Job satisfaction among TANF leavers.

Scott, Jeff

Using means tests, ANOVA, contingency methods and polytomous logistic regression techniques, I analyze job satisfaction survey data provided by former welfare recipients in Illinois. Mean job satisfaction in the sample is high. Wages, work hours, professional status, having employer sponsored health care and being in good health have significant positive effects on job satisfaction. Contrary to popular assumptions regarding welfare dependency, time on welfare positively affects post-TANF job satisfaction. I discuss implications of these findings in the context of policy debates regarding TANF reauthorization.

Keywords: welfare reform, the working poor, job satisfaction

**********

The 1996 Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) and the subsequent implementation of state Temporary Assistance for Needy Families (TANF) programs have drawn attention to the lives of the working poor. Caseloads fell dramatically after welfare reform (Blank, 2001; Moffit, 1999), and many who left the rolls found jobs (Loprest, 1999; U.S. General Accounting Office, 1999). The work experiences of TANF leavers, specifically their incomes and job tenures, have been closely monitored (Acs, Loprest, & Roberts, 2001). Expected to ease the transition from welfare to work, access to health insurance and child care programs have been studied extensively (Currie, & Yelowitz, 2000; Garret, & Holahan, 2000). The impact of parental employment on children's behavior also has received consideration (Duncan, & Chase-Lansdale, 2001).

Despite all the attention provided to understanding dimensions of work in the era of welfare reform, a basic issue has been overlooked: Do TANF leavers like their jobs? The answer to this question is important from a policy perspective, as previous congressional debates about work and welfare and current discussions regarding PRWORA reauthorization have proceeded in an atmosphere of competing assumptions regarding the poor.

A long-held, popular belief is that poor people are unmotivated and unwilling to meet the demands of the jobs available to them (Lewis, 1959; 1966). The argument as it relates to welfare is that dependency on government assistance primarily results from the reluctance of able-bodied adults to work continuously in the bottom segment of the labor market (Mead, 1986; 1992; Murray, 1984). Corresponding to this viewpoint, the PRWORA welfare reform policy emphasis was focused on eliminating entitlement to assistance, reducing welfare caseloads, enforcing work requirements and promoting self-sufficiency, although a provision to create jobs was notably absent from the legislation (Katz, 2001).

While sometimes conceding that welfare dependency is problematic and that reform was in order, some have downplayed the significance of work in terms of poverty reduction, partly because many of the poor have extensive employment histories. Edin and Lein (1997) revealed that before PRWORA many people used cash assistance to complement income earned from employment, a finding that undercut the notion that welfare reform was necessary to promote work and self-sufficiency. Alternatively, liberals stressed the need for a new covenant between welfare recipients and the government, with the former providing work effort in return for job creation, improved support services and tax credits to protect against poverty conditions (Elwood, 1988). Their mantra has been to "make work pay."

If lack of personal motivation among the poor is prominent, one would anticipate that dissatisfaction with "bad" jobs partially accounts for the erratic work records observed within the low-income population (Glazer, 1969). In contrast, those emphasizing restricted opportunities contend that the working poor typically take pride in their work efforts and derive personal satisfaction from their jobs, though they may rightfully object to the limited financial returns produced by working (DeParle, 2004, Newman, 1999). Both perspectives make assumptions about poor people's work perceptions that have not been adequately established. As such, this study assesses job satisfaction and its determinants among a sample of former TANF recipients in one state.

Literature Review

Low-Income Work After TANF

Most former TANF recipients left welfare for work-related reasons. Analyzing data from a national sample of welfare leavers, Loprest (1999) found that 69% stopped getting public assistance either because they obtained a job or their earned income from work improved. Similarly, Acs, Loprest, and Roberts (2001) synthesized findings from a set of 15 federally funded, state level TANF leaver studies and discovered that 71% of leavers were employed at some point in the year following their exits. Nonetheless, leavers have generally received low wages. Shortly after PRWORA, the first state TANF reports showed that a substantial number of working leavers earned less than $6 an hour (Parrott, 1998). More recent monitoring at the state level revealed that median hourly wage levels remained low, between $7 and $8 (Acs et al., 2001).

Mirroring recent employment and earnings trends, income levels among the poor have risen. In addition, the 1999 poverty rate for female-headed households, the population typically most reliant on cash assistance, was at its lowest level in 40 years (Haskins, & Primus, 2001). But income gains among TANF leavers have been modest. In fact, the most robust finding across leaver studies is that families almost always remained low-income after leaving welfare (Acs et al., 2001; Grogger, Karoly, & Klerman, 2002; Loprest, 1999; Polit, Widom, Edin, Bowie, London, Scott, & Venezuela, 2001).

These limited earnings likely relate to the kind of jobs available to those moving from welfare to work. Studies have found that leavers normally gained employment in the service sector, where work commonly involved sales, food preparation or clerical support (Loprest, 1999; Parrot, 1998). Research has also shown that jobs available to leavers typically lack fringe benefits, including paid vacation and sick days, retirement pensions and health insurance (Acs et al., 2001).

Information about the job tenures of former welfare recipients is mixed and sometimes limited because observations were made shortly after welfare reform began and before the economic recession that hit five years later. Loprest (1999) found that almost three quarters of working leavers had been on their current jobs for less than a year and a third for less than six months, whereas Polit et al. (2001) indicated that many held the same jobs for at least 19 months during a two year study period. On the other hand, numerous leaver studies highlight job instability as a significant problem: In eight states that were part of the Acs et al. (2001) report, only 37 percent of leavers were employed in all four quarters in the year following TANF exit.

Underlying these sporadic employment patterns are the obstacles faced by the working poor. Issues related to childcare have been linked to leaver's employment outcomes (Julnes, Halter, Anderson, Frost-Kumpf, Schuldt, Staskon, & Ferrara, 2000). The disappearance of jobs from the inner city can create a transportation burden for those leavers who cannot find work in close proximity to their homes (Mancuso, Lieberman, Lindler, & Moses, 2001). Health and mental problems are often correlated with living in poverty conditions and could also stand in the way of employment consistency (Kalil, Schweingruber and Seefeldt, 2001). Even when leavers manage to overcome barriers and maintain employment, experiences of food and housing crises are not uncommon (Acs et al., 2001; DeParle, 2004). Despite these limitations, employers have noted the productivity of leavers. Meyer (1999) found that a majority of employers express positive attitudes about workers who formerly received welfare and are compelled to continue hiring them in the future.

Job Satisfaction

Research on job satisfaction has been popular among educational and industrial psychologists, and a vast literature has evolved (Locke, 1976). Measured in different ways and within various employment settings, job satisfaction has been consistently identified as an important predictor of work behavior. Most notably, employees who choose to quit (Ackerlof, Rose, & Yellen, 1988; Freeman, 1978; McEvoy, & Cascio, 1985) or be absent from work (Clegg, 1983; Drago, & Wooden, 1992) have low job satisfaction.

Many behavioral researchers have investigated the sources of job satisfaction. A problem common to such studies is that working people often report fairly high job satisfaction levels. Therefore, it is sometimes difficult to obtain statistically significant results when treating job satisfaction as a dependent variable. Nevertheless, researchers have discovered job satisfaction determinants related to both personal and work-related variables.

Being female, married and having good health have all been associated with higher levels of job satisfaction (Clark, 1996; Clark, Oswald, & Warr, 1996; Idson, 1990; Meng, 1990). A positive relationship between age and job satisfaction is a common finding, although there has been debate as to whether it is linear (Janson, & Martin, 1982; Kalleberg, & Loscocco, 1983) or U-shaped (Clark, Oswald, & Warr 1996; Warr, 1992). Race seems less useful in predicting job satisfaction (Bartel, 1981; Idson, 1990), and distinctions in job satisfaction among workers with varying educational backgrounds are unclear. It has been demonstrated that education translates into high earnings and upward mobility, but its correlation with job satisfaction usually has been negative (Clark, 1996; Clark, Oswald, & Warr, 1996; Meng, 1990). A possible explanation is that people who receive advanced educations have higher job expectations, which if unfulfilled result in diminished satisfaction with work (Clark, & Oswald, 1996).

Job satisfaction is also produced by certain workplace conditions. For instance, having flexibility and control over one's work activity has been connected to higher job satisfaction levels (Idson, 1990). A direct relationship between wages and job satisfaction has not been found, however workers tend to measure their earnings and benefits in relation to their peers or the market's "going rate", and the correlation between perceived equity of a job's economic returns and job satisfaction is strongly positive (Cappelli, & Sherer, 1988; Clark, & Oswald, 1996). Additionally, Brown and McIntosh (2003) noted that several studies link elevated job satisfaction levels to having a senior position, reception of job training, perceived opportunity for advancement and job tenure.

Other job attributes may diminish job satisfaction. Workers at large firms have reported relatively low job satisfaction (Clark, 1996; Clark, Oswald, & Warr, 1996). Correspondingly, Idson (1990) has observed that big corporations are more apt to enforce rigid work rules, a practice which detracts from employee satisfaction. Numerous studies have shown that union jobs produce less satisfaction (Borjas, 1979; Freeman, 1978; Meng, 1990; Schwochau, 1987). Some have argued that union workers choose to voice workplace discontent while nonunion workers just decide to quit jobs they don't like (Freeman, 1980; Freeman, & Medoff, 1984). This exit-voice theory has been challenged by Gordon and Denisi (1995), who found no significant ties between union membership and job satisfaction, as well as by Bender and Sloan (1998), who concluded that dissatisfaction among union workers simply stems from antagonistic industrial relations between labor and management.

The need for further research

Extensive research has been done on welfare reform, but levels of post-TANF job satisfaction have not been well assessed. Likewise, job satisfaction determinants among the leaver population have not been identified. Typically, job satisfaction studies are focused broadly on national samples or narrowly targeted on selected professions. As a result, connections between job satisfaction and socio-economic status are not very clear (Ritter & Anker, 2002). This article addresses these gaps in the literature by examining job satisfaction levels and factors associated with job satisfaction among former welfare recipients in one state.

Methods

Sample and Interviewing

The sample was drawn from 8,804 adults who exited the Illinois TANF program in December of 1998. Consistent with other federally funded leaver studies (National Research Council, 1999), eligible survey respondents were those whose welfare cases had remained closed for at least two months following their initial TANF exits. The resulting non-proportional, stratified sample consisted of 500 Cook County (Chicago area) leavers and 501 leavers from the rest of the state (downstate). The downstate portion of the sample was further proportionately stratified by selected urban and rural areas. To account for the intentional over-sampling, I used weighted data for this study.

Six to eight months after their TANF exits, interviews were obtained from 514 (51.3%) sample members. Illinois Department of Human Services (IDHS) administrative data allowed for comparisons between respondents and non-respondents: They were virtually the same in terms of gender, age, TANF history, and household composition. Hispanics were slightly underrepresented as were the highly uneducated, but most disparities between respondents and non-respondents were not significant. Some sample members were continuously unemployed and were excluded from the study as were three respondents who refused to report job satisfaction data. This left a subsample of 434 working leavers.

The survey consisted of questions drawn from an earlier IDHS instrument as well as from similar TANF studies conducted in Michigan, Wisconsin and South Carolina. Once constructed, the instrument was pre-tested with a random sample of leavers having characteristics similar to those of the study respondents. The study was administered by trained survey research staff using Computer Assisted Telephone Interviewing (CATI).

Variables

Working leavers were asked the following question: "How satisfied or dissatisfied (are/were) you with your current or most recent job?" A five-point scale measured job satisfaction levels (very satisfied, somewhat satisfied, neither satisfied nor dissatisfied, somewhat dissatisfied or very dissatisfied) and produced the study's dependent variable.

I followed a framework used by Clark (1996) in which the dependent variables thought to predict job satisfaction are classified as either personal characteristics (race/ethnicity, family size and structure, education, personal health, etc.) or work-related characteristics (wages, job tenure, hours worked per week, etc.). In addition, I used a third classification composed of welfare-related variables (time on welfare and reason for exit). I grouped the variables in this manner so that the findings could be reported efficiently, and to observe whether one set of variables or another was more important in terms of predicting job satisfaction.

Analysis

First, I obtained frequency distributions of job satisfaction and of the categorized respondent characteristics treated as independent variables in subsequent analyses. This permitted assessment of overall job satisfaction in the sample and yielded a demographic profile of early, working TANF leavers in Illinois.

I next considered group differences in mean job satisfaction by applying independent samples T-tests (on dichotomous variables) and analyses of variance techniques (on variables having more than two categories). With cross tabulation procedures, I measured the association of job satisfaction with each of the individual, welfare and work characteristics. Given the manner in which they are categorized, job satisfaction and most of the characteristic variables could reasonably be treated as ordinal. Moreover, dichotomous measures can be appropriately treated as ordinal when using contingency methods (Agresti, & Finley, 1997). The corresponding test statistic for ordinal by ordinal cross classification is gamma, which is parallel to chi-square, but preferable in this case because it is insensitive to small expected cell frequencies. In relation to job type, residential location, race/ethnicity and living arrangements, all of which cannot be considered ordinal variables, job satisfaction was treated as nominal in order to accommodate nominal by nominal cross classifications.

After testing for mean differences and assessing the strength of the association between job satisfaction and the selected characteristic variables, further analysis was conducted with a multivariate application. The ordered dependent variable complicated the selection of a regression technique. Since there were only five job satisfaction categories and because responses were concentrated in two categories, an ordinary least squares (OLS) routine was not advisable (Agresti, 2002; Menard, 2001). Instead, I constructed a polytomous logistic regression model to estimate the effects of the independent variables on the log odds of having higher rather than lower job satisfaction scores. The Wald statistic generated for each independent variable in the model has a chi-square distribution with one degree of freedom and was used to determine if an effect was statistically significant. The model itself performed fairly well, as indicated by the highly significant model chi-square and the Nagelkerke pseudo R-square (.209).

Results

The distribution of job satisfaction in the sample is reflected in Table 1. Almost 80% of respondents report being very or somewhat satisfied with their job, while just over 15% report job dissatisfaction at some level.

Table 2 presents frequency distributions of the categorized respondent characteristics treated as independent variables in later analyses. Respondents typically are African American, single mothers, caring for one or two children. Three fourths of the sample attained at least a high school degree or its equivalent. Before exiting the rolls mainly due to enhanced income or caseworker diversion, 43% received welfare for a period longer than five years. It appears that work following welfare usually did not pay well, was not full-time, offered few benefits and was unstable. Sample members most often worked in the service industry, earned median wages of just over $7 an hour and held their current or most recent job for less than six months. Only 22% benefited from employer sponsored health plans, and a smaller number (15%) received on-the-job training. When asked what they like best about their jobs, 38% of the sample reported social aspects, a fourth appreciated the type or difficulty of work and a fifth valued economic returns.

For each characteristic, job satisfaction means are reported as are the percentages somewhat or very satisfied with their jobs (Table 3). Most significant relationships appearing in Table 3 carry over to the polytomous logistic regression of job satisfaction, the results of which are presented in Table 4. Based on the information contained in these tables, I now consider the job satisfaction findings specific to each of the three characteristic types.

Job Satisfaction and Individual Characteristics

A significant positive relationship exists between job satisfaction and self-reported health. As can be seen in Table 3, of those believing themselves to be in very good health, 84% say they are somewhat or very satisfied on the job, with a mean satisfaction score of 4.21. A much lower percentage of respondents in poor or very poor health (60% and 67% respectively) report higher levels of job satisfaction, and their job satisfaction means (3.30 and 3.64 respectively) are the lowest among the subgroups. Controlling for the other characteristics in the model, the regression analysis provides further evidence of a link between health and job satisfaction. Health effects in Table 4 are statistically significant, and their signs as expected are negative. As such, perceived poor health significantly decreases the probability of reporting higher levels of job satisfaction.

For the most part, the influence of other personal characteristics on job satisfaction is minimal. There are gender differences regarding job satisfaction, yet due to the small number of men in the leaver sample these are not statistically significant. Family size and composition does not appear to impact job satisfaction. There is little variation in job satisfaction among those having more or less education. Caucasians report slightly lower job satisfaction levels than did Blacks and Latinos, but meaningful distinctions in job satisfaction in terms of race and ethnicity cannot be made. Likewise, there is little evidence to imply regional differences in job satisfaction.

Job Satisfaction and Welfare Characteristics

The Table 3 and Table 4 data point to a positive relationship between time spent on welfare and job satisfaction. Mean job satisfaction is lowest (3.86) for those who had been on welfare for less than two years prior to exit, while the mean satisfaction score for individuals who received welfare longer than five years is 4.15. Furthermore, the Table 3 cross tabulations indicate a statistically significant difference of almost 15 percentage points between short- and long-term welfare recipients. When other variables are held constant in the regression, the same picture emerges. Here, the significant positive coefficient on TANF tenure suggests greater job satisfaction the longer a person has been on welfare.

Connections between self-reported reasons for leaving welfare and post-TANF job satisfaction are less clear. At the bivariate level, there is only statistical evidence that being diverted off of welfare by a caseworker leads to subsequent job satisfaction. However, this relationship disappears in the regression analysis. Time limits and work requirements, the two most distinctive features of welfare reform, do not significantly affect post-TANF job satisfaction.

Job Satisfaction and Employment Characteristics

Taken together, the bivariate results and the regression analysis provide strong support for the hypothesis that more pay is associated with greater job satisfaction. Workers with earnings at or below the 1999 federal and state minimum wage ($5.15) have the lowest mean satisfaction score (3.28) displayed in Table 3. Likewise, just 63.5% of the lowest income group report being somewhat or very satisfied with their jobs, as compared to well over 80% for the higher earnings groups. Income from work also turns out to be a very strong, positive job satisfaction predictor in the Table 4 regression results.

Leaver job satisfaction also is related to the number of hours worked. In Table 3, leavers working less than 20 hours per week have a mean satisfaction score of 3.87, and about 77% of them report higher levels of job satisfaction. For those working more regularly but not full time, mean job satisfaction (3.62) and the percentage reporting a higher job satisfaction level (67.6%) is lower. Job satisfaction appears highest among the workers whose schedules approach, meet or exceed fulltime status. These findings imply a U-shaped relationship, and therefore, squared hours worked per week was included in a preliminary regression analysis as an independent variable. However, the corresponding coefficient was not significant and its direction provided no further evidence of a nonlinear relationship.

The leavers in this study are employed in various types of jobs, and a few distinctions in job satisfaction can be made. Job satisfaction is most pronounced among the small segment of the sample whose work is classified as professional/technical. Mean job satisfaction (4.80) and the percentage somewhat or very satisfied (91.6) are higher for these workers in comparison to leavers employed in all other fields. The largest group of leavers found jobs in the service industry, and mean job satisfaction (3.83) is considerably lower for this group, of which a quarter reported dissatisfaction with work. With service sector work as the reference category in the regression, the coefficients for the higher status jobs are positive and significant.

In comparison to leavers who are uninsured, on Medicaid or who have purchased private health coverage independently, job satisfaction is significantly higher when employers contribute to health care expenses. Table 3 shows that mean job satisfaction is 4.24 among leavers insured through employer-sponsored health plans and 3.95 for leavers not receiving this work-related benefit. This relationship is significant at the bivariate level, but not in the regression analysis.

While on the surface the data in Table 3 and Table 4 suggest several more relationships between job satisfaction and work characteristics, statistical evidence to support the following claims is weak. First, leavers with just a single employer are marginally more satisfied than those holding multiple jobs. Also, job satisfaction is a little higher among sample members who had received job training. Recent employment as well as stable jobs held for a year or longer both produce comparable amounts of job satisfaction. Finally, leavers who travel 30 minutes to an hour to work have higher job satisfaction than do leavers with shorter and longer commutes.

Discussion

The terms of employment for many welfare recipients are forced in relation to work requirements connected to TANF. Moreover, work available to leavers is often found in the service industry and tends to pay substandard wages with few, if any, tangible fringe benefits. Consequently, when the context of jobs taken by TANF leavers is considered, the reports of high job satisfaction in this study are striking.

Several factors might explain the unbalanced distribution of job satisfaction in the sample. For instance, Marxist theory holds that individuals are inspired to work by the intrinsic aspects of any job and take pleasure in accomplishing job tasks that might otherwise be regarded as mundane (Marx, 1990). It also is important to remember that 15% of the sample was continuously unemployed and not a part of the study. There may be a self-selection bias when working leavers are examined exclusively. In addition, an uncommonly strong economy coincided with TANF implementation. We know this played a significant role in the caseload decline (Blank, 2001), and as well it may have influenced leaver work perceptions. The debate as to whether PRWORA was helpful or harmful still rages, but at the onset of welfare reform everyone, including those receiving cash assistance, was in agreement--getting welfare was dehumanizing and change was needed (Elwood, 1988). Job satisfaction among leavers may be high simply because past experiences in welfare offices have been so bad.

Discovering high job satisfaction levels in the sample is especially interesting given the widespread notion that the poor lack interest in working. A popular belief that fueled the original debate on the need to reform the American welfare system was that dependency on the government erodes positive attitudes about work. A key finding in this study suggests this assertion lacks empirical credibility: Leavers with the longest TANF tenures are more likely to report higher job satisfaction levels. Similarly, the job satisfaction among leavers documented in this article is inconsistent with some stereotypes regarding the hierarchical classification of American jobs and the people who hold them. It seems that low-prestige jobs are not necessarily limited in terms of providing personal fulfillment. To some this finding might be used to challenge the practice of making policy based on unfounded claims about the work attitudes of the poor. Welfare reform proponents might favor a different interpretation. Since welfare recipients have been shown to enjoy work, even when it is found at the bottom of the labor market, some might be inclined to advocate tougher work requirements and more abbreviated time limits on benefit reception. Proponents of such policy changes should be cautioned that high job satisfaction does not remedy the inadequate material rewards produced by work after welfare.

Unlike research on different employee populations, I find very few connections between the personal characteristics of TANF leavers and job satisfaction. An important exception is health status, as those suffering from poor or very poor health are significantly less likely to report higher levels of job satisfaction. In addition, some evidence suggests that lack of private health care undermines job satisfaction among TANF leavers. Taken together, these findings underscore a fundamental issue confronting American society--the need to extend quality and affordable health care to all citizens.

Overall, since factors such as race, education and family composition do not appear to be determining factors, it is more fruitful to consider job satisfaction in relation to the structural nature of leaver employment for purposes of identifying social welfare policy issues. Of course, one would expect high pay, benefit reception and favorable working conditions to translate into elevated job satisfaction levels. This study confirms such presumptions. There is limited evidence to suggest that employer-sponsored health plans boost job satisfaction, but, more notably, the finding that better pay is related to greater job satisfaction deserves attention. Scholars and policymakers, noting the inability of work alone to protect TANF leavers against poverty, have endorsed the idea of raising the national minimum wage in the process of reauthorizing welfare reform legislation. Increasing the Earned Income Tax Credit (EITC) at both the federal and state levels would clearly benefit low-income workers in this respect as well. In addition, this study suggests that these policy decisions could improve the subjective feelings held by TANF leavers about their jobs.

Largely because so many people left the rolls in the wake of PRWORA, welfare reform has been perceived as a great success. Hence, now that it has come time to reauthorize the legislation, welfare reform advocates contend that the core components of the original law should be broadened in scope. Specifically, some lawmakers want to elevate the hours of work activity required by individuals for benefit reception and raise the expectations of the states to engage people in such activities. The clear objective is to force poor people into consistent attachment to the mainstream job market, but the underlying assumption is that they don't like working any more than they are required. The results presented above suggest otherwise, as working full-time or at least 30 hours per week had a significant positive impact on leaver job satisfaction. While reauthorization debate is likely to center around making TANF recipients work more hours whether they like it or not, policy choices would be more appropriately directed at promoting the availability of full-time employment opportunities.

Expanded demand for labor in the service sector has coincided with the movement of people off of welfare and into jobs, but employment conditions at the bottom of the labor market are such that work and poverty are still not mutually exclusive. Under these circumstances, welfare-to-work programs represent an uneven contract between the federal government and the poor. Following the 1996 welfare reforms, many recipients exited the rolls and entered the workforce, but low wages and employment instability have limited their rewards. Even so, this study shows that leavers are satisfied with the jobs they have taken. The determinants of the job satisfaction observed in the leaver sample indicate that policy approaches to enhance the material rewards of low-income work may also enhance the subjective rewards of working.

Limitations and Research Implications

The survey data used for this study were collected from individuals a short period after they had exited the welfare system. While this has been a common approach for evaluating welfare reform, longitudinal studies are necessary to document the long-term policy impacts. It is especially important to pay attention to job satisfaction over time as it tends to change in relation to age, job tenure and professional development (Locke, 1976).

A common theoretical framework underlies many job satisfaction studies, as personal characteristics and aspects of the work environment are thought to affect job satisfaction. Less attention has been given to the role of politics, culture and community (Seashore & Taber, 1975). This article makes a contribution by considering a few variables related to welfare reform, an issue that is clearly political and cultural. Finding that time on welfare translates into higher job satisfaction is notable and worth discussion, but one might pose other questions about welfare reform and job satisfaction. For instance, does reception of certain in-kind benefits or access to other social programs have a bearing upon job satisfaction? The social policy setting in Illinois is relatively progressive. Besides Food Stamps and the requisite transitional Medicaid, Illinois boasts a well developed child care subsidy system, broad state health insurance programs, improved public transportation and a state-level earned income tax credit. Intended to ease the burdens of the working poor, these social supports probably promote job satisfaction at some level.

Because the leavers' personal characteristics did not seem to affect their job satisfaction, I focused on influential structural variables of employment, such as wages, benefits, and working conditions. While job training was examined as a potential job satisfaction predictor, I did not fully investigate the significance of advancement opportunities more generally. This omission is important given that job ladders have become shorter due to the choices employers often make in response to economic globalization and deregulation (Appelbaum, Bernhardt, & Murnane, 2003).

Because job satisfaction is subjectively determined, some scholars have questioned its meaning (Hodson, 1991). Although used with great frequency by job satisfaction scholars and relied upon in this study, one-dimensional job satisfaction measures are especially vulnerable to such criticism. Some researchers, consequently, have disaggregated job satisfaction through factor analysis into several components, such as pay, work intensity, social conditions, long-term prospects and other facets (Schwochau, 1987). Building on the findings presented and discussed in this article, research exploring multidimensional job satisfaction among leavers would improve our understanding about the nature of post-TANF employment.

References

Acs, G., Loprest, P., & Roberts, T. (2001). Final synthesis report of findings from ASPE "leavers'" grants. Retrieved on October 1, 2004, from http://aspe.hhs.gov/hsp/leavers99/synthesis02/index.htm

Agresti, A., (2002). Categorical data analysis: Wiley series in probability and statistics [2nd ed.]. Hoboken, NJ: Wiley Interscience.

Agresti, A., & Finlay, B. (1997). Statistical Methods for the Social Sciences [3rd ed.]. Upper Saddle River, NJ: Prentice Hall.

Akerlof, G. A., Rose, A. K., & Yellen, J. L. (1988). Job switching and job satisfaction in the U.S. labor market. In Brookings papers on economic activity (No. 2, pp. 495-582). Washington, D.C.: The Brookings Institution.

Appelbaum, E., Bernhardt, A., & Murnane, R. J. (2003). Low-Wage America: How employers are reshaping opportunity in the workplace. New York: The Russell Sage Foundation.

Bartel, A. (1981). Race differences in job satisfaction: A reappraisal. Journal of Human Resources, 14, 21-40.

Bender, K., & Sloane, P. (1998). Job satisfaction, trade unions, and exit-voice revisited. Industrial and Labor Relations Review, 51, 222-240.

Blank, R. M. (2001). Welfare and the economy. Washington D. C.: The Brookings Institution.

Borjas, G. J. (1979). Job satisfaction, wages, and unions. Journal of Human Resources, 14(1), 21-40.

Brown, D., & McIntosh, S. (2003). Job satisfaction in the low wage service sector. Applied Economics, 35, 1241-1254.

Capelli, P., & Sherer, P. (1988). Satisfaction, market wages, and labor relations: An airline study. Industrial Relations, 27(1), 56-73.

Clark, A. E. (1996). Job satisfaction in Britain. British Journal of Industrial Relations, 34 (2), 189-217.

Clark, A. E., & Oswald, A. J. (1996). Satisfaction and comparison income. Journal of Public Economics, 61,359-381.

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

Clegg, C. W. (1983). Psychology of employee lateness, absence, and turnover: A methodological critique and an empirical study. Journal of Applied Psychology, 68, 88-101.

Currie, J., and Yelowitz, A. (2000). Health insurance and less skilled workers. In D. Card, & R. M. Blank (Eds.), Finding jobs: Work and welfare reform (pp. 233-261). New York: Russell Sage.

DeParle, J. (2004). American dream: Three women, ten kids, and a nation's drive to end welfare. New York: Viking Books.

Drago, R., & Wooden, M. (1992). The determinants of labor absence: Economic factors and workgroup norms across countries. Industrial and Labor Relations Review, 45, 764-778.

Duncan, G. J., & Chase-Lansdale, P. L. (2001). Welfare reform and child wellbeing. In R. B. Blank, and R. T. Haskins (Eds.), The new world of welfare (pp. 391-417). Washington, D. C.: The Brookings Institution.

Edin, K., & Lein, L. (1997). Making ends meet: How single mothers survive on welfare and low-wage work. New York: Russell Sage Foundation.

Ellwood, D. T. (1988). Poor support: Povertiy in the American family. New York: Basic Books.

Freeman, R. B. (1978). Job satisfaction as an economic variable. American Economic Review, 68, 135-141.

Freeman, R. B. (1980). The exit-voice trade off in the labor market: Unionism, job tenure, quits. In National Bureau of Economic Research Working Papers. Retrieved on October 8, 2004, from http://ideas.repec.org/e/pfr23.html

Freeman, R. B., & Medoff, J. L. (1984). What do unions do? New York: Basic Books.

Garret, B. & Holahan, J. (2000). Welfare leavers, Medicaid coverage and private health Insurance. Retrieved on September 29, 2004, from http//:newfederalism.urban.org/html/series_b/b13/b13.html

Glazer, N. (1969). Beyond income maintenance. The Public Interst, 16, 102-122.

Gordon, M. E., & Denisi, A. S. (1995). A re-examination of the relationship between union membership and job satisfaction. Industrial and Labor Relations Review, 48(2), 222-236.

Groger, J. Karoly, L. A., & Klerman, J. A. (2002). Consequences of welfare reform: A research synthesis. Retrieved on October 1, 2004, from http://www.acf.hhs.gov/programs/opre/welfare_reform/reform_cover.html

Haskins, R., & Primus, W. (2001). Welfare reform and poverty. Retrieved on September 10, 2004, from http://www.brookings.edu/es/research/projects/wrb/publications/pb/pb04.htm

Hodson, R. (1991). Workplace behaviors: Good soldiers, smooth operators, and saboteurs. Work and Occupations, 18(3), 271-290.

Idson, T. L. (1990). Establishment size, job satisfaction and the structure of work. Applied Economics, 22(8), 1007-1018.

Janson, P., & Martin, J. K. (1982). Job satisfaction and age. Social Forces, 60(4), 1089-1102.

Julnes, G., Halter, A., Anderson S., Frost-Kumpf, L., Schuldt, R., Staskon, F., & Ferrara, B. (2000). Illinois study of former TANF clients, final report. Springfeild, Illinois: Institute for Public Affairs, University of Illinois at Springfield, and School of Social Work, University of Illinois at Urbana-Champaign.

Kalil, A., Schweingruber, H., & Seefeldt, K. (2001). Correlates of employment among welfare recipients: Do psychological characteristics and attitudes matter? American Journal of Community Psychology, 29, 701-723.

Kalleberg, A., & Loscocco, K. A. (1983). Age differences in job satisfaction. American Sociological Review, 48(1), 78-90.

Katz, M. B. (2001). The price of citizenship: Redefining the American welfare state. New York: Owl Books.

Lewis, O. (1959). Five families: Mexican case studies in the culture of poverty. New York: Basic Books.

Lewis, O. (1966). La vida: A Puerto Rican family in the culture of poverty. New York: Random House.

Locke, E. A. (1976). The nature and causes of job satisfaction. In M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 1279-1349). Chicago: Rand McNally.

Loprest, P. (1999). Families who left welfare: Who are they and how are they doing? Retrieved September 23, 2004, from http://www.urban.org/UploadedPDF/discussion99-02.pdf

Mancuso, D. C., Lieberman, C. J., Lindler, V. L., & Moses, A. (2001). Examining circumstances of individuals and families who leave TANF: Assessing the validity of administrative data, final 18-month report. Burlingame, CA: SPHERE Institute.

Marx, K. (1990). Capital: A critique of political economy [vol. 1]. New York: Penquin Books.

McEvoy, G. M., & Cascio, W. F. (1985). Strategies for reducing employee turnover: A meta-Analysis. Journal of Applied Psychology, 70, 342-353.

Mead, L. M. (1986). Beyond entitlement: The social obligations of citizenship. New York: Free Press.

Mead, L. M. (1992). The new politics of poverty: The nonworking poor in America. New York: Basic Books.

Menard, S. (2001). Applied logistic regression analysis: Quantitative applications in the social sciences [2nd ed.]. Thousand Oaks, CA: Sage.

Meng, R. (1990). The relationship between unions and job satisfaction. Applied Economics, 22(12), 1635-1648.

Meyer, J. A. (1999). Assessing welfare reform: Work pays. The Public Interest, 136, 113-120.

Moffit, R. A. (1999). The effect of pre-PRWORA waivers on AFDC caseloads and female earnings, income and labor force behavior. In S. H. Danziger (Ed.), Economic Conditions and Welfare Reform (pp. 19-47). Kalamazoo, MI: Upjohn Institute for Employment Research.

Murray, C. (1984). Losing ground: American social policy 1950-1980. New York: Basic Books.

National Research Council (1999). Evaluating welfare reform: A framework and review of current work. Interim report. Panel on Data and Methods for Measuring the Effects of Changes in Social Welfare Programs. R. A. Moffit & M. Ver Ploeg, eds. Committee on National Statistics, Commission on Behavioral and Social Sciences and Education. Washington D.C.: National Academy Press.

Newman, K. S. (1999). No shame in my game: The working poor in the inner city. New York: Russell Sage Foundation.

Parrott, S. (1998). Welfare recipients who find jobs: What do we know about their employment and earnings? Retrieved on October 1, 2004, from http://www.cbpp.org/11-16-98wel.htm

Polit, D. F., Widom, R., Edin, K., Bowie, S., London, A. S., Scott, E. K., & Valenzuela, A. (2001). Is work enough? The experiences of current and former welfare mothers who work. Retrieved on October 1, 2004, from http://www.mdrc.org/publications/74/execsum.html

Ritter, J. A., & Anker, R. (2002). Good jobs, bad jobs: Worker's evaluations in five countries. International Labour Review, 141 (4), 331-358.

Seashore, S. E., & Taber, T.D. (1975). Job satisfaction indicators and their correlates. American Behavioral Scientist, 18(3), 333-368.

Schwochau, S. (1987). Union effects on job attitudes. Industrial and Labor Relations Review, 40, 209-224.

United States General Accounting Office. (1999). Welfare reform: Information on former recipients' status. Retrieved September 23, 2004, from http://aspe.hhs.gov/hsp/leavers99 /staterpts/GAO-Welf-Reform-Info-Former-he99048.pdf

Warr, P. B. (1992). Age and occupational well being. Psychology and Aging, 7, 37-45.

JEFF SCOTT

University of Illinois at Urbana-Champaign

School of Social Work

Table 1: Reported job satisfaction distribution

Level N %

1. Very dissatisfied 32 7.3
2. Somewhat dissatisfied 34 7.8
3. Neutral 22 5.1
4. Somewhat satisfied 158 36.3
5. Very satisfied 189 43.6
 434 100.0

Table 2: Frequency distribution of individual
and employment characteristics

Individual Characteristics N %

Gender
 Male 12 3
 Female 422 97

Living Arrangements
 Married/living w/partner 69 16
 Single, never married 261 60
 Other 104 24

# of Children in Home
 0 24 6
 1 155 36
 2 141 32
 3 75 17
 4 or more 39 9

Race/Ethnicity
 African American 286 66
 Latino 37 9
 Caucasian 109 25

Education
 < high school degree 116 27
 High school degree/GED 133 31
 Postsecondary 181 42

Health Condition
 Very good 226 52
 Good 153 35
 Poor 39 9
 Very Poor 15 3

Residential Location
 Chicago 258 59
 Cook County suburbs 48 11
 Metro suburbs 26 6
 Downstate urban 64 15
 Rural 37 9

Work Values
 Economic 87 20
 Type/difficulty of work 109 25
 Social aspect 163 38
 Convenience of schedule 77 18
 Autonomy 37 9
 Physical environment 51 12

Welfare Characteristics

TANF Tenure
 < 2 years 123 28
 2 to 5 years 125 29
 > 5 years 186 43

Reason for TANF Exit
 Time limits 101 24
 Work requirements 119 27
 Enhanced income 211 49
 Diversion 225 52
 Hourly Wages

Job Characteristics N %

Hourly Wages
 [less than or equal to] $5.15 42 10
 $5.16 to $9.99 338 78
 [greater than or equal to] $10.00 54 12

Hours Worked Per Week
 [less than or equal to] 20 44 10
 21 to 30 73 17
 31 to 40 272 63
 > 40 45 10

Job Tenure
 [less than or equal to] 6 months 235 54
 6 months to a year 119 27
 a year or more 75 17

Minutes to Work
 [less than or equal to] 30 285 66
 31 to 60 46 11
 > 60 103 24

Job Type
 Professional/ technical 11 3
 Managerial 24 6
 Clerical 71 16
 Service 145 33
 Sales/cashier 58 13
 Industrial 39 9
 Craft /construction 9 2
 Other 20 5
 Missing 57 13

Employer-Sponsored Health Plan
 Yes 97 22
 No 337 78

Number of Jobs Held at a Time
 1 415 96
 > 1 17 4

On-the-Job Training
 Yes 65 15
 No 367 85

Note. Respondents were allowed to
report more than one reason for TANF
exit and more than one work value.
Categories associated with these characteristics
are not mutually exclusive.
For a few characteristics, there were
small amounts of missing data. Where
category Ns do not sum to 434 (Health
Condition and Job Tenure), percentages
likewise do not sum to 100.

Table 3: Mean job satisfaction and bivariate associations of job
satisfaction with individual and employment characteristics

 Mean Job % Somewhat or
 Satisfaction Very Satisfied

Overall 4.01 79.9

Individual Characteristics

Gender
 Male 3.85 66.7
 Female 4.01 80.1

Living Arrangements
 Married/living w/partner 4.06 79.1
 Single, never married 4.03 77.9
 Other 3.95 80.9

# of Children in Home
 0 4.25 83.4
 1 3.87 74.1
 2 4.06 82.2
 3 4.11 86.5
 4 or more 4.07 77.5

Race/Ethnicity
 African American 4.06 81.5
 Latino 4.06 80.6
 Caucasian 3.87 76.8

Education
 < high school degree 3.93 76.9
 High school degree/GED 4.06 82.6
 Postsecondary 4.01 79.7

Health Condition
 Very good 4.21 *** 84.0 *
 Good 3.93 79.8
 Poor 3.30 60.0
 Very Poor 3.64 66.6

Residential Location
 Chicago 4.09 81.9
 Cook County suburbs 3.69 70.0
 Metro suburbs 4.26 84.6
 Downstate urban 3.90 78.2
 Rural 3.95 76.3

Work Values
 Economic 4.12 83.9
 Type/difficulty of work 4.19 84.4 *
 Social aspect 4.04 82.1
 Convenience of schedule 4.00 81.5
 Autonomy 4.07 70.3
 Physical environment 4.12 82.7

Welfare Characteristics

TANF Tenure
 < 2 years 3.86 * 69.9 *
 2 to 5 years 4.07 82.6
 > 5 years 4.15 84.3
Reason for TANF Exit
 Time limits 4.16 89.1
 Work requirements 3.93 78.2
 Enhanced income 4.00 80.1
 Diversion 4.11 84.5 *

Job Characteristics

 Hourly Wages
 [less than or equal] to $5.15 3.28 *** 63.5 **
 $5.16 to $9.99 4.07 81.2
 [greater than or equal to] 10.00 4.23 83.7

Hours Worked Per Week
 [less than or equal] 20 3.87 ** 77.3 **
 21 to 30 3.62 67.6
 31 to 40 4.11 82.8
 > 40 4.19 82.6

Job Tenure
 [less than or equal] 6 months 4.06 83.0
 6 months to a year 3.86 74.8
 a year or more 4.05 78.3
 Minutes to Work
 [less than or equal] 30 4.00 78.6
 31 to 60 4.21 89.1
 > 60 3.96 77.9

Job Type
 Professional/ technical 4.80 91.6
 Managerial 4.17 80.0
 Clerical 4.22 87.3
 Service 3.83 75.2
 Sales/cashier 3.94 76.3
 Industrial 3.94 79.5
 Craft /construction 4.54 90.0
 Other 4.34 95.0
 Missing 3.92 75.4

Employer-Sponsored Health Plan
 Yes 4.24 * 84.6 **
 No 3.95 78.6

Number of Jobs Held at a Time
 1 4.02 80.5
 > 1 3.86 70.6

On-the-Job Training
 Yes 4.23 84.8
 No 3.97 79.0

* p < .05. ** p < .01. *** p < .001.

Note: For work values and reason for
TANF exit, a series of separate analyses
were conducted for each of the response
categories as they are not mutually
exclusive.

Table 4: Ordered Logistic Regression of Job Satisfaction
on individual and employment characteristics

 B (SE B)

Individual Characteristics

Gender
 (Female)
 Male .254 (.638)

Living Arrangements
 (Other)
 Married/living w / partner .023 (.229)
 Single, never married .059 (.185)

Number of Children in Home .072 (.092)

Race/Ethnicity
 (Caucasian)
 African American -.137 (.290)
 Latino .052 (.425)

Education
 (Postsecondary)
 < high school degree .159 (.266)
 High school degree/GED .212 (.243)

Health Condition
 (Very good)
 Good -.583 ** (.221)
 Poor -1.444 *** (.357)
 Very Poor -.935 (.555)

Residential Location
 (Rural)
 Chicago .000 (.413)
 Cook County suburbs -.750 (.477)
 Metro suburbs .554 (.545)
 Downstate urban -.165 (.418)

Work Values
 (Social aspect)
 Economic .127 (.272)
 Type/difficulty of work .487 (.253)
 Convenience of schedule .449 (.270)
 Autonomy .605 (.374)
 Physical environment .051 (.314)
 Other .456 (.358)

Welfare Characteristics

TANF Tenure .053 * (.022)

Reason for TANF Exit
 (Time limits)
 Work requirements -.262 (.222)
 Enhanced income -.330 (.205)
 Diversion .162 (.207)

Job Characteristics

Hourly Wages .141 ** (.050)

Hours Worked Per Week .022 * (.011)

Job Tenure (months) -.004 (.010)

Minutes to Work -.001 (.004)

Job Type
 (Service)
 Professional / technical 2.553 ** (.951)
 Managerial .019 (.454)
 Clerical .377 (.314)
 Sales / cashier -.069 (.324)
 Industrial -.424 (.374)
 Craft / construction 1.445 (.863)
 Other .504 (.491)
 Missing .157 (.322)

Employer-Sponsored Health .280 (.270)

Holding > 1 job at a Time -.233 (.499)

On-the-Job Training .309 (.282)

Model Information
N 426.68
Chi Square (model fit) 91.480 ***
Nagelkerke .209

* p < .05. ** p < .01. *** p < .001.