What Makes a Job Better? Job-Changers Pull Back the Curtain and Share Survey Evidence
When workers change jobs, they may find themselves faced with numerous questions: How is the healthcare coverage? Will I be able to pick up my kids from school? How long before a promotion? Because workers value characteristics of jobs beyond their paychecks, differences in job quality cannot be measured by wage differentials alone.1 However, limitations in available data have led to a focus on differences in determinants of wages and salaries that obscures the role of other job characteristics in determining job quality.
In this article, we use survey data from the Federal Reserve’s Survey of Household Economics and Decisionmaking (SHED) to analyze how changes in monetary and nonmonetary job characteristics relate to workers’ assessment of their new job’s overall quality. Specifically, we use new survey questions asked of individuals who changed jobs to see (1) whether the respondent said that the overall job was better, the same, or worse than the respondent’s previous job and (2) how the change in self-assessed job quality relates to other self-assessed changes in pay and benefits, opportunities for advancement, the respondent’s interest in the work, the physical demands of the job, and the work–life balance associated with the job.
We test the relative importance of job characteristics by assessing how well changes in a single, self-assessed job characteristic predict changes in self-assessed overall job quality. First, we show how accurately changes in each characteristic summarize changes in overall job quality. Predicting overall job quality using only that characteristic gives an assessment of how accurate it would be to make assessments using only that characteristic, as is commonly done using measures of workers’ pay.2 Second, we estimate a linear probability model to show how much more likely a respondent is to say that their new job is better conditional on a change in a particular characteristic, holding the others fixed. Finally, we examine the correlations in changes between job characteristics.
We find strong evidence that workers value characteristics beyond pay and benefits in evaluating their new jobs, although pay and benefits remain important predictors of job quality. Predicting whether a worker considered a job was better or not using only changes in pay and benefits gives an incorrect inference 30 percent of the time. Interest in the work is particularly predictive of a worker’s assessment of overall job quality, even after controlling for changes in pay and benefits. Holding other characteristics fixed, an improvement in a worker’s interest in the work is associated with a 27-percent increase in the likelihood of the employee considering the job to be better overall. For pay and benefits, the increase is smaller, at 19 percent. Work–life balance is also highly predictive, with improved work–life balance being associated with an 18-percent increase in the likelihood of an employee considering the job to be better.
Across demographic groups, the changes in characteristics that most affect overall assessments of job quality are similar, but we do find evidence that characteristics vary in predicting overall job quality for different subgroups. Changes in work–life balance are around 50 percent more predictive of overall changes in job quality for women than for men. However, our results are too imprecise to detect if changes in work–life balance are meaningfully more predictive of overall changes in job quality for parents. Also, the results show that improvements in pay and benefits are more associated with overall self-assessed job quality improvements for job changers over 45 years of age. Finally, for older job changers, changes in interest in the work are particularly predictive of overall self-reported job quality compared with younger workers.
The correlations between changes in job characteristics are stronger for workers with lower levels of education compared with workers with a bachelor’s degree or higher. Improvements in pay and benefits stand out with a robust, positive correlation with improvements in other characteristics for workers with a high school education or less. The weaker levels of correlation in job characteristics for more educated workers suggest that improvements in pay and benefits tend to accompany improvements in work–life balance, opportunities for advancement, and interest in the work to a greater degree among workers with a high school education or less relative to workers with a bachelor’s degree or higher.3 The correlation results suggest that pay and benefits changes for less educated job changers understate the overall change in self-assessed job quality compared with job changes for workers with a bachelor’s degree or higher.
The survey questions we use were added to the SHED in 2021, and this article presents results through 2023, a period when labor markets were relatively tight and inflation was higher than in previous years. Analyzing responses since 2021 illustrates the importance of other job characteristics in determining changes in self-reported job quality, when workers had more bargaining power to determine working conditions, pay, and benefits. Employers may also have been making tradeoffs between pay, benefits, and other job characteristics to attract workers, while maintaining profitability, while other input prices were increasing.
There are several advantages to surveying job changers. The first is being able to ask simple and understandable questions. Individuals who change jobs may be able to more easily compare their overall job satisfaction and how specific characteristics in their new job compare to those in their old one. In a survey, we can ask respondents about each characteristic directly, as opposed to trying to infer changes indirectly or asking respondents to assess their job’s quality relative to alternative jobs, the knowledge of which is likely to depend on their personal experiences. The second is that we examine workers’ preferences about jobs that they actually moved to, which allows us to go beyond hypothetical comparisons. Third, job changers are a particularly relevant group, as they are the individuals actively comparing jobs. Hence, they are most likely to have their decisions affected by any job-quality shifts. Lastly, by focusing on job changers, we are also able to analyze actual changes in particular characteristics for the same worker.
In contrast to work surveying all workers or asking about hypothetical job changes, SHED data represents individuals who change jobs, and these individuals may differ from workers more generally in both observable and unobservable ways. In addition, by focusing on job changes, we identify differences in characteristics between jobs among job changers, rather than levels of satisfaction with characteristics. Finally, changes in job characteristics are being evaluated after a job change and are subjective. These assessments could be overly positive if they are affected by confirmation bias, where the worker is more likely to say the job is better to justify their decision to switch, or overly negative if they are focused on negative characteristics of their current job, which are more salient than the downsides of their previous position.
The most direct antecedents in the literature are studies asking workers about job characteristics, both actual and hypothetical. Previous research has used surveys of workers to collect information about job characteristics and workers’ willingness to pay for job characteristics based on hypothetical job choices.4 Similarly, other work has estimated job quality based on workers’ evaluations of job characteristics combined with a directly elicited rating of the importance of the characteristic to the worker.5 In contrast, we draw conclusions about individuals who are on the margin between two different jobs, since we observe them changing jobs. This approach blends the existing survey literature with studies of job search behavior.6
This work also provides additional evidence contributing to the broader literature about the importance of nonmonetary job characteristics. This literature suggests that a substantial proportion of job quality is due to characteristics other than pay.7 There is also a strong public policy interest in the concept of measuring workers’ job quality as this knowledge can inform the creation of policies that promote “good jobs.”8
Data description and results
The data discussed in this article come from the 2021, 2022, and 2023 versions of SHED, an annual survey of more than 11,000 adults.9 The survey contains extensive information about respondents’ financial well-being, as well as their employment situations and sources of income. The survey is nationally representative and based on a primarily address-based probability sample, where households that do not initially have internet access are provided with internet access to complete the survey.10 All analyses use weights designed to match the demographic characteristics of the national population according to the Current Population Survey.11
The specific questions that we focus on were new in 2021 and asked of job changers. Job changers are defined as individuals who were employed at the time of the survey, reported starting a new job in the last year, and whose current main job is different from their main job a year before. Job changers were presented with a series of characteristics alongside the question, “Are each of the following better, the same, or worse at the main job you have now than the one you had a year ago?” The characteristics are the following: pay or benefits, opportunities for advancement, your interest in the work, physical demands of the job, and work–life balance. Respondents were then asked, “Overall, is the main job you have now better, the same, or worse than the one you had a year ago?”
A notable difference between all workers in the sample and those who change jobs was that job changers were much more likely to be young. We also saw small differences by education and income, which interestingly showed opposite effects. Higher-educated individuals were more likely to change jobs, while individuals with lower family income were also more likely to change jobs. There were not many differences between the years in terms of who changed jobs, but 2022 had the smallest share of job changers laid off, perhaps reflecting the strong job market that year.13 Whether or not an individual was laid off from a job may have driven differences in their job search and subsequent job quality. In the Differences between voluntary and involuntary job changes section, we provide a separate analysis comparing job changers who experienced a layoff to those who did not.
How well do job characteristics predict overall job quality?
To understand which job characteristics are most predictive of a worker saying that their new job was better overall, we compute error rates of each characteristic. These error rates are fractions of observations where (1) the characteristic is better, but the job is the same or worse or (2) the characteristic is the same or worse and the job is better overall.16 (See table 3.) We also separately calculate the share of job changes that have a better job characteristic but the same or worse overall self-assessed job quality and the share of job changes that have the same or worse job characteristic but better self-assessed overall job quality. The error rate is the smallest for interest in the work, followed by pay and benefits, advancement opportunities, and work–life balance. Changes in physical demands appear less predictive of whether a job is better.17 The error rates associated with the job characteristic being worse or the same but the job quality being better are higher than those where the job characteristic is better but the self-assessed job quality is worse.18
Interest in the work stands out as very predictive in determining overall self-reported job quality, with 90 percent of job changes where interest in the work improved being associated with a job improvement overall and only 21 percent of job changes where interest in the work declined being associated with the job being better overall. The other job characteristics have sizeable shares of job changers who experienced a decline in that job characteristic but still felt the job was better overall. For example, over 40 percent of jobs where work–life balance, physical demands, or pay and benefits were worse were still seen as overall better jobs.
Correlations between job characteristics
Overall, changes in job characteristics are modestly positively correlated.19 Table 4 shows a correlation matrix of improvements in each of the characteristics among job changers. Again, a greater interest in the work was most correlated with the job being better overall, followed by pay and benefits, advancement opportunities, and work–life balance. Improvements in pay and benefits are positively correlated with improvements in all other characteristics. The pay and benefit changes associated with a job change are most strongly correlated with opportunities for advancement and are less correlated with interest in the work, physical demands, and work–life balance. Analyzed together, the positive correlations are consistent with differences in overall quality across jobs, as opposed to an equilibrium where jobs offer compensating differentials for lower quality characteristics.20
Grouping the data by education shows weaker correlations between job characteristics for more educated workers. (See table 5.) For workers with a bachelor’s degree or higher, unlike workers overall, work–life balance and physical demands have very low correlations with pay and benefits. In contrast, the correlation levels between job characteristic changes are larger for those with a high school education or less, although the general patterns are similar. Increases in self-assessed job quality are most correlated with interest in the job, followed by pay, work–life balance, and opportunities for advancement. Changes in pay and benefits are most correlated with opportunities for advancement but are also relatively correlated with interest in the work. However, one difference is that, for those with a high school degree or less, changes in work–life balance are more correlated with changes in pay, interest in the work, opportunities for advancement, and overall job quality than for those with a bachelor’s degree or higher.
The correlations suggest that the variation in job characteristics are more likely to reflect overall job quality for workers with a high school degree or less compared to workers with a bachelor’s degree or higher. If one characteristic, like pay and benefits, improves, then it is more likely that other characteristics will improve as well. One implication is that differences in pay and benefits understate differences in overall job quality for less educated workers. Finally, it is also important to note that previous research suggests that workers without a bachelor’s degree generally have lower levels of pay, less workplace flexibility, and higher physical demands in their jobs.21
Which job characteristics do workers value when changing jobs?
Improvements in pay and benefits as well as work–life balance are also very predictive of a worker saying that a job is better, holding other characteristics constant. Better pay and benefits makes an individual 19 percentage points more likely to say the job is better, and better work–life balance makes an individual 18 percentage points more likely to say the job is better. (See table 6.) Better work opportunities are associated with a 9-percentage-point increase in the likelihood of the job being better, but changes in physical demands have a small and statistically undetectable effect.22
Next, we show evidence that parents particularly value changes in work–life balance by separately estimating the model on samples based on the presence of children in the household. Parents of a child under 13 have a 23-percentage-point greater probability of saying a job is better, conditional on work–life balance being better. (See table 6.) The point estimate is smaller for nonparents, with only a 17-percentage-point increase in the likelihood of saying the job is better based on work–life balance being better, although the standard errors of these estimates limit the strength of these conclusions.
The estimates also suggest that mothers value changes in work–life balance more than fathers. In table 6, mothers of children under 13 have a higher coefficient of 0.29 compared with 0.19 for fathers.
Women also have a larger estimated coefficient on work–life balance than men. Previous research has documented gender differences in value placed on certain job characteristics, finding that women with young children were willing to give up more in salary to avoid irregular schedules and to work from home than women without young children or men with children.23 Others have found that women have a higher willingness to give up more in salary in exchange for paid time off and lower physical demands.24
Conclusion
We use a large, nationally representative survey (SHED) to show that job changes lead to improvements in job characteristics beyond pay and benefits. Focusing on pay and benefits alone leads to incorrect predictions about job changes 30 percent of the time, as workers use more than pay and benefits to evaluate job quality. The most predictive characteristic of self-reported job quality is interest in the work. The results of this article suggest that parents, in particular mothers, weigh work–life balance more heavily in assessing whether their new job is better than do nonparents. Pay and benefits changes are less predictive of overall self-reported job quality changes for older workers compared with younger workers, with interest in the work being more predictive for older workers.
We also show large differences in the correlations between different job characteristics by worker education level. Changes in job characteristics are more positively correlated for workers with a high school degree or less compared with those with a bachelor’s degree or higher. The differences in correlations suggest that using only one characteristic, like pay and benefits, to assess job quality understates differences in overall job quality by more for less educated workers, relative to those with a bachelor’s degree or higher.
The results improve understanding of differences in workers’ pay, which are imperfect proxies for job quality. Workers’ emphasis on factors like their interest in the work, holding pay fixed, helps to explain the large amount of effort that businesses devote to their culture beyond strict productivity enhancements. The results in this article also validate previous research focusing on moving workers without bachelor’s degrees into higher-paying jobs, since those jobs tend to also come with other, better characteristics.
Katherine Lim, and Mike Zabek, "What makes a job better? Survey evidence from job changers," Monthly Labor Review, U.S. Bureau of Labor Statistics, June 2026, https://doi.org/10.21916/mlr.2026.16
References
1 For a recent survey of the compensating wage differentials literature, see Kurt Lavetti, “Compensating wage differentials in labor markets: empirical challenges and applications.” Journal of Economic Perspectives 37, no. 3 (Summer 2023), pp. 189–212, https://doi.org/10.1257/jep.37.3.189.
2 The academic literature following John M. Abowd, Francis Kramarz, and David N. Margolis, “High wage workers and high wage firms,”Econometrica 67, no. 2 (March 1999), pp. 251–333, https://www.jstor.org/stable/2999586 explicitly focuses on job changers’ earnings differences between firms without considering differences in other characteristics, primarily due to data availability. Policy programs are also often focused on creating pathways from low- to high-paying jobs, including Harry J. Holzer, Julia I. Lane, David B. Rosenblum and Fredrik Andersson, Where are all the good jobs going? What national and local job quality and dynamics mean for U.S. workers (Russell Sage Foundation, 2011) and Keith Wardrip, Kyle Fee, Lisa Nelson, and Stuart Andreason, Identifying opportunity occupations in the nation’s largest metropolitan economies (U.S. Federal Reserve System, September 9, 2015), https://www.clevelandfed.org/publications/cd-reports/2015/sr-20150909-identifying-opportunity-occupations.
3 Previous research by Jonathan Rothwell and Steve Crabtree, Not just a job: new evidence on the quality of work in the United States (Gallup, 2019), https://www.luminafoundation.org/wp-content/uploads/2019/11/not-just-a-job-new-evidence-on-the-quality-of-work-in-the-united-states.pdf, on the distribution of job characteristics by income suggests that high-income workers have better job characteristics overall, so perhaps their job changes are less likely to result in improvements across all characteristics, even if they have better job characteristics.
4 Nicole Maestas, Kathleen Mullen, David Powell, Till von Wachter, and Jeffrey Wenger, “The value of working conditions in the United States and the implications for the structure of wages,” American Economic Review 113, no.7 (July 2023), pp. 2007–47, https://doi.org/10.1257/aer.20190846.
5 Rothwell and Crabtree, Not just a job: New evidence on the quality of work in the United States.
6 See Robert Hall and Andreas Mueller, “Wage dispersion and search behavior: the importance of nonwage job values,” Journal of Political Economy, 126, no. 4 (August 2018), pp. 1594–1637, https://doi.org/10.1086/697739.
7 See Isaac Sorkin, “Ranking firms using revealed preference,” The Quarterly Journal of Economics 133, no.3, (January 2018), pp. 1331–1393, https://doi.org/10.1093/qje/qjy001.
8 See Batia Katz, William Congdon, and Jessica Shakesprere, Measuring job quality current measures, gaps, and new approaches (The Urban Institute, April 2022) and William Congdon, Molly Scott, Batia Katz, Pamela Loprest, Demetra Smith Nightingale, and Jessica Shakesprere, Understanding good jobs: a review of definitions and evidence (The Urban Institute, July 2020), https://www.urban.org/research/publication/understanding-good-jobs-review-definitions-and-evidence.
9 Report on the economic well-being of U.S. households in 2021 (U.S. Board of Governors of the Federal Reserve System, May 2022), https://www.federalreserve.gov/publications/files/2021-report-economic-well-being-us-households-202205.pdf.
10 Prior to 2009, some members were recruited via random digit dialing. For more information, see Report on the economic well-being of U.S. households in 2021 (U.S. Board of Governors of the Federal Reserve System, May 2022).
11 Jeff Larrimore, Maximilian Schmeiser, and Sebastian Devlin-Foltz, Should you trust things you hear online? Comparing SHED and Census Bureau survey results (U.S. Board of Governors of the Federal Reserve System, October 2015), https://doi.org/10.17016/2380-7172.1619 show that weighted results from a previous vintage of the survey based on the same online panel are comparable to several other commonly used surveys produced by the U.S. Census Bureau.
12 See Rick Penn and Victor Huang, “Job openings reach record highs in 2022 as the labor market recovery continues,” Monthly Labor Review, May 2023, https://doi.org/10.21916/mlr.2023.10 for evidence on strong labor market data in 2022.
13 See Victor Huang and En Ping Cheng, “Job openings and hires decline in 2023 as the labor market cools,” Monthly Labor Review, September 2024, https://doi.org/10.21916/mlr.2024.18 for evidence on strong labor market data in 2022 followed by weaker data in 2023.
14 When we examine differences by whether the job changer was laid off, we find that 53 percent of job changers who were laid off had a better job overall, compared with 73 percent of job changers who were not laid off (see chart 2).
15 U.S. Bureau of Labor Statistics, “Nonfarm business sector: Unit labor costs for all workers [prs85006112]” (Federal Reserve Bank of St. Louis (FRED database), retrieved on December 4, 2023), https://fred.stlouisfed.org/series/PRS85006112.
16 One can think of the error rate in this context as instances where the change in the quality of the specific job characteristic does not match the change in the quality of the overall job.
17 Differences between the error rates are all statistically significant at the 5-percent level, and all but the difference between opportunities and work–life balance are significant at the 1-percent level.
18 When we separate the error rate analysis by laid-off status, we find similar qualitative patterns across the characteristics (table 8). Error rates are generally lowest for interest in the work. The better error rate is higher for those laid off than for those not laid off, while the worse error rate is lower. This is consistent with those who are laid off having a lower share of overall job improvements.
19 Although we find modest positive correlations across pay and other job amenities, a recent paper finds evidence of a negative relationship between pay and common amenities among job changers. See Anders Humlum, Mette Rasmussen and Evan K. Rose, “Firm premia and match effects in pay vs. amenities,” working paper 33884 (National Bureau of Economic Research, May 2025), http://doi.org/10.3386/w33884.
20 Kurt Lavetti, “Compensating wage differentials in labor markets: empirical challenges and applications.”
21 It may be less costly to improve characteristics when they begin at low levels. This would occur if the cost function for job characteristics is convex. Nicole Maestas, Kathleen Mullen, David Powell, Till von Wachter, and Jeffrey Wenger, Working conditions in the United States: results of the 2015 American Working Conditions Survey (RAND, 2017), https://www.rand.org/pubs/research_reports/RR2014.html. We find that changes in job characteristics are also more correlated for job changers who were laid off, compared with those who were not (table 9). This may partially reflect differences in layoff propensities by education and, likewise, the education results may reflect differences in the share of voluntary job changes.
22 When we separate our analysis by whether the person had been laid off, we find that results are mostly similar except that pay and benefits being better is more strongly associated with a better job overall for those who were laid off, relative to those who were not, and the coefficient is similar to the effect of interest in the work (table 10).
23 Alexandre Mas and Amanda Pallais, “Valuing alternative work arrangements,” American Economic Review 107, no. 12 (December 2017), pp. 3722–59, https://doi.org/10.1257/aer.20161500.
24 Nicole Maestas, Kathleen Mullen, David Powell, Till von Wachter, and Jeffrey Wenger, “The value of working conditions in the United States and the implications for the structure of wages.”
25 Nicole Maestas, Kathleen Mullen, David Powell, Till von Wachter, and Jeffrey Wenger, Working conditions in the United States: results of the 2015 American Working Conditions Survey.
26 The coefficient is also consistent with perceptions that millennial and Generation Z workers prioritize work–life balance more than older generations.
27 However, the two coefficients themselves are statistically indistinguishable.
28 Rothwell and Crabtree, Not just a job: new evidence on the quality of work in the United States.
29 Since the question about layoffs does not specify what job someone was leaving, some individuals in the group reporting a layoff may have actually left their previous main job voluntarily and been laid off from a different job. This job could have been a secondary job or another main job that they held for less than a year.
30 The sample used in this article is too small to study differences jointly by layoff status and other covariates, like education.
