The research team based its model on the contemporary process perspective of ego depletion, which differs from the older idea that self-control operates like a finite tank of fuel that can be emptied. In this view, sustained regulatory demands can leave a person less motivated to continue restraining impulses, less attentive to cues that signal a need for control, and more responsive to immediate rewards or relief. Abusive supervision could create exactly these demands because employees may have to suppress negative emotions, inhibit retaliatory impulses, monitor interpersonal threats, and maintain appropriate conduct despite feeling mistreated. The proposed mechanism therefore concerns changes in motivation and attention, not simply exhaustion after making difficult choices.

The analysis separated the proposed process across three weekly waves. In week one, 700 valid participants evaluated their supervisor’s abusive behavior and supplied demographic information. In week two, 516 valid respondents who had completed the first survey reported pay satisfaction and ego depletion. In week three, those who had completed the second questionnaire reported compulsive buying and compulsive eating. Matching the records with unique user IDs and removing invalid responses left 445 complete cases. The one-week intervals were intended to place the predictor, proposed mechanism and outcomes in temporal order while limiting the passage of time during which new experiences could arise.

All five focal constructs were assessed with established questionnaires and seven-point response scales, and the researchers controlled for gender, age, education and monthly income. Abusive supervision was measured with Tepper’s 15-item scale, ego depletion with a five-item scale, compulsive buying with a 13-item scale, compulsive eating with a 15-item scale and pay satisfaction with a nine-item measure. Their reported internal consistency was high, with Cronbach’s alpha values ranging from 0.925 to 0.961 for those instruments. Confirmatory factor analysis favored the five-factor measurement model over alternatives that combined constructs. An unmeasured latent method factor also produced only modest improvements in model fit, suggesting that a single common method was unlikely to account fully for the observed relationships, although this check cannot remove all concerns associated with self-report.

In the direct statistical tests, abusive supervision had a positive association with ego depletion after adjustment for the controls, with a coefficient of 0.804 and a standard error of 0.066. Ego depletion, in turn, was positively related to compulsive buying, with a coefficient of 0.319 and a standard error of 0.048, and to compulsive eating, with a coefficient of 0.466 and a standard error of 0.038. The corresponding bivariate correlations were also positive: abusive supervision correlated with ego depletion at r = 0.504, with compulsive buying at r = 0.293 and with compulsive eating at r = 0.395. Ego depletion correlated with compulsive buying at r = 0.357 and with compulsive eating at r = 0.560. These estimates support the proposed sequence, but they describe statistical associations rather than experimental effects or changes caused by exposure to abusive supervision.

Bootstrap tests indicated that ego depletion carried part of the association between abusive supervision and each consumption outcome. The estimated indirect effect for compulsive buying was 0.183, with a 95% bias-corrected confidence interval from 0.082 to 0.302, while the indirect effect for compulsive eating was 0.310, with a 95% interval from 0.213 to 0.420. Neither interval included zero. In other words, employees who reported more abusive supervision also tended to report more self-regulatory impairment, which in turn was associated with more difficulty resisting buying and eating urges. The study did not establish whether the consumption behaviors temporarily reduced distress or accumulated into long-term harm, nor did it test whether one consumption behavior caused the other.

Pay satisfaction emerged as a selective moderator rather than a general answer to the effects of workplace mistreatment. It represents employees’ affective evaluation of their compensation, and the authors reasoned that satisfaction might signal greater perceived financial latitude, making an activated buying urge easier to enact. The interaction between pay satisfaction and ego depletion positively predicted compulsive buying, with a coefficient of 0.125 and p below 0.01. Simple-slope estimates showed that ego depletion was related to compulsive buying at both low and high pay satisfaction, but the relationship was stronger at high pay satisfaction: the estimated slope rose from 0.243 to 0.554, with both p values below 0.001. Because monthly income was included as a control, the result was not simply a reflection of higher-paid workers reporting more buying.

The same pattern did not emerge for compulsive eating. The interaction term was positive but did not reach the conventional statistical threshold, with a coefficient of 0.053 and p = 0.069. Simple slopes were positive at low pay satisfaction, at 0.344, and at high pay satisfaction, at 0.477, but their difference was not statistically supported. An integrated latent-variable structural equation model produced the same distinction. For compulsive buying, the conditional indirect effect of abusive supervision through ego depletion rose from 0.263 at low pay satisfaction to 0.430 at the mean and 0.596 at high pay satisfaction; the index of moderated mediation was 0.144, with a 95% confidence interval of 0.057 to 0.230. For compulsive eating, the corresponding index was 0.037 and its 95% interval, from -0.040 to 0.115, included zero. The authors interpret this contrast as evidence that financial latitude is more closely matched to the requirements of buying than to those of eating, which may be more proximally influenced by cravings, physiological cues, affect regulation and habitual responses.

The findings remain constrained by the study’s design and setting. Every focal variable was self-reported, and the researchers did not verify abusive supervision, spending or eating through supervisors, records or behavioral measures. A time-lagged, three-wave structure helps establish temporal ordering and reduce consistency motives, but it cannot rule out unmeasured confounding or demonstrate causality. Weekly intervals may also miss rapid within-person fluctuations, while the sample of full-time Chinese employees—59% women, 60% aged 26 to 35, and 84% holding a bachelor’s degree or higher—limits generalization to other workplaces and populations. The authors also note that they did not directly measure affordability, financial constraint, reward salience or access to consumption opportunities, so the explanation for pay satisfaction’s role remains inferential. Nevertheless, the results provide evidence that a workplace stressor can be associated with private-life behaviors through impaired self-control, and that the contextual conditions governing those behaviors need not be identical.

Subject of Research: A three-wave observational survey of abusive supervision, ego depletion, pay satisfaction and compulsive consumption among full-time employees in China.

Article Title: Beyond the workplace: abusive supervision and employees’ compulsive consumption behaviors

Article References: Wang, M., Zhang, Y., Lang, Y., & Liu, Y. (2026). Beyond the workplace: abusive supervision and employees’ compulsive consumption behaviors. Frontiers in Psychology, 17, 1885985. https://doi.org/10.3389/fpsyg.2026.1885985

Image Credits: AI Generated

DOI: 10.3389/fpsyg.2026.1885985

Keywords: abusive supervision, compulsive buying, compulsive eating, ego depletion, self-regulation, pay satisfaction, workplace stress, employee behavior, consumption, survey research, Beyond, workplace