General job satisfaction from a developmental perspective: exploring choice--job matches at two career stages.

The authors explored predictions of general job satisfaction at early and middle adulthood and uncovered several findings about developmental processes associated with job satisfaction. Tests of life-span career theory propositions revealed that neither choice--job congruence nor gender added significantly to predictions of job satisfaction at 2 career stages. Earlier occupational choice and current job added to predictions of midcareer (modal age 43 years) job satisfaction, especially for men. The predictability of job satisfaction is apparently influenced by the career stage when satisfaction is appraised.

**********

General job satisfaction, the overall attitude of liking or disliking a job, is a universal and essential aspect of adult career development. It is a global, "bottom-line" approach to satisfaction as contrasted with a facet approach, defined as a constellation of attitudes indicating a liking or disliking of several facets of a job. Theorists and practitioners seem to accept the assumption that nearly everybody seeks satisfaction in his or her work. Satisfaction and success are considered as the two summary markers of adult vocational adjustment (Crites, 1969; Osipow & Fitzgerald, 1996).

Enhancing people's job satisfaction and diminishing their prospects for dissatisfaction are important goals for career counselors. For example, L. Gottfredson (1996) identified general job satisfaction as one of five global career status outcomes pursued in career counseling--the others were performance, persistence, economic stability, and identity. With adults, counselors typically intervene to help resolve current problems with jobs that they dislike. Counselors generally intervene to help adolescents choose and enter jobs that have promise of high satisfaction and to avoid jobs that have promise of low satisfaction. Thus, career counselors depend on understanding the dynamics of job satisfaction across the working life span--how job satisfaction develops at different stages and what conditions influence it.

Job satisfaction is one of the most widely studied constructs in the social sciences (Dorman & Zapf, 2001; Spector, 1997). Such widespread empirical investigation has led to increasingly complex interpretations and uncertainty about theoretical meanings (Hobson, 1991). Investigators, therefore, must specify their theoretical perspective. Our perspective is the life-span developmental career theory (Super, Savickas, & Super, 1996), a theory that "focuses on career development--the life course of adapting to work and working conditions" (p. 130). According to life-span theory, people's careers progress through five predictable and major life stages: growth or childhood, exploration or adolescence, establishment or young adulthood, maintenance or middle adulthood, and disengagement or old age. Each stage is named for the principal developmental task associated with the chronological age. These tasks are defined as "societal expectations about preparing for, engaging in and reflecting upon a productive work life" (Super et al., 1996, p. 131).

Life-span career theorists directly address general job satisfaction in the following formal proposition: "The degree of satisfaction people attain from work is proportional to the degree to which they have been able to implement self-concepts" (Super et al., 1996, p. 125). Satisfaction at any point in an individual's career depends on the successful implementation of occupational self-concepts, that is, "translating one's idea of oneself into occupational terms" (Super et al., 1996, p. 139). One form of occupational self-concepts is occupational choices. Developmental theorists construe expressed occupational choices as a public way to implement an individual's occupational self-concept, a manifestation of a subjective sense of purpose or goals rather than as an objective indicator of personality traits. If a person becomes engaged in work that matches his or her occupational self-concept, then he or she is likely to experience general job satisfaction. Specifically, the match between expressed occupational choices and the kind of work that a person enters contributes to the person's general job satisfaction.

Life-span career theorists view adult career development as "a continuing process of improving the match between self and situations" (Super et al., 1996, p. 139). The match between self-concepts and work situations is a never-completed project extending across the adult working career. People change their self-concepts to accommodate changing social conditions (e.g., in the workplace, community, and family). Like-wise, work conditions change in terms of labor market opportunities, organizational structures, and task content. At any particular age or stage, a person's satisfaction reflects the contemporary match between his or her general purposes and generalized job realities.

After reviewing early empirical studies on job satisfaction, Crites (1969) linked general job satisfaction with age and discovered a developmental trend in vocational adjustment. The trend was called a U-shaped "satisfaction cycle" (Crites, 1969) that started at about age 20 with high satisfaction, dipped to low satisfaction at about age 30, and then gradually rose to high satisfaction by midcareer. More recent evidence about the cycle is mixed, some supportive (e.g., Clark, Oswald, & Warr, 1996) and some revealing a linear relationship (e.g., Kalleberg & Loscocco, 1983). As people develop and adapt, their purposes more closely match the work that occupies their time and effort; in other words, the match becomes increasingly harmonious. Therefore, job satisfaction is expected to increase as people progress through career life stages.

Although women and men are generally assumed to experience similar career life stages and developmental tasks, gender role demands affecting their work life are markedly different (Levinson, 1996). After analyzing gender differences in job satisfaction, Mannheim (1993) summarized two perspectives on gender differences. The structural perspective explains observed gender differences as due to conditions, such as pay and promotion, covarying with gender rather than due to gender itself. The socialization perspective holds that the differential socialization of men and women results in psychological differences in the motivation to work, which, in turn, influences job satisfaction. As compared with men, women's job satisfaction may have different correlates, such as workplace social support (Harris, Moritzen, Robitschek, Imhoff, & Lynch, 2001) and gender concentration within an occupation (Fricke & Beehr, 1992; Smart & Ethington, 1987). Therefore, matches for women may differ from those for men.

The purpose of this study was to examine predictions of general job satisfaction during the early and middle adult stages for both men and women in an age cohort who grew up in predominantly working-class families. Specifically, we address three research questions:

1. Is the accuracy of predictions of general job satisfaction for men and women in early adulthood improved by adding a predictor, called the degree of match between an individual's high school occupational choice and current occupation?

2. Is the accuracy of the predictions of the general job satisfaction for men and women in middle adulthood improved by adding a predictor, called the degree of match between an individual's early adulthood occupational choice and current occupation?

3. Do general job satisfaction and choice--job matches change from early to middle adulthood?

For each stage (modal ages of 25 years and 43 years), we address the question of whether the match between earlier occupational choices and the present job will add to predictions of job satisfaction independent of gender, job tenure, and the type of choice and type of current job. In other words, testing the developmental proposition is limited to examining the unique contribution of the choice--job match to predicting job satisfaction after considering factors assumed to be prior conditions in the life span. We first addressed the question by considering the participants as one group and then addressed the question separately for men and women.

Predictors of general job satisfaction were tested by controlling statistically for gender, job tenure, earlier occupational choice, and current job. Controlling for gender allows for examination of Mannheim's (1993) structural explanation of gender differences in job satisfaction, namely, that any differences due to interactions with job tenure, type of job, and earlier choices will be observed. We also controlled for job tenure, that is, the length of time in the current job. An alternative interpretation of tenure is "accentuation of type over time" (Spokane, Meir, & Catalono, 2000, p. 158), that is, a progressive strengthening of qualities manifested by a personality type.

Controlling for earlier occupational choice allows for exploration of a competing proposition, called the "dispositional approach" (Staw, Bell, & Clausen, 1986), that maintains that general job satisfaction is predicted by people's relatively stable personality dispositions. Hesketh (2000) acknowledged that some personalities are more likely to be satisfied regardless of their fit with their work. For example, people who consistently express purposes and preferences that are "people-oriented" or "social" are more likely to express satisfaction with any work situation than are people whose purposes and preferences are "things-oriented" or "mechanical/realistic."

Controlling for current job allows for examination of a more parsimonious and competing proposition, called the "present status model" (Tinsley, 2000), that maintains that job satisfaction is predicted by the contemporary work environment alone. Explained another way, some kinds of work produce job satisfaction irrespective of the extent of matching with personal characteristics (Hesketh, 2000). Because people adapt to the job market, their general job satisfaction at a later career stage may be influenced by the change in match from an earlier career stage. Consequently, the first match in early adulthood was entered as a mediator of job satisfaction in middle adulthood.

Developmental theory suggests that choice--job matches change with maturity and experience and change to become more harmonious with the work situation. Therefore, the third question addresses group changes across stages in general job satisfaction and choice--job matches.

Research Review

Integrative Reviews

Person-environment (P-E) match, or congruence, has been thoroughly examined over the past three decades, as exemplified by several comprehensive literature reviews (e.g., Assouline & Meir, 1987; Edwards, 1991; Spokane, 1985; Spokane et al., 2000; Tinsley, 2000; Tranberg, Slane, & Ekeberg, 1993). Thoughtful reactions to Tinsley's synthesis have clarified the strengths and limitations of empirical evidence relating to what is generally labeled a P-E interaction (PEI) model (Tinsley, 2000). According to Tinsley, job satisfaction is the most frequently studied outcome associated with P-E congruence. Most P-E studies of job satisfaction focus on interest-job congruence and often use interest inventory scores such as the P (Person) index. Only a few investigators tested Super's (Super et al., 1996) idea of self-concept or occupational choice as the P index. Several early studies surprisingly used expressed choices as measures of environment, which elicited caution from Spokane (1985) on the grounds that "more rigorous environmental measures" (p. 332) are needed. We would add that measures derived from theory are also essential.

Integrative reviews of P-E congruence studies have yielded mixed conclusions. Tinsley (2000) found that "with very few exceptions, these studies report that congruence is positively related to indices of employee well-being (e.g., job satisfaction, job involvement, and trust) and negatively related to indices of employee discontent (e.g., absenteeism and turnover"; p. 150). He suggested that the evidence may even underestimate the efficacy of the P-E fit model because of research design flaws. On the other hand, Tranberg et al. (1993) performed a meta-analysis of 22 studies of interest-job congruence with satisfaction and concluded that the overall mean congruence--satisfaction correlation was not statistically significant and that "interest congruence alone does not predict satisfaction" (p. 261). Spokane et al. (2000) observed that recent studies have concluded that congruence seems to be a sufficient, though not a necessary, condition for job satisfaction, with median correlations in the .25 range. They listed job satisfaction among several correlates of congruence observed over many studies, although they noted that "the evidence is mixed on this variable" (Spokane et al., 2000, p. 177). They also concluded that gender has generally not been found to be associated with congruence and that some environments were generally more satisfying than were others.

Longitudinal Studies Using Choices

Among the recent studies reviewed by Spokane et al. (2000) were three longitudinal studies of college graduates' job satisfaction that used expressed choices as the P (Person) measure and current job as the E (Environment) measure. Smart, Elton, and McLaughlin (1986) analyzed longitudinal data from more than 10,000 individuals who entered college in 1971 and who were employed full-time in 1980. They found that congruence between college major and the job held 9 years later significantly predicted intrinsic job satisfaction. The effects of congruence on extrinsic and overall job satisfaction were different for men and women. Elton and Smart (1988) explored the relationship between congruence and extrinsic job satisfaction for 1,869 college graduates (from the same data set as used by Smart et al., 1986) who entered 13 occupations representing all Holland environments except Realistic. Their findings suggested that both gender and the occupation entered influenced extrinsic job satisfaction. Overall, people with the highest congruence level were less dissatisfied than were people with the lowest congruence level. Smart and Ethington (1987) studied job satisfaction for 2,165 women (again from the data set used by Smart et al., 1986) who were employed in three kinds of jobs: male dominated, sex balanced, and female dominated. Women employed in sex-balanced jobs expressed more satisfaction with the intrinsic and overall nature of jobs than did women in female-dominated occupations. These three studies suggest that congruence between expressed choice and the job held later is predictive of college graduates' job satisfaction and that those predictions vary somewhat by gender.

Using an archival data set that contained observations from the 1930s and the 1950s, Leong and Boyle (1997) studied predictors of midlife career adjustment, including general job satisfaction. They found that men's type of job was more stable (48% the same) than was women's (28%) and that, over time, both men's and women's jobs became more congruent with their interests. Midlife job satisfaction was predicted at a significant level by values and personality factors but not by interest-job congruence.

Job Satisfaction and Age

The evidence about changes in job satisfaction over a life span is also mixed. A few representative studies and summaries illustrate this conclusion. Large-sample, cross-sectional surveys seem to have found a functional relationship between age and job satisfaction. For example, Clark et al. (1996) found a U-shaped relationship in interviews with 5,192 British employees ages 16 to more than 60 years. Although the design was cross-sectional, the authors speculated that changes in societal expectations with increasing age are important to age differences. Kalleberg and Loscocco (1983) used data from a 1972-1973 U.S. national survey of 1,391 individuals ages 16 to more than 61 years and found that age was positively related to job satisfaction. Although conceding that differences between the samples may account for the differences found in the relationships, the authors argued that the more general processes of aging and development were at work. Weaver (1980) examined job satisfaction among 4,709 American workers from 1972-1978. Although job satisfaction remained the same over the years, it was associated with both age and occupation.

Method

Participants

Participants, a total of 249, were 1973 graduates from three contiguous Wisconsin high schools who voluntarily completed questionnaires when in 12th grade (1973), again 7 years after graduation (1980), and again 25 years after graduation (1998). Complete data for 12th-grade and both follow-up surveys were available for 169 individuals, or 75% of the individuals who supplied 12th-grade data. Individuals who did not complete one of the questionnaires were eliminated only from analyses requiring those data.

Approximately 85% of all the graduates contacted in the spring of 1980 responded to the follow-up questionnaire survey in the spring of that year. Thanks to 25th anniversary class reunion committees, we were able to locate 1998 addresses for 228 graduates (6 were deceased and accurate addresses were not available for 15). Graduates who did not respond to two letters were called and invited to complete the survey by telephone. About 75% of those contacted provided responses to the survey. Attrition was greater for men than for women and for graduates who had left their home communities than for those who had stayed. We compared the responses that we had received with the responses for one class's 20th reunion and found that the people who did not respond to us also did not respond to their classmates.

The sample is remarkable among follow-up studies for retention of farming and working-class members. Approximately 31% grew up in farm families, 61% in working-class families, and 8% in managerial-professional families.

Few graduates sought employment far away from their hometowns. In 1980, the vast majority lived and worked within 40 miles of their high school; in 1998, about three fourths of all respondents lived within 40 miles of their original high schools. The respondents who no longer lived within 40 miles of their high school were working in at least 20 other states and two foreign countries.

Setting

The following material about the setting and history is substantially the same as appears in Jepsen and Choudhuri (2001), because both that study and this study used the same data archive. Participants grew to adulthood in a rural setting with particular geographic, cultural, and economic features. The three contiguous communities are located in a rolling stretch of rural Wisconsin that was cleared of thick hardwood forest and settled in the early nineteenth century by European, primarily German, immigrant farmers. Evidence of the Germanic cultural influence can be seen today in the form of locally popular foods and festivals. According to 1970 and 1980 census data, the area's population is overwhelmingly White and of European descent. The local economy of this area is dominated by two factors: It is the state's most industrialized area, measured by proportion of factory jobs, and it is one of Wisconsin's four major "cheese regions," although the dairy farms are relatively small on average, owing to early problems with clearing the land. All three communities are within a 1-hour commute of several large cities that have diversified businesses and industries and public technical colleges and universities.

Although the communities share boundaries, each community has unique features. School A was located in a thriving manufacturing and dairy city with a 1970 population of 2,600. Metal and foundry industries produced farm machinery, lawn and garden equipment, and parts for other machines. The success of two large companies has supported a secure and well-paid workforce. School B was located in a village of 900 (according to the 1970 census) with an economy built around dairy farming and supporting businesses. School C was also located in a dairy farming and manufacturing region. The district includes four villages that range in the number of inhabitants from 250 to 1,000 (according to the 1970 census). Perhaps more people from this area than from the other two communities commuted to industrial plants in nearby cities.

Period in History

Participants entered high school in the fall of 1969 and experienced a unique period in social and economic history that had an impact on their careers, as two examples make clear. The boys left high school shortly after the military draft was discontinued, thus only a few served in the military and none experienced combat. The girls left high school to enter a labor market that was open to women as never before. In 1970, the state workforce included 37.2% women, but by 1980 the proportion had grown to 43%. In 1970, the professional specialty occupations included 45.3% women, and by 1980 more than half (50.3%) were women. Had this study been initiated 10 years earlier, we would be reporting very different kinds of occupational careers; had it been conducted 10 years later, the effects of the tight labor market and farm crisis in the 1980s would surely have been evident.

Instruments

Data used in this study were responses to selected questions on three instruments: (a) the Occupational Information Inventory (OII), a 13-item questionnaire administered under classroom conditions in March 1973; (b) the Follow-Up Survey 80 (FUS 80) a four-page survey mailed in spring 1980; and (c) the Follow-Up Survey 98 (FUS 98), an eight-page survey mailed in summer 1998. Participants' completion of each survey followed their being informed about the uses of the data and giving their consent for their responses to be used in the study. Additional descriptions of the instruments can be found in Jepsen (1975) and Jepsen and Choudhuri (2001).

We used these instruments to assess five variables of interest in this study. Occupational choices were elicited on both the OII and the FUS 80. Participants supplied answers to questions about the occupation that they expected to enter in the future. Responses were classified into Holland three-letter codes by trained raters using the Dictionary of Holland Occupational Codes (DHOC; G. Gottfredson & Holland, 1996). Raters were instructed to match a participant's expressed choice to one occupation and record the matching three-letter Holland code. Two trained raters independently classified all 12th-grade occupational choices. Interrater agreement is the preferred method for assessing consistency for rating nominally scaled data such as jobs based on Holland types (Tinsley & Weiss, 1975). The interrater agreement on first-letter codes for all available choices (N = 216) was 89.8%. We reconciled differences for final coding. All 1980 choices were classified by the second rater and verified by us. Because many participants listed their 1980 current job as their 1980 occupational choice, the reliability of 1980 occupational choice was partially dependent on the classifications of current jobs.

Current job, job tenure, and job satisfaction were measured by responses to question on the FUS 80 and the FUS 98. Current jobs were those listed as full-time and included military occupations and homemaker. Each current job was classified according to Holland's typology using the DHOC, with the aid of the Dictionary of Occupational Titles (DOT; U.S. Department of Labor, 1991). Both the 1980 and 1998 survey questions asked participants to describe three facets of their current jobs: (a) job name, (b) job tasks and duties, and (c) employer. Each facet was used as a check on the other two in assigning an occupational title to the participant's response.

Interrater agreement on classification of current job was established for four raters. Job ratings by a first rater were verified for all cases by a second rater, and the first author served as a tiebreaker. Agreement between the first two raters was assessed for a 16% random sample of 27 cases. The percentage of agreement on assigning jobs to one-letter Holland codes was 92.6%. A fourth rater rated the same 27 cases, using the same procedures, and agreed with the consensus rating on 92.6% of the occupations.

Job tenure is the number of years on the present job as reported on the FUS 80 and the FUS 98. There are no reliability estimates for these reports.

Job satisfaction was assessed by the Job Satisfaction Blank (JSB; Hoppock, 1935), which was included in both follow-up surveys. This scale measures global job satisfaction by summing ratings across four items and is one of the earliest, best known, and most widely used measures of general job satisfaction. McNichols, Stahl, and Manley (1978) evaluated JSB construct validity, concurrent validity, convergent validity, and reliability for more 29,000 workers and concluded that the measure performs well, and its use is suggested when lengthier instruments are not warranted. Hoppock reported an initial split-half reliability of .93 (N = 310), and several subsequent studies have found similar reliabilities. Evidence of validity is that other job satisfaction measures are correlated highly with the JSB (Breeden, 1993; Lofquist & Dawis, 1984).

The match or congruence between occupational choice and current job was assessed by the Congruence, or C, index introduced by Brown and Gore (1994). The C index offers three advantages: (a) All three letters of the Holland code are accounted for and the letters weighted differentially; (b) C allows fine distinctions among persons with similar, but out of order, three-letter codes; and (c) C is relatively easy to calculate.

Congruence was measured for each participant at two stages: (a) modal age 25 years in 1980, the match between participants' 12th-grade occupational choice and their current job at age 25, and (b) modal age 43 years in 1998, the match between occupational choice at age 25 and current job at age 43. All matches were made using Holland's (1997) hexagon model. Participants' occupational choices in 12th grade and at age 25 indicated their subjective self-concept at that time. In terms of life-span theory, each C index indicated the degree to which a participant's self-concept was implemented several years later in the work environment of his or her current job.

Analysis

Preliminary data analyses included the derivation of descriptive statistics and bivariate Pearson product--moment correlations among the continuous variables of job tenure, job satisfaction, and congruence at each of two occasions (1980--modal age 25 years; 1998--modal age 43 years). All 12th-grade (n = 155) and age 25 (n = 169) occupational choices and age 25 (n = 168) and age 43 current jobs (n = 167) were classified by Holland types, and their categorical distributions were examined. A dependent t test was used to analyze differences across time in job satisfaction and congruence by comparing age 25 and age 43 scores.

Six hierarchical regression models--two each for total sample, women, and men--tested job satisfaction predictions at both age 25 years and age 43 years. For the total sample regression model, five independent predictor variables were entered separately and sequentially into the model: gender, occupational choice, job tenure, current job, and congruence, as well as interactions among occupational choice, current job, and congruence. The choice and current job predictors were dummy coded, as recommended for categorical predictor variables (Kleinbaum, Kupper, & Muller, 1988). Each predictor was coded as five dummy variables used to represent each participant's membership in the six Holland types.

We entered predictors in a sequence parallel to the temporal order that participants experienced them. The contribution of gender to predictions of general job satisfaction was tested as a mediating variable in the regression model and by examining separate regression models for each gender group. All predictors except gender were entered in the same order into the separate regression models for women and men. Because this is an exploratory study, the alpha level was set at .05, although significant relationships may be identified that are not practically meaningful at that level. Therefore, effect scores, Cohen's (1988) [f.sup.2], are reported to provide a basis for evaluating the practical implications.

Results

The match between occupational choice and current job depends, in part, on the independent distributions of occupational choice and current job across the Holland types. Both occupational choices and current jobs are distributed unevenly across the six Holland types and between genders, which is not surprising in light of similar distributions in the 1970 national census (L. Gottfredson, 1996). Among 12th graders, 6% of women and 64% of men chose Realistic occupations, and 44% of women and 12% of men chose Social occupations. Seven years later (1980), the proportion of women and men who were working in Realistic-type current jobs was 26% and 71%, respectively, whereas 35% of women and 11% of men occupied Social-type current jobs. In 1980, the distribution of both men's and women's occupational choices (what they expected to do in 5 years) across Holland types was almost identical to the distribution of their current jobs, suggesting that occupational choices at age 25 were an extension of their current job. Indeed 40% of the participants changed the Holland type of their job between 1974 (1 year out of high school) and 1980.

The distribution of current jobs at age 43 revealed several changes over 18 years. The trend for men was to change from Realistic occupations to Enterprising occupations and for women to leave Social occupations and enter Enterprising occupations, with the result that approximately the same proportion of men and women at age 43 (25% and 24%, respectively) were holding Enterprising current jobs. At age 43, men were still predominantly engaged in Realistic occupations (59%), followed by Enterprising occupations (25%), whereas women were distributed among Social occupations (27%), Realistic occupations (24%), Enterprising occupations (24%), and Conventional occupations (21%). Investigative and Artistic occupations were rare (less than 5%) at both ages 25 years and 43 years.

The bivariate correlation coefficients among gender, job tenure, job satisfaction, and choice--job congruence at age 25 (in 1980) and age 43 (in 1998) were generally not statistically significant (see Table 1). The exception was a .26 correlation (p < .01) between job tenure at age 43 and congruence at age 43. Gender, when coded as a binary variable, was not correlated with either congruence or job satisfaction at either age 25 or 43. In summary, most of the individual measures were not significantly related. Thus, the threat of multicolinearity in the regression model was reduced.

We tested the multivariate prediction of job satisfaction at modal age 25 years (in 1980) using hierarchical regression analyses. Five independent variables were entered into the model in the order in which they were assumed to be experienced: (a) gender, (b) 12th-grade occupational choice, (c) age 25 job tenure, (d) age 25 current job type, and (e) age 25 choice-job match or congruence. See Table 2 for a summary of the findings. When age 25 job tenure was added to the previous predictors (i.e., gender and 12th-grade occupational choice), the ability to predict for the total sample increased significantly (p = .044). In addition, the ability to predict men's age 25 job satisfaction increased significantly (p = .041) by entering the age 25 job type as a predictor after gender, 12th-grade occupational choice, and age 25 job tenure were already in the model. These findings must be considered with caution because two significant differences from among so many significance tests are probably within the 5% error rate. Job satisfaction score effect sizes for the whole regression equations are total sample, [f.sup.2] = .07; women, [f.sup.2] = .08; and men, [f.sup.2] = .20. According to Cohen (1988), the first two effect sizes should be considered small, and the third one should be considered medium. Job satisfaction for men in early adulthood seems to be affected by all the predicators, but the effect is not high for women.

Multivariate predictions of midlife adult (age 43 years) job satisfaction were more remarkable. A hierarchical model was tested with six independent variables entered in the following order: (a) gender, (b) age 25 congruence, (c) age 25 occupational choice, (d) age 43 job tenure, (e) age 43 current job type, and (f) age 43 congruence. The results are provided in Table 3. For all participants, the first-letter Holland code of age 25 occupational choice and age 43 current job added significantly to predictions of age 43 job satisfaction when all prior entries were controlled. In other words, age 25 occupational choice made significant additional contributions to the prediction when gender and age 25 congruence were already in the model. At age 43, current job made significant additional contributions to predicting job satisfaction when gender, age 25 congruence, age 25 occupational choice, and age 43 job type were already in the model. Other predictors did not add significantly to the prediction of age 43 job satisfaction.

The predictions for the adults in midlife for the total sample applied to male but not to female participants. For men, age 43 job satisfaction was predicted by age 25 occupational choice when age 25 congruence was taken into account. Age 43 current job was a significant predictor of job satisfaction when age 25 congruence, age 25 occupational choice, and age 43 job tenure were included in the prediction model. The age 43 job satisfaction score effect sizes for the whole regression equations were total sample, [f.sup.2] = .14; women, [f.sup.2] = .10; and men, [f.sup.2] = .39. As in 1980, predictions of men's satisfaction seemed to be more powerful than predictions of women's, especially when the length of time (18 years) is considered.

The significant contribution of age 25 occupational choice and current job to predicting age 43 job satisfaction both for the entire sample and for men only was further explored by conducting several post hoc one-way analyses of variance to detect differences among Holland type of age 25 occupational choice and age 43 current job. See Table 4 for results. Because only a few participants were engaged in current Investigative and Artistic jobs, they were not included in the analyses. Based on Scheffe post hoc analyses, participants with Realistic-type age 25 occupational choices and Realistic-type age 43 current jobs had the lowest age 43 job satisfaction, and participants with Social-type age 43 current jobs tended to have the highest age 43 job satisfaction.

Participants' general job satisfaction and choice-job match (congruence) in early adulthood (age 25) were compared with satisfaction and match in midlife (age 43). The group mean job satisfaction scores were essentially the same at age 25 (M = 20.4, SD = 3.8) and at age 43 (M = 20.7, SD = 3.1), yielding a dependent t of 0.54, p = .59. On the other hand, the group mean congruence scores were higher at age 43 (M = 13.3, SD = 4.3) than at age 25 (M = 11.2, SD = 4.2), which is a group difference that exceeded chance (dependent t = 4.25, p < .001). Apparently the choice--job match increased over the adult years, as developmental theory suggests, whereas participants' degree of general job satisfaction remained the same. Of course, these are group comparisons across time and do not apply to an individual's changes.

The correlation between age 25 and age 43 job satisfaction scores was .19 (p < .02) for the total group (n = 147), .13 (p = .21) for women (n = 98), and .33 (p < .02) for men (n = 49). Considering that less than 4% of the total group variance in age 43 job satisfaction was explained by age 25 job satisfaction, the job satisfaction stability for the total sample was not high. Stability among men was somewhat higher, because nearly 11% of the variance in age 43 job satisfaction was explained by age 25 job satisfaction. A stability index of this magnitude for an attitudinal variable over 18 years seems noteworthy.

Discussion

Design Strengths and Limitations

The results should be interpreted in the context of several design strengths and limitations. First, the sample was heterogeneous with regard to job satisfaction but not occupational choices or current jobs, thus the "attenuation of range on the important variables" (Tinsley, 2000, p. 151) is a limited threat. The population came from a region with a relatively narrow range of cultural and economic differences. Although this has the advantage of ruling out alternative explanations of the findings, it reduces generalizability; that is, the findings are most applicable to people who grow up in majority White cultures and working-class or farming families.

The relatively high level of participation suggests that the sample represents the original population of high school graduates as well as has been the case in most follow-up studies. Although not large, the size of the sample in our study is at the median among those satisfaction studies reviewed by Tranberg et al. (1993) and includes more working-class participants than do most studies reviewed. On the other hand, the loss of about one fourth of the participants in the post-high school surveys may have influenced job satisfaction score distributions, because survivors of follow-up studies tend to report higher satisfaction (Rounds, McKenna, & Hubert, 2000). Low numbers of participants entered Artistic and Investigative jobs, thus limiting inferences from the results to the other four types.

The design takes into account the direct prediction of satisfaction from both occupational choice (the person variable) and current job (the environment variable), thus exploring competing parsimonious explanations. Established measures of job satisfaction and congruence were used, although only one measure of each was included.

The congruence results may be attenuated by the fact that congruence is derived from measures of both occupational choice and current job and thus "inherits" the error in both these measures. Interrater disagreement for classifying choices and jobs independently ranged between 7% and 10%, therefore the disagreement rate for congruence was very likely to be even greater. Compared with the 20-30% disagreement rates reported by Rounds et al. (2000), the percentage in this study is not unusually high. Regrettably, many studies using judges' ratings simply do not report error rates, thus making it difficult to assess how much these errors generally influenced predictions in earlier studies (Spokane et al., 2000).

Finally, the design is longitudinal and prospective. Data were gathered from the same individuals on three occasions, thus predictor variables are separated by several years from criterion variables, which reduced the confound of participants' recollections of prior responses.

Implications for Understanding Job Satisfaction

This longitudinal study has revealed several relations that have implications for understanding job satisfaction across the life span. First, choice-job congruence does not add significantly to predictions of general job satisfaction (when earlier occurring variables are controlled) in either early adulthood or midlife or for men or women separately. Thus, the proposition that originates in developmental theory that satisfaction is proportional to self-concept implementation is not supported by data in this study. Our findings are similar to Leong and Boyle's (1997) results for a mixed-gender group using inventoried interests rather than choices over a similar span of years but during a different period in history.

The finding that congruence does not add to general job satisfaction predictions contributes nonsignificant findings to the mixed conclusions summarized by Spokane et al. (2000) and differs from earlier studies reporting significant predictions from expressed choices (e.g., Smart et al., 1986). The studies by Smart and colleagues often used multiple measures of expressed choice. The single expressed occupational choice used in this study may not sufficiently represent an adolescent's purposes as specified by life-span career theory. Additional assessments of expressed occupational choices may be warranted. For example, Swaney and Prediger (1985) have shown that "data noise"--occupational choice variables whose clarity and salience are undetermined--reduced predictions of satisfaction.

Second, the Holland type of an individual's prior choice added to midcareer job satisfaction predictions for the whole sample and for men in particular. Holland types for age 25 occupational choice added significantly to predictions

of age 43 general job satisfaction. These findings are consistent with recent trends in predicting job satisfaction at midcareer from personality factors (e.g., Judge & Larsen, 2001; Leong & Boyle, 1997; Tokar & Subich, 1997) and support a simplified test of the dispositional approach. Respondents who chose Social occupations at age 25 tended to report higher age 43 job satisfaction than did respondents who chose Realistic occupations.

We also discovered some support for a broadly interpreted "present status model" to explain midcareer job satisfaction, because age 43 current job added to predictions of job satisfaction. Respondents in Social, Enterprising, and Conventional types of occupations tended to report higher job satisfaction than did respondents in Realistic type occupations. Common features of the several jobs classified within Holland's theoretical environments contribute to predicting job satisfaction.

Third, gender was not directly associated with either congruence or job satisfaction and did not mediate predictions of congruence. On the other hand, predictions of job satisfaction were significant for men but not for women, although gender comparisons were not applied.

Fourth, the age when job satisfaction is appraised apparently does influence these predictions. Early adult (modal age 25 years) job satisfaction was less predictable than was midlife adult (modal age 43 years) job satisfaction, even though the latter predictions from occupational choice spanned 18 years compared with 7 years for the earlier period. Apparently, job satisfaction becomes more predictable at midcareer than it is in early adulthood.

Finally, although average job satisfaction scores remained the same over the adult years, average congruence increased significantly. Job satisfaction is not associated with age for this sample, in contrast to several cross-sectional studies that have revealed small, but significant, age-satisfaction relationships. Change in choice--job congruence seems to reflect the continual adaptive processes of matching choices and jobs. Evidently, experiences during the first 7 years after high school modified participants' choices so that choice--job congruence was increased for the next several years.

These findings only hint at practical applications, but two general career counseling strategies for use with adults are offered. Career counselors may find that for working-class adults, in particular, midcareer general job satisfaction is related to the type of current job and the adapted goals that they set early in their adult career rather than the choice--job match. In other words, the match may be less important than the independent job circumstances and the adult's earlier personal goals. Exploring aspects of the person's current job and the person's history of career goals separately may be a more productive strategy than approaching the match issue. Perhaps redirecting the worker's focus toward sources of satisfaction in the social aspects of a job (e.g., coworkers) and away from the job conditions (e.g., production tasks and schedules) may be fruitful. Finally, there may be accumulated resentment, or even grief, over sacrificing goals or losing jobs that affects satisfaction at midcareer.

In summary, general job satisfaction remains a complex phenomenon, the dynamics of which are only partially clarified by empirical research. Nevertheless, changes in choice--job matches over time seem to summarize the adaptive processes emphasized by adherents of developmental theory, although the specific contribution of match to satisfaction is not supported by evidence in this study.

TABLE 1

Intercorrelations Among Gender, Job Tenure, Congruence, and Job
Satisfaction at Modal Ages 25 and 43

Variable                   1       2     3      4     5

Age 25

 1. Gender (a)               --

 2. Age 25 job tenure   .09 (b)    --

 3. Age 25
    congruence (c)     -.02 (b)  -.02    --

 4. Age 25 job
    satisfaction       -.05 (b)  -.12  -.10     --

Age 43

 1. Gender (a)               --

 2. Age 25
    congruence (c)     -.02 (b)    --


 3. Age 43
    job tenure          .04 (b)  -.09    --

 4. Age 43
    congruence (d)      .06 (b)  -.02   .26 **    --


 5. Age 43 job
    satisfaction       -.03 (b)  -.05  -.13     -.03  --

Note. For age 25, n = 138 to 155; for age 43, n = 154 to 169.
Individuals who did not complete one of the questionnaires were
eliminated only from analyses requiring those data.
(a) Women = 1; men = 2. (b) Point-biserial correlation coefficient. (c)
Congruence between career choice in 12th grade and occupation in 1980.
(d) Congruence between career choice at age 25 and occupation at
age 43.
** p <.01.

TABLE 2

Hierarchical Regression Analyses Predicting Age 25 Job Satisfaction for
Total Sample and Female and Male Participants

Predictor                             B       t      p

Total sample (n = 137)

 Block 1: Gender                    0.978   1.000  .319

 Block 2: 12th-grade
   occupational choice (OC)         0.635   0.858  .393

 Block 3: Age 25 job tenure        -0.279  -1.881  .062

 Block 4: Age 25 current job (CJ)   0.017   0.032  .975

 Block 5: Age 25 congruence        -0.066  -0.303  .762

Block 6:

    Age 25 Congruence x 12th-
      Grade OC                     -0.031  -0.439  .661

    Age 25 Congruence x Age
      25 CJ                         0.035   0.496  .620

    Age 25 Congruence x 12th-
      Grade OC x Age 25 CJ         -0.002  -0.202  .840

Women (n = 92)

 Block 1: 12th-grade
  occupational choice (OC)          2.115   1.688  .095

 Block 2: Age 25 job tenure        -0.371  -1.960  .053

 Block 3: Current job (CJ)         -0.245  -0.351  .726

 Block 4: Age 25 congruence         0.477   1.073  .287

 Block 5:

   Age 25 Congruence x 12th-
     Grade OC                      -0.172  -1.442  .153

   Age 25 Congruence x Age
     25 CJ                          0.025   0.229  .819

   Age 25 Congruence x 12th-
     Grade OC x Age 25 CJ           0.005   0.355  .723

Men (n = 45)

 Block 1: 12th-grade
  occupational choice (OC)          0.182   0.151  .880

 Block 2: Age 25 job tenure        -0.045  -0.172  .864

 Block 3: Age 25 current job
  (CJ)                              0.459   0.410  .684

 Block 4: Age 25 congruence        -0.099  -0.289  .774

 Block 5:

   Age 25 Congruence x 12th-
     Grade OC                      -0.080  -0.569  .573

   Age 25 Congruence x Age
     25 CJ                          0.004   0.037  .971

   Age 25 Congruence x 12th-
     Grade OC x Age 25 CJ           0.018   0.709  .483

Predictor                          [DELTA][R.sup.2]  p([DELTA][R.sup.2]

Total sample (n = 137)

 Block 1: Gender                         .003               .518

 Block 2: 12th-grade
   occupational choice (OC)              .014               .165

 Block 3: Age 25 job tenure              .030               .044

 Block 4: Age 25 current job (CJ)        .005               .423

 Block 5: Age 25 congruence              .012               .206

Block 6:                                 .003               .945

    Age 25 Congruence x 12th-
      Grade OC

    Age 25 Congruence x Age
      25 CJ

    Age 25 Congruence x 12th-
      Grade OC x Age 25 CJ

Women (n = 92)

 Block 1: 12th-grade
  occupational choice (OC)               .010               .350

 Block 2: Age 25 job tenure              .039               .058

 Block 3: Current job (CJ)               .000               .985

 Block 4: Age 25 congruence              .002               .633

 Block 5:                                .023               .551

   Age 25 Congruence x 12th-
     Grade OC

   Age 25 Congruence x Age
     25 CJ

   Age 25 Congruence x 12th-
     Grade OC x Age 25 CJ

Men (n = 45)

 Block 1: 12th-grade
  occupational choice (OC)               .027               .279

 Block 2: Age 25 job tenure              .013               .453

 Block 3: Age 25 current job
  (CJ)                                   .094               .041

 Block 4: Age 25 congruence              .017               .381

 Block 5:                                .014               .888

   Age 25 Congruence x 12th-
     Grade OC

   Age 25 Congruence x Age
     25 CJ

   Age 25 Congruence x 12th-
     Grade OC x Age 25 CJ

TABLE 3

Hierarchical Regression Analyses Predicting Age 43 Job Satisfaction for
Total Sample and Female and Male Participants

Predictor                             B       t      p

Total sample (n = 154)

 Block 1: Gender                    0.617   1.046  .298

 Block 2: Age 25 congruence        -0.048  -0.786  .433

 Block 3: Age 25
  occupational choice (OC)          0.246   0.397  .692

 Block 4: Age 43 job tenure        -0.019  -0.602  .548

 Block 5: Age 43 current job (CJ)   0.904   1.861  .065

 Block 6: Age 43 congruence
  (Age 25 OC + Age 43 CJ)           0.100   0.629  .531

Block 7:

  Age 43 Congruence x
   Age 25 OC                        0.022   0.338  .736

  Age 43 Congruence x
   Age 43 CJ                       -0.032  -0.647  .519

  Age 43 Congruence x Age 25
   OC x Age 43 CJ                  -0.004   0.624  .534

Women (n = 96)

 Block 1: Age 25 congruence        -0.040  -0.446  .657

 Block 2: Age 25 OC                 0.649   0.746  .457

 Block 3: Age 43 job tenure         0.004   0.099  .921

 Block 4: Age 43 CJ                 1.276   2.010  .048

 Block 5: Age 43 congruence
  (Age 25 OC + Age 43 CJ)           0.340   1.438  .154

 Block 6:

  Age 43 Congruence x
   Age 25 OC                       -0.004  -0.468  .641

  Age 43 Congruence x
   Age 43 CJ                       -0.068  -1.069  .288

  Age 43 Congruence x
   Age 25 OC x Age 43 CJ            0.005  -0.060  .952

Men (n = 58)

 Block 1: Age 25 congruence        -0.015  -0.169  .867

 Block 2: Age 25 OC                 0.740   0.673  .504

 Block 3: Age 43 job tenure        -0.032  -0.670  .506

 Block 4: Age 43 CJ                 0.154   0.183  .855

 Block 5: Age 43 congruence
  (Age 25 OC + Age 43 CJ)           0.022   0.088  .930

 Block 6:

  Age 43 Congruence x
   Age 25 OC                       -0.148  -0.955  .344

  Age 43 Congruence x
   Age 43 CJ                        0.035   0.406  .687

  Age 43 Congruence x
   Age 25 OC x Age 43 CJ            0.018   0.872  .387

Predictor                          [DELTA][R.sup.2]  p([DELTA][R.sup.2])

Total sample (n = 154)

 Block 1: Gender                         .005               .387

 Block 2: Age 25 congruence              .002               .555

 Block 3: Age 25
  occupational choice (OC)               .049               .006

 Block 4: Age 43 job tenure              .009               .234

 Block 5: Age 43 current job (CJ)        .048               .005

 Block 6: Age 43 congruence
 (Age 25 OC + Age 43 CJ)                 .000               .978

Block 7:                                 .008               .708

  Age 43 Congruence x
   Age 25 OC

  Age 43 Congruence x
   Age 43 CJ

  Age 43 Congruence x Age 25
   OC x Age 43 CJ

Women (n = 96)

 Block 1: Age 25 congruence              .000               .914

 Block 2: Age 25 OC                      .023               .138

 Block 3: Age 43 job tenure              .002               .645

 Block 4: Age 43 CJ                      .032               .082

 Block 5: Age 43 congruence
   (Age 25 OC + Age 43 CJ)               .002               .636

 Block 6:                                .034               .363

  Age 43 Congruence x
   Age 25 OC

  Age 43 Congruence x
   Age 43 CJ

  Age 43 Congruence x
   Age 25 OC x Age 43 CJ

Men (n = 58)

 Block 1: Age 25 congruence              .027               .219

 Block 2: Age 25 OC                      .098               .016

 Block 3: Age 43 job tenure              .033               .152

 Block 4: Age 43 CJ                      .103               .009

 Block 5: Age 43 congruence
  (Age 25 OC + Age 43 CJ)                .003               .665

 Block 6:                                .016               .785

  Age 43 Congruence x
   Age 25 OC

  Age 43 Congruence x
   Age 43 CJ

  Age 43 Congruence x
   Age 25 OC x Age 43 CJ

TABLE 4

Analyses of Variance for the First-Letter Holland Codes of Age 25
Occupational Choice (OC) Code and Age 43 Current Job on Age 43 Job
Satisfaction (Total Sample and Male Participants)

Source           df       SS     MS      F

Total sample

 Age 25 OC (a)     3    101.88  33.96  3.79 *

 Error           155  1,387.79   8.95

Men

 Age 25 OC (b)     2     57.78  28.89  4.23 *

 Error            57    389.07   6.83

Total sample

 Age 43 current
  job code (a)     3    202.74  67.58  7.94 **

 Error           151  1,284.90   8.51

Men

 Age 43 current
  job code (c)     3    111.75  37.25  6.23 **


 Error            56    335.10   5.98

                    Scheffe Post Hoc
Source                 Comparisons

Total sample

 Age 25 OC (a)   S(21, 48)>R(19, 77) *

 Error

Men

 Age 25 OC (b)   S(22, 71)>R(19, 91) *

 Error

Total sample     S(21, 58)>R(19, 25) **

 Age 43 current
  job code (a)   E(21, 71)>R(19, 25) **

 Error           C(21, 36)>R(19, 25) *

Men

 Age 43 current
  job code (c)   E(22, 13)>R(19, 38) **

 Error

(a) Only Realistic, Social, Enterprising, and Conventional types are
included in this analysis. (b) Only Realistic, Social, and Enterprising
types are included in this analysis. (c) Only Realistic, Investigative,
Social, and Enterprising types are included in this analysis.
* p < .05.** p < .01.

References

Assouline, M., & Meir, E. I. (1987). Meta-analysis of the relationship between congruence and well-being measures. Journal of Vocational Behavior, 31, 319-332.

Breeden, S. A. (1993). Job and occupational change as a function of occupational congruence and job satisfaction. Journal of Vocational Behavior, 43, 30-45.

Brown, S. D., & Gore, P. A., Jr. (1994). An evaluation of interest congruence indices: Distribution characteristics and measurement properties. Journal of Vocational Behavior, 45, 310-327.

Clark, A., Oswald, A., & Warr, P. (1996). Is job satisfaction U-shaped in age? Journal of Occupational and Organisational Psychology, 69, 57-81.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.

Crites, J. O. (1969). Vocational psychology. New York: McGraw-Hill.

Dorman, C., & Zapf, D. (2001). Job satisfaction: A meta-analysis of stabilities. Journal of Organisational Behavior, 22, 483-504.

Edwards, J. R. (1991). Person-job fit: A conceptual integration, literature review and methodological critique. In C. L. Cooper & I. T. Robertson (Eds.), International review of industrial and organisational psychology (Vol. 6, pp. 283-357). New York: Wiley.

Elton, C. F., & Smart, J. C. (1988). Extrinsic job satisfaction and person-environment congruence. Journal of Vocational Behavior, 32, 226-238.

Fricke, M. A. M., & Beehr, T. A. (1992). A longitudinal investigation of interest congruence and gender concentration as predictors of job satisfaction. Personnel Psychology, 45, 99-117.

Gottfredson, G., & Holland, J. L. (1996) Dictionary of Holland occupational codes (3rd ed.). Odessa, FL: Psychological Assessment Resources.

Gottfredson. L. (1996). Gottfredson's theory of circumscription and compromise. In D. Brown, L. Brooks, & Associates (Eds.), Career choice and development (3rd ed., pp. 179-232). San Francisco: Jossey-Bass.

Harris, J. I., Moritzen, S. K., Robitschek, C., Imhoff, A., & Lynch, J. L. A. (2001). The comparative contribution of congruence and social support on career outcome. The Career Development Quarterly, 49, 314-323.

Hesketh, B. (2000). The next millennium of "fit" research: Comments on "The congruence myth: An analysis of the efficacy of the person-environment fit model" by H. E. A. Tinsley. Journal of Vocational Behavior, 56, 190-196.

Hobson, R. (1991). Workplace behaviors: Good soldiers, smooth operators, and saboteurs. Work and Occupations, 18, 271-290.

Holland, J. L. (1997). Making vocational choices (3rd ed.). Odessa, FL: Psychological Assessment Resources.

Hoppock, R. (1935). Job satisfaction. New York: Harper.

Jepsen, D. A. (1975). Occupational decision development over the high school years. Journal of Vocational Behavior, 7, 225-237.

Jepsen, D. A., & Choudhuri, E. (2001). Stability and change in 25-year occupational career patterns. The Career Development Quarterly, 50, 3-19.

Judge, T. A., & Larsen, R. J. (2001). Dispositional affect and job satisfaction: A review and theoretical extension. Organisational Behavior and Human Decision Process, 86, 67-98.

Kalleberg, A. L., & Loscocco, K. A. (1983). Aging, values and rewards: Explaining age differences in job satisfaction. American Sociological Review, 48, 78-90.

Kleinbaum, D. G., Kupper, L. L., & Muller, K. E. (1988). Applied regression analyses and other multivariate methods (2nd ed.). Belmont, CA: Duxbury Press.

Leong, F. T. L., & Boyle, K. A. (1997). An individual differences approach to midlife career adjustment: An exploratory study. In M. E. Lachman & J. B. James (Eds.), Multiple paths of midlife development: Studies on successful midlife development (pp. 411-451). Chicago: The University of Chicago Press.

Levinson, D. J. (1996). The seasons of a woman's life. New York: Knopf.

Lofquist, L. H., & Dawis, R. V. (1984). Research on work adjustment and satisfaction: Implications for career counseling. In S. D. Brown & R. W. Lent (Eds.), Handbook of counseling psychology (pp. 216-237). New York: Wiley.

Mannheim, B. (1993). Gender and the effects of demographics, status and work values on work centrality. Work & Occupations, 20, 3-22.

McNichols, C. W., Stahl, M. J., & Manley, T. R. (1978). A validation of Hoppock's Job Satisfaction Measure. Academy of Management Journal, 21, 737-742.

Osipow, S. H., & Fitzgerald, L. F. (1996). Theories of career development (4th ed.). Boston: Allyn & Bacon.

Rounds, J., McKenna, M. C., & Hubert, L. (2000). Tinsley on Holland: A misshapen argument. Journal of Vocational Behavior, 56, 205-215.

Smart, J. C., Elton, C. F., & McLaughlin, G. W. (1986). Person-environment congruence and job satisfaction. Journal of Vocational Behavior, 29, 216-225.

Smart, J. C., & Ethington, C. A. (1987). Occupational sex segregation and job satisfaction of women. Research in Higher Education, 26, 202-211.

Spector, P. E. (1997). Job satisfaction: Applications, assessment, cause, and consequences. Thousand Oaks, CA: Sage.

Spokane, A. R. (1985). A review of research person-environment congruence in Holland's theory of careers. Journal of Vocational Behavior, 26, 306-343.

Spokane, A. R., Meir, E. I., & Catalono, M. (2000). Person-environment congruence and Holland's theory: A review and reconsideration. Journal of Vocational Behavior, 57, 147-179.

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

Super, D. E., Savickas, M. L., & Super, C. M. (1996). The life-span, life-space approach to careers. In D. Brown, L. Brooks, & Associates (Eds.), Career choice and development (3rd ed., pp. 121-178). San Francisco: Jossey-Bass.

Swaney, K., & Prediger, P. J. (1985). The relationship between interest-occupation congruence and job satisfaction. Journal of Vocational Behavior, 26, 13-24.

Tinsley, H. E. A. (2000). The congruence myth: An analysis of the efficacy of the person-environment fit model. Journal of Vocational Behavior, 56, 147-179.

Tinsley, H. E. A., & Weiss, D. J. (1975). Interrater reliability and agreement of subjective judgments. Journal of Counseling Psychology, 22, 358-376.

Tokar, D. M., & Subich, L. M. (1997). Relative contributions of congruence and personality dimensions to job satisfaction. Journal of Vocational Behavior, 50, 482-491.

Tranberg, M., Slane, S., & Ekeberg, S. E. (1993). The relation between interest congruence and satisfaction: A meta-analysis. Journal of Vocational Behavior, 42, 253-264.

U.S. Department of Labor. (1991). Dictionary of occupational titles (4th ed.). Indianapolis, IN: JIST Works.

Weaver, C. N. (1980). Job satisfaction in the United States in the 1970s. Journal of Applied Psychology, 65, 364-367.

David A. Jepsen, Division of Counseling, Rehabilitation and Student Development, The University of Iowa; Hung-Bin Sheu, Department of Counseling and Personnel Services, The University of Maryland--College Park. This article is part of a project titled "Occupational Career Development From Rural Beginnings," supported by timely and generous grants from the University of Iowa Graduate College and the Spencer Foundation. The data and conclusions are solely the responsibility of the authors. This article is based on a paper presented at the 2001 annual conference of the American Counseling Association, San Antonio, Texas. Correspondence concerning this article should be addressed to David A. Jepsen, N338 Lindquist Center, University of Iowa, Iowa City, IA 52242 (e-mail: [e-mail]).