth. The general satisfaction subscale is simply a summation of the 20 items and can be viewed as a composite of all the facets of job satisfaction. The manual indicates that the internal consistency measures calculated for these scales, on the basis of a wide variety of occupational groups, produces a median reliability coefficient of .86 for the intrinsic satisfaction scale, .80 for the extrinsic scale, and .90 for the general satisfaction scale.

We formed scales for extrinsic, intrinsic, and general satisfaction for each respondent using the MSQ manual as our guide to which of the items were scored on each scale. We compared the means and standard deviations of the intrinsic, extrinsic, and general satisfaction scales in the sample with those presented in the manual to determine if the sample differed substantially from other normative groups. The MSQ manual reports the means for these three scales as 47.14, 19.98, and 74.85, respectively, on the basis of a sample of 1,723 individuals representing a variety of occupational areas. The respective values obtained in the present sample were 49.37, 21.23, and 78.64. The standard deviations were 7.42, 4.78, and 11.92 for the normative group, and 6.73, 4.56, and 10.78 for the present sample. Compared with the normative group, the sample of MZA twins expressed slightly higher satisfaction on each of the three scales; although these differences were statistically significant for all three scales, the magnitude of the differences was not substantial. The differences between the standard deviations for the sample and the normative group were quite modest, further suggesting that the twin sample was quite similar to the larger population sample.

In addition to the satisfaction items and subscales, a single item designed to measure overall job satisfaction was included in the questionnaire. This measure was also scored on a 5-point scale (1 = very dissatisfied and 5 = very satisfied).

Finally, each job was assigned several scores derived from the 4th edition of the Dictionary of Occupational Titles (DOT; U.S. Department of Labor, 1977) by Roos & Treiman (1980). These scores have been used by Gerhart (1987) in developing relatively objective non-self-report-based indices of job characteristics. Using the method described by Gerhart (1987) and Roos & Treiman (1980), jobs are assigned scores on the relative complexity, motor skills, physical demands, and undesirable working conditions associated with their respective tasks, duties, and responsibilities. This scoring system represents average levels of complexity, motor skill demands, and so forth for entire occupations rather than for specific jobs. Thus, there is considerable room for departure from these values because of interoccupational job differences that would introduce error into this variable in addition to rater error. The scoring system is factorially derived, so that scores on the complexity variable reflect DOT items such as functional complexity in relation to data, required intelligence, and required abstract reasoning, whereas scores on the physical demands variable reflect items involving such features as eye–hand coordination, climbing, hazards, and so forth.

In the present study, each job was coded independently by two raters who were blind to the twin pairing. Jobs that were based on a full sample of individuals who completed a work-history questionnaire as part of a larger study (monozygotic twins, dizygotic twins and their spouses, N = 146) were coded. The correlations between the two raters for the complexity, motor skills, physical demands, and working conditions scales across all jobs were .85, .69, .77, and .70, respectively. When the cases representing the housewives category were removed from these analyses, similar levels of interrater agreement were obtained. These values are consistent with the reliabilities reported by Gerhart (1987). Scores were subsequently obtained for each of the four scales by averaging across the two raters. We report the means and standard deviations for the satisfaction and DOT-based job ratings in Table 1.



 

Procedure

The data were analyzed in two phases. In Phase 1, all items and subscales were adjusted for the effects of age and sex (prior to computing the intraclass correlations) following the procedures outlined by McGue and Bouchard (1984). These authors have pointed out that when the variables of interest are correlated with age or sex (or with both) and are not adjusted, the twin correlations represent overestimates of genetic influence. We used the data based on the full sample set of monozygotic and dizygotic twins (N = 146) to determine the correlation of each of the MSQ items with age and sex. This information was used to correct the variables in terms of age and sex. We then computed F tests to determine if the coefficients were significantly different from zero. To determine if the intraclass correlation for the intrinsic satisfaction subscale differed from that of the extrinsic satisfaction composite, we performed a one-tailed test for the equality of two dependent correlations (Snedecor & Cochran, 1967, p. 295; Steiger, 1980).

Phase 2 analyses involved determining job similarity by calculating intraclass correlations for the age- and sex-adjusted DOT-based measures of complexity, motor skills, physical demands, and working conditions. As we discuss later, these DOT-based scores were partialed out of the satisfaction items and subscales to determine if a significant genetic component for satisfaction remained after holding the jobs constant via these DOT-based scores.

In Table 2 we report the mean squares between pairs, mean squares within pairs, and the intraclass correlation coefficient for the MSQ job satisfaction variables and composites on the basis of the sex- and age-adjusted scores. The prediction that general satisfaction would demonstrate a significant heritability was confirmed by these data. The intraclass correlation for the adjusted scores was .309 (p < .05) for general satisfaction. When this analysis was completed after removing any pair that included a housewife, the intraclass correlation was .304 (p < .08) for general satisfaction on the basis of a sample of 21 pairs. This estimate is obviously close to that observed using the full data set, but has a reduced significance value because of the smaller sample. The intraclass correlation for the single-item overall satisfaction scale, however, was not statistically significant (r = .166).



 

The age- and sex-adjusted intraclass correlation for the intrinsic satisfaction scale was also significantly different from zero (.315, p < .05), whereas the adjusted intraclass correlation for the extrinsic satisfaction scale was only .109 and not significantly different from zero. The prediction that intrinsic satisfaction would show a stronger heritability than the extrinsic satisfaction scale was not confirmed (z = 1.04, ns).

Reviewing the data on an item-by-item basis, five of the satisfaction items demonstrated significant intraclass correlations. These items involved satisfaction with the chance to be somebody in the community (Item 4), the way my boss handles people (Item 5), doing things for other people (Item 9), the freedom to use my own judgment (Item 15), and the feeling of accomplishment I get from my job (Item 20). These data confirmed the hypothesis that there is a significant genetic component to intrinsic job satisfaction as well as to general job satisfaction.

Next, we explored the issue concerning whether there were similarities in the target jobs held or described by the twins and whether partialing out features associated with these jobs would diminish the heritabilities obtained for job satisfaction. One factor that might have accounted for the results was that the twin pairs might have self-selected similar job environments (so that they experienced similar reinforcers in those environments). Such events could conceivably have led their satisfaction levels to coincide. Plomin and colleagues (1980) described situations in which individuals are not passive recipients of their environments but actively seek environments compatible with their genetic propensities. They called this process “active” genotype–environment covariance. Scarr and McCartney (1983) similarly pointed out that this form of self-seeking behavior or environmental matching may be genetically influenced.

To explore the issue of active genotype–environment covariance, we first calculated the intraclass correlations using the DOT-derived scores as dependent variables. If there were a genetically based propensity to seek out similar jobs, the heritabilities for these scores would be significant. The intraclass correlations for complexity, motor skills, and physical demands were .443, .356, and .338, respectively (all p s = .05). The intraclass correlation obtained for working conditions was not significant. These results represent evidence that the twins held jobs that were similar in terms of their complexity level, motor skill requirements, and physical demands. They are also compatible with the hypothesis that there is a genetic component in terms of the jobs that are sought and held by individuals.

To assess the question concerning whether the heritabilities obtained for the job satisfaction variables were due, in part, to the propensities of the twins to hold similar jobs, we partialed out job complexity, motor skills, and physical demand scores from the various job satisfaction measures and recalculated the intraclass correlations for satisfaction. Only slight changes in the intraclass correlations were observed. The statistically significant correlations remained significant. For example, the intraclass correlation for the general satisfaction scale (holding the complexity, motor skills, physical demands, and working conditions variables constant) was .289, a minor change from the .309 value observed earlier. Examination of the correlations between the complexity, motor skills, physical demands, and working conditions factors and the satisfaction variables revealed no significant relationships, which explains the very modest changes observed in the intraclass correlations after partialing.

The intraclass correlations of .309 for general job satisfaction and .315 for intrinsic satisfaction can be interpreted as broad heritabilities or as the proportion of variance resulting from genetic factors. The correlation computed for extrinsic satisfaction was neither significantly different from zero nor significantly different from the value obtained for intrinsic satisfaction.

It is important to understand that, although there is a significant genetic component to intrinsic and general satisfaction, this component is not overwhelming. Approximately 70% of the total variance in this variable is explained by environmental and other factors (e.g., error variance). The data are supportive of the consistency model articulated by Staw and Ross (1985). The hypothesis that the heritability for intrinsic satisfaction would be significantly higher than the heritability for extrinsic satisfaction was not supported, but the data were in the specified direction. A larger sample with greater power might substantiate this prediction in the future.

It is interesting to speculate as to why the one-item overall satisfaction measure failed to demonstrate a significant heritability. One possibility is that this item simply lacked sufficient reliability. Another possible explanation is that the variance of this item was somewhat constrained because of a leniency effect (because it has the smallest variance of any item on the inventory). A similar question involves why there is a significant heritability for certain items and not others. A possible explanation is that these differences occurred simply because of sampling error. That is, the various intraclass correlations for the items differed only because of chance fluctuations.

It is also appropriate to comment on the tendencies of the MZ twins to hold similar jobs, as indicated by their significant intraclass correlations for complexity, motor skills, and physical demands. These data are consistent with the hypothesis that these twin pairs seek out environments that are compatible with their particular genetic makeups. This process might be mediated through intellectual mechanisms. That is, relatively intelligent individuals might seek out environments that are relatively complex and challenging. This idea is reinforced by the recent comment by McCormick (1987) in which he suggests that “people tend to gravitate into occupations that are compatible with their own intellectual levels” (p. 45). Future research needs to explore this and other mechanisms in pinpointing the processes by which genetic elements influence job choices.

What are the implications of this research? First, it appears that the organization may have somewhat less “control” over job satisfaction than is commonly believed, particularly with respect to intrinsic satisfaction. Although job enrichment efforts, quality circles, and other environmental changes might be made to enhance intrinsic job satisfaction, the data suggest certain “boundaries” for each individual with regard to job satisfaction. Individuals appear to bring important predispositions to the job that may be more difficult to modify than heretofore acknowledged. These data, however, should not be construed to mean that job enrichment and other interventions will have no effect; they simply indicate that similar rank ordering on job satisfaction variables of individuals placed in different job environments are likely to obtain. Job enrichment efforts may, however, have the intended effect of raising mean levels of job satisfaction for the individuals involved, even though rank ordering of individuals is preserved.

A second implication of this research is that prediction of future job satisfaction or dissatisfaction with a different job may be possible from knowledge of current satisfaction. This implication is unique in two ways: First, we have not typically considered job satisfaction as a criterion for prediction using traditional selection methods (Pulakos & Schmitt, 1983). Second, we have not traditionally used job satisfaction as a predictor of future satisfaction, although it would seem reasonable to do so.

The ultimate goal of behavior genetic research is to specify the mode of inheritance and the mechanisms that lead to the expression of specific phenotypes. It seems plausible that several personality factors afford viable pathways, at least in terms of explaining the mechanism for the heritability of job satisfaction. For example, Tellegen (1982) has developed a number of trait measures that reflect “normal” personality factors. One factor, labeled positive affect, reflects the capacity that individuals have to experience job, enthusiasm, and feelings of trust and gratification. Another factor, negative affect, reflects a tendency to feel fearful, worried, suspicious, and dissatisfied, and to act in ways that perpetuate these feelings. Both these factors have demonstrated high heritabilities on the basis of twin studies (Tellegen et al., 1988). It may be that the genetic pathway operates via these, or similar, personality dispositions. On the other hand, the relationships between personality and job satisfaction may be limited (cf. Furnham & Zacherl, 1986), especially when the individuals studied hold the same job or occupation.

There are several factors to consider when evaluating these results. The modest sample size and degree to which generalizations to the nontwin population are allowable (on the basis of this analysis) deserve comment. For example, the sample studied includes a higher proportion of women than men than is typically found in the general population. The parents of monzygotic twins represent a random sample of the population, but this may not be true for MZA twins. Nevertheless, occupations included in this study represent quite a diverse population of jobs, so that some generalizability would appear to be warranted.

The research illustrates the utility of using twins to study organizationally based phenomena. Further research efforts might consider, for example, studying twin pairs longitudinally across different job environments to form additional estimates of genetic and environmental contributions to such variables as satisfaction and performance. We agree with Staw and Ross's (1985) contention that additional research dealing with the influence of specific traits (e.g., intelligence and personality) on job satisfaction is needed. These efforts could aid us in disentangling the various sources of variance that contribute to job satisfaction and other organizationally based phenomena.

Arvey, R. D., & Gross, R. H. (1977). Satisfaction levels and correlates of satisfaction in the homemaker job. Journal of Vocational Behavior, 10, 13–24.

Bouchard, T. J., Jr. (1984). Twins reared apart and together: What they tell us about human diversity. In S.Fox (Ed.), The chemical and biological bases of individuality (pp. 147–184). New York: Plenum.

Bouchard, T. J., Jr. (1987). Diversity, development and determinism: A report on identical twins reared apart. In M.Amelang (Ed.), Bericht über den 35. Kongress der Deutschen Gesellsschaft für Psychologie in Heidelberg 1986 [Proceedings of the 35th Congress of the German Society for Psychology]. Göttingen, Federal Republic of Germany: Hogrefe.

Bouchard, T. J., Jr., & McGue, M. (1981). Familial studies of intelligence: A review. Science, 212, 1055–1059.

Furnham, A., & Zacherl, M. (1986). How people make their own environments: A theory of genotype–environment effects. Child Development, 54, 424–435.

Gerhart, B. (1987). How important are dispositional factors as determinants of job satisfaction? Implications for job design and other personnel programs. Journal of Applied Psychology, 72, 366–373.

Goldsmith, H. H. (1983). Genetic influences on personality from infancy to adulthood. Child Development, 54, 331–355.

Hay, D. A. (1985). Essentials of behavior genetics. Melbourne, Australia: Blackwell Scientific.

Locke, E. A. (1976). The nature and causes of job satisfaction. In M. D.Durincetti (Ed.), Handbook of industrial and organizational psychology (pp. 1297–1349). Chicago: Rand McNally.

Lykken, D. T. (1978). The diagnosis of zygosity in twins. Behavior Genetics, 8, 437–473.

Martin, N. G., Eaves, L. J., Heath, A. C., Jardine, R., Feingold, L. M., & Eysenck, H. J. (1986). Transmission of social attitudes. Proceedings of the National Academy of Science, 83, 4364–4368.

McCormick, E. J. (1987). Job component validity: A reply to the letter from metanalysis. The Industrial–Organizational Psychologist, 25, 45–47.

McGue, M., & Bouchard, T. J., Jr. (1984). Adjustment of twin data for the effects of age and sex. Behavior Genetics, 14, 325–343.

McGue, M., & Bouchard, T. J., Jr. (in press). Genetic and environmental determinants of information processing and special mental abilities: A twin analysis. In R. J.Sternberg (Ed.), Advances in the psychology of human intelligence.

Muchinsky, P. M. (1983). Psychology applied to work: An introduction to industrial and organizational psychology. Homewood, IL: Dorsey Press.

Nichols, R. C. (1978). Twin studies of ability, personality and interests. Homo, 29, 158–173.

Plomin, R., DeFries, J. C., & McClearn, G. E. (1980). Behavioral genetics: A primer. San Francisco: W. H. Freeman.

Pulakos, E. D., & Schmitt, N. (1983). A longitudinal study of a valence model approach for the prediction of job satisfaction of new employees. Journal of Applied Psychology, 68, 157–165.

Roos, P. A., & Treiman, D. J. (1980). Worker functions and worker traits for the 1970 U.S. census classification. In A. R.Miller, D. J.Treiman, P. S.Cain, & P. S.Roos (Eds.), Work, jobs and occupations: A critical review of the dictionary of occupational titles (Appendix F). Washington, DC: National Academy Press.

Rowe, D. C. (1987). Resolving the person–situation debate: Invitation to an interdisciplinary dialogue. American Psychologist, 42, 218–227.

Scarr, S., & McCartney, K. (1983). How people make their own environments: A theory of genotype–environment effects. Child Development, 54, 424–435.

Snedecor, G. W., & Cochran, W. G. (1967). Statistical methods. Ames: Iowa State University Press.

Staw, B. M., Bell, N. E., & Clausen, J. A. (1986). The dispositional approach to job attitudes: A lifetime longitudinal test. American Science Quarterly, 31, 56–77.

Staw, B. M., & Ross, J. (1985). Stability in the midst of change: A dispositional approach to job attitudes. Journal of Applied Psychology, 70, 469–480.

Steiger, J. H. (1980). Tests for comparing elements of a correlation matrix. Psychological Bulletin, 87, 245–251.

Teasdale, T. W., & Owen, D. R. (1984). Heredity and familial environment in intelligence and educational level—A sibling study. Nature, 309, 620–622.

Tellegen, A. (1982). Brief manual for The Differential Personality Questionnaire. Unpublished manuscript, University of Minnesota, Minneapolis.

Tellegen, A., Lykken, D. T., Bouchard, T. J., Jr., Wilcox, K. J., Segal, N. L., & Rich, S. (1988). Personality similarity in twins reared apart and together. Journal of Personality and Social Psychology, 54, 1031–1039.

U.S. Department of Labor. (1977). Dictionary of occupational titles (4th ed.). Washington DC: Government Printing Office.

Weiss, D. J., Dawis, R. V., England, G. W., & Lofquist, L. H. (1967). Manual for the Minnesota Satisfaction Questionnaire.. Minnesota studies in vocational rehabilitation. (XXII), Minneapolis: Industrial Relations Center, University of Minnesota.

Wernimont, P. F. (1966). Intrinsic and extrinsic factors in job satisfaction. Journal of Applied Psychology, 50, 41–50.

Willerman, L. (1979). The psychology of individual and group differences. San Francisco: W. H. Freeman.

Submitted: June 18, 1987 Revised: March 28, 1988 Accepted: March 10, 1988