Methods: Data and samples: We draw on data from three waves (2013-2018) of the China Health and Retirement Longitudinal Study (CHARLS), a nationally representative longitudinal survey of mid-aged and older Chinese that began in 2011. Given that the respondents’ relative deprivation information was collected from 2013, we extracted data from the CHARLS 2013 (Wave 2) to 2018 (Wave 4) and limited the respondents who attended at least two waves from 2013. The final sample size was 6,035.
Measures: In this study, relative deprivation was measured by relative socio-economic status benchmarked by five references groups (i.e., relative socio-economic status compared with relatives, schoolmates, colleagues, neighbors, and other people in the city). The answers range from 1 to 3, with a higher score indicating a higher relative socio-economic status. Health was estimated by depressive symptoms and self-rated health. Depressive symptoms were evaluated with the ten questions version of the ten-item Center for Epidemiologic Studies Depressive Scale (CES-D). Self-rated health was assessed by self-reported general health.
Analysis: First, to explore relative deprivation patterns, we used latent class analysis (LCA) to identify unobserved subgroups based on relative socio-economic status among five reference groups for the full sample and subsamples (i.e., male, female, urban, and rural groups) separately. Second, parallel process latent growth curve models (PPM) were adopted to examine how these identified latent groups were associated with depressive symptoms and self-rated health trajectories in later life for the full sample and subsamples.
Results: The results of LCA identified two latent groups (the deprived and non-deprived groups). Similar patterns were found from the full sample and subgroups. The PPM results from the full sample indicated that the deprived group, compared to the non-deprived group, showed a higher baseline level of depressive symptoms and a lower baseline level of self-rated health. However, compared with the non-deprived group, the deprived group had a flatter rate of growth of depressive symptoms in later life. Similar results were found in females and rural residents.
Conclusions and Implications: Relative deprivation has long-term impacts on health. Besides, non-deprived females and rural residents are at high risk of mental health deterioration in the long term. These results suggest that reducing relative deprivation could facilitate health, and policies aimed at promoting mental health should not only take the deprived groups into account but also consider the non-deprived females and rural residents.