Study Examines Labor Conditions and Depressed Affect Among Chinese Workers

In a repeated cross-sectional analysis of Chinese workers aged 18 to 59 years, researchers observed positive associations between long weekly working hours and recent feelings of depressed or down affect. The investigation, which harmonized nine waves of the Chinese General Social Survey conducted between 2010 and 2023, evaluated how distinct labor conditions relate to worker mental health. Analyzing adjusted percentage-point differences through weighted linear probability models, the authors identified an adjusted prevalence difference of 2.25 percentage points, with a 95 percent confidence interval ranging from 0.95 to 3.55, for employees working more than 44 hours per week compared to baseline reference levels. For individuals working 60 hours or more per week, the adjusted prevalence difference was 2.82 percentage points, with a 95 percent confidence interval spanning 1.12 to 4.51 percentage points.

The study evaluated whether working hours demonstrated consistent patterns under different analytical conditions. While both weekly working-hour thresholds showed positive associations in pooled models, sensitivity checks revealed variations across survey periods. Specifically, when data from survey years 2011 and 2023 were excluded simultaneously from the models, the adjusted estimate for working more than 44 hours per week attenuated to 1.40 percentage points, yielding a 95 percent confidence interval from -0.06 to 2.85 percentage points that crossed the null value. In contrast, the estimate for workers reporting 60 hours or more per week remained positive at 2.48 percentage points, with a 95 percent confidence interval from 0.45 to 4.51 percentage points. The investigators highlighted that these overlapping working-hour thresholds were not designed to test a dose-response relationship.

Beyond working hours, the research team examined employment-relation forms by evaluating casual arrangements, positions without a fixed employer, and labor-dispatch work. When analyzed as a combined employment-relation category compared against fixed-employer employees, the adjusted prevalence difference for the primary outcome was 3.31 percentage points, accompanied by a 95 percent confidence interval between 0.71 and 5.92 percentage points. However, this association proved sensitive to how the mental health outcome was defined. When evaluated under a stricter outcome threshold, the combined estimate attenuated to 0.88 percentage points, with a 95 percent confidence interval from -0.29 to 2.05 percentage points and a p-value of 0.135.

Disaggregating the employment-relation categories relative to employees with fixed employers further illuminated differing patterns. Casual employment or working without a fixed employer yielded a category-specific estimate of 4.87 percentage points. By comparison, labor-dispatch arrangements showed an adjusted estimate of 0.73 percentage points, and the direct contrast between casual work and labor dispatch remained imprecise. Because of this imprecision and outcome sensitivity, the researchers characterized the specific estimate for labor dispatch as inconclusive, emphasizing that combining diverse non-standard employment types into a single broad label can mask meaningful differences between work arrangements.

The investigation also evaluated contract formalization and workplace task autonomy, completing the assessment of four separate labor dimensions. Models evaluating written-contract status generated an imprecise estimate, failing to establish a clear directional association. Meanwhile, the estimate for task autonomy was close to zero with a narrower confidence interval, indicating little detectable relationship with recent depressed affect within this analytical framework. Consequently, each of the four examined dimensions—working time, employment relations, contract formalization, and task autonomy—demonstrated distinct patterns of point estimates and sensitivity.

To conduct these assessments, the researchers harmonized nine survey waves from 2010 to 2023, relying on a single questionnaire item measuring whether respondents felt depressed or down at least sometimes during the previous four weeks. Analytical models incorporated survey-year indicators and province indicators, adjusting standard errors for clustering at the provincial level. The authors noted that because the study relies on repeated cross-sectional surveys rather than longitudinal panels, these contemporaneous associations cannot determine temporal order or establish causal effects. Nevertheless, the contrasting findings across domains indicate that future longitudinal studies should measure these four labor dimensions separately rather than grouping them under broad labor-condition categories.

Subject of Research: Labor conditions and depressive affect among Chinese workers

Article Title: Labor-condition indicators and recent depressed/down affect among Chinese workers, 2010–2023

Article References: Zhang, Y., & Wan, G. (2026). Labor-condition indicators and recent depressed/down affect among Chinese workers, 2010–2023. BMC Public Health. https://doi.org/10.1186/s12889-026-30175-5

Image Credits: AI Generated

DOI: 10.1186/s12889-026-30175-5

Keywords: labor conditions, mental health, depressive affect, Chinese General Social Survey, working hours, employment relations, occupational health, cross-sectional study, Labor-condition, indicators, recent, depressed