This study uses data from the China Health and Retirement Longitudinal Survey (CHARLS), a large-scale interdisciplinary tracking project carried out jointly by the China Social Science Survey Center at Peking University and the Peking University Youth League Committee. As part of CHARLS, a national baseline survey was conducted in 2011–2012, followed by three national follow-up surveys in 2013, 2015, 2018 and 2020, and a life course survey in 2014. It was conducted using the Probability Proportionate to Size (PPS) sampling method on 150 counties and 450 communities in 28 provinces, making the sample nationally representative. Using the particular survey data from the “China Middle-aged and Elderly Life Course Survey” in 2014 and 2018, this study horizontally merges the two databases by matching the elderly identity code variables. Childhood adversity came from the 2014 Chinese Middle-aged and Elderly Life Course Survey, while depression, social participation and control variables came from the 2018 China Health and Elderly Care Tracking Survey. In this study, childhood adversity is a key variable, so we excluded samples that did not participate in the 2014 special survey and those aged 45–59, resulting in a reduction of 11,751 samples. Subsequently, we excluded 1,297 samples with missing or abnormal values for the three main variables: childhood adversity, depression, and social participation. Ultimately, we obtained a final sample of 6,704 older adults aged 60 and above.

The data were analyzed using STATA15.1 software, and single-factor analysis was used to compare the depression degrees of older adults based on their essential characteristics. An analysis of childhood adversity and depression in older adults was conducted using multiple linear regression models. The robustness of the relationship between childhood adversity and depression was tested in order to alleviate endogenous problems. To analyze the “dose–response” effects of 7 childhood adversity items, the “number of childhood adversity experiences” was transformed into a dummy variable and analyzed with a multiple linear regression model. Furthermore, the moderating effect was examined by combining the product term of social participation with the seven items of childhood adversity.

A total of 6,704 older adults were surveyed in this study, of which 3,325 (49.60%) were males and 3,379 (50.4%) were females; their ages ranged from 60 to 118, with an average age of 70.24 years old; 5,222 (77.89%) belonged to agricultural households and 1,482 (22.11%) were from urban households; 2,053 (30.62%) were illiterate and 2,993 (44.64%) in primary school, 1,658 (24.73%) in junior high school and above; 2,355 (35.13%) not in marriage (including widowed, divorced and unmarried), 4,349 (64.87%) in marriage; 6,187 (92.29%) working in agriculture as type of work before retirement; 121(1.8%) working in government departments/institutions/non-profit organizations, 396 working in enterprises/individuals (5.91%); 2,051 (30.59%) drinking alcohol, 4,653 (69.41%) not drinking alcohol; 3,466 (51.70%) without chronic diseases, 1,925 (28.71%) with one chronic disease, 1,313 (19.59%) with two or more chronic diseases; 1,908 with annual income <6.84 ~ 9.49, and 1,668 with annual income >9.49. 2,217 (33.07%) had no childhood adversity, 1,977 (29.49%) had one childhood adversity, 1,343 (20.03%) had two childhood adversity, and 1,167 (17.41%) had three or more childhood adversity.

Based on univariate analysis of Table 1, there are significant differences in depression degree among older adults based on their gender, age, education level, marital status, type of household, income, type of work before retirement, alcohol consumption, chronic disease, and childhood adversity experience (p < 0.05). Female older adults were more likely to suffer from depression than male older adults; The young old had a higher depression degree than the oldest old. Those who are illiterate have the highest degree of depression, while those who have a high school education or more have the lowest. The unmarried were more depressed than the married. There is a higher depression degree among older adults in agricultural households than among older adults in urban households. There was a greater prevalence of depression among those who worked in agriculture before retirement. Those who did not drink alcohol were more depressed than those who drank alcohol. Depression rates were highest among older adults with two or more chronic diseases and lowest among older adults without chronic diseases. The depression level of older adults who had faced “the caregiver with severe physical disabilities” was the highest, and the depression level of older adults who had faced a “low-income family economic situation” was the lowest. Table 1 shows the results.

Multiple linear regression was conducted using depression score as the dependent variable and seven childhood adversities as the independent variables. Model 1 showed that after considering childhood adversities, lack of mutual assistance within the neighborhood, lonely due to lack of friends, two weeks or more of depression in the caregiver, the caregiver with severe physical disabilities, and having a poor relationship with parents, older adults were more likely to be depressed (p < 0.01). Adding moderating and controlling variables to model 1, the results show that childhood adversity’s coefficients have all declined but still have strong explanatory power for depression among older adults. Among older adults, social participation, age, education, Logarithm of the annual income of older adults, alcohol consumption, urban household registration, marriage, and work type before retirement are non-agricultural (working for government departments, institutions, non-profit organizations, or individuals) decreased depression risk (p < 0.05). The higher number of chronic diseases would enhance the tendency to depression in older adults (p < 0.05). Table 2 shows the results.

According to the model above, the effect of childhood adversity on depression in older adults may be endogenous since older adults with a higher propensity to be depressed are perhaps more likely to report experiencing more childhood adversity, causing selectivity bias. Propensity score matching was used in the present study to mitigate the endogeneity problem caused by selectivity bias, including nearest neighbor, radius, and kernel matching. According to three matching methods, Table 3 shows the average treatment effects on the treated (ATT) of the seven childhood adversities on depression in older adults. All T-values are significant at 1%. The similarity of the results obtained by different matching methods suggests that ATT estimates are robust, suggesting that childhood adversities are still relevant to depression.

The “dose–response” effect of childhood adversity was tested with a multiple linear regression model using a dummy variable representing the number of childhood adversity experienced by the sample as the independent variable and depression score as the dependent variable. Table 4 showed that childhood adversity had a significant dose–response effect on depression in older adults. Older adults with one childhood adversity scored higher on depression by 1.440 units; Older adults with two childhood adversities scored higher on depression by 2.646 units; And older adults with three childhood adversities scored higher on depression by 4.122 units. Specifically, the more childhood adversity an older adult experiences, the more significant the impact on their depression levels.

Using the total number of social participation activities in the sample of 6,704 as the moderating variable, we examined the moderating effect of social participation on childhood adversity and depression. To demonstrate the differences in moderating effects for different age groups of older adults, the sample was divided into two groups: the young old and the oldest old. Results showed that among older adults of all ages, the product terms “poor family economic situation” and “social participation” negatively correlated with depression (p < 0.05). Older adults of all ages were positively associated with “lack of mutual assistance within the neighborhood” and social participation, as well as the product term “lack of good playmates” and social participation. It was found that the product term “loneliness due to lack of friends” and social participation were positively associated with depression among the oldest old (p < 0.1), which is statistically significant. Table 5 shows the results.

According to the study, female older adults are more likely to be depressed, which corresponds to traditional Chinese medicine’s theory that “women are more depressed” (Wang et al., 2022). Possibly, this is because women are by nature sensitive, and they have weaker emotional stability and the ability to resist pressure (Hu, 2021). It has been found that the oldest-old have a lower depression level. According to “social–emotional selection theory,” older adults are better at reconciling the conflict between behaviors and ideas due to years of accumulation and experience fewer negative emotions (Chen and Jordan, 2018). Unmarried older adults have a higher risk of depression, mainly because they face greater pressure from the public, which negatively affects their psychological well-being (Li, 2021). As a result of their low incomes, agricultural household registration, and farming type of work before retirement, older adults with agricultural household registration are more likely to suffer from depression since the above-mentioned old adults cannot obtain more social resources easily. There is a wide gap between urban and rural economic and healthcare resources, making it difficult for them to obtain relevant medical care and social support (Angelini et al., 2019; Yuan et al., 2023); A higher level of education leads to a lower level of depression among older adults because it allows them to understand their situation from the correct perspective of culture and economics. Meanwhile, education can also increase their employment opportunities, improve their incomes and socioeconomic status, and lower their depression levels (Back and Lee, 2011). The study indicates that older adults who do not drink alcohol have a higher degree of depression, which is consistent with existing research (Cai et al., 2022). It is mainly because drinking alcohol can help older adults reduce stress so that they can “drown their sorrows,” and moderate drinking can help them cope with depression. Thus, we need to focus on improving the mental health of those mentioned above key older adults by implementing targeted early intervention strategies.

According to multiple linear regression results, older adults who experienced childhood adversity in a family or social setting are more depressed, which is consistent with existing studies (Domènech-Abella et al., 2021; Cross et al., 2023). A propensity score matching test revealed that childhood adversity still had an explanatory effect on depression in older adults despite some endogenous issues in the original model. Compared with older adults without childhood adversity, those who experienced childhood adversity were more likely to experience depression. This result is consistent with the findings reported in previous studies (Merrick et al., 2017). Furthermore, the dose–response results showed that childhood adversities had a more significant impact on depression levels in older people when the number of adversities was high. This implies that an increase in the number of childhood adversities exacerbates the negative impact on children’s psychological development, leading to higher outcome risks (Finkelhor et al., 2011). This conclusion highlights the cumulative effect of multiple childhood adversities, which is more detrimental to mental health than any single adversity, thereby significantly affecting the psychological and personality health of older adults.

Results of the moderating effect study suggest that social participation appears to reduce the negative impact of “poor family economic situation” on depression in early childhood environments for older adults of all ages. This is mainly because social participation in later life can benefit older adults and may compensate for a lack of social participation in their youth due to insufficient economic resources. In the meantime, various forms of social participation can enhance their sense of self-efficacy and social identity, weakening the inferiority complex resulting from low-income family economic standing in childhood. Additionally, it will increase the possibility of older adults getting social support, which will aid them in alleviating depression. For instance, community volunteer activities and interest groups can provide opportunities for older adults to interact and collaborate with others, thereby expanding their social support network and improving their mental health. Recreational activities such as dancing and fitness not only promote physical health but also enhance emotional and psychological well-being through social interaction.

However, social participation can exacerbate the problem of “a lack of mutual assistance within the neighborhood” and “a lack of good playmates” during the early stages of life. As a result of this counterintuitive finding, we can observe that social participation regulation is heterogeneous in response to different influences of childhood adversity and depression. According to studies, when designing the living environment for older adults with emotional disorders, it is essential to consider the emotional instability of the older adults due to illness so that an environment can be stable and familiar (Tu, 2016). Researchers will examine the environment’s ability to evoke positive emotions and strengthen inherent memories (Liu and Zhu, 2019). In addition, some studies have found that music can assist older adults with cognitive disorders by evoking their happy memories to prevent and degrade their agitated behavior (Pan, 2013). Memories and emotions are triggered by social environments and specific characteristics (Ellis et al., 2011), and similar environments evoke the same memories and emotions. Inferentially, childhood adversity, which occurred in the social environment in the early years, as well as “lack of mutual assistance within the neighborhood” and “lack of good playmates,” have profoundly affected children’s mental health. In their later years, older adults usually engage in social participation behaviors in their social environment. During this time, more social participation behaviors are likely to evoke negative memories of the trauma experienced in the social environment during the early years of older adults, which will further exacerbate their depression tendencies and damage their mental health. At the same time, social participation will aggravate the depression degree of “lonely due to lack of friends” in the early social environment for the oldest-old; the reasons are similar to the mechanism of the above two regulatory effects. Nevertheless, this adjustment has a significant effect only on depression among the oldest-old. There is a possibility that “lonely due to lack of friends” is more harmful to children than “lack of good playmates.” Compared to the young-old with the same childhood adversity, the oldest-old feel more loneliness due to lack of friends in the early years. Meanwhile, more of the oldest-old are experiencing physical and degenerative changes, and social participation will require more energy. In conjunction with the above factors, the adjustment of social participation worsens the psychological condition of the oldest-old who are “lonely due to lack of friends” during their early years. Due to their low psychological resilience, older adults who have experienced childhood adversity are more likely to suffer from depression if they participate excessively in social activities, so they should be encouraged to participate moderately in social activities.

In this study, there is strong evidence to suggest that social participation moderates the relationship between childhood adversity and depression in older adults, which suggests that social participation may not only cushion the impact of childhood adversity but may also regulate some of the influence between childhood adversity and depression. Therefore, this study proposes the following recommendations: Clinical interventions can guide older adults toward moderate social participation to enhance resilience and reduce depressive symptoms, especially for depression patients with a history of childhood adversity, by providing personalized treatment plans. Increasing moderate social participation enables older adults to access more social support and emotional interaction, thereby enhancing self-efficacy and psychological resilience, effectively alleviating symptoms of depression. Additionally, governments and communities should collaborate to create opportunities and platforms for older adults to engage in social activities. Tailored social participation activities should be designed specifically for older adults with different childhood adversities to avoid triggering negative memories. For example, organizing low-intensity, low-pressure social activities such as cultural and recreational events, craft workshops, etc., can help older adults build new positive memories and social connections, promoting psychological well-being. Furthermore, additional psychological protection measures should be provided for older adults reporting childhood adversity, such as establishing professional psychological counseling and support services, offering regular mental health assessments and interventions to help them manage the impact of early-life trauma. Through these preventive measures, the health risk chain from childhood to late adulthood can be broken, improving the psychological health status of older adults and reducing the likelihood of depression occurrence.

There are some limitations to this study as well. Firstly, it is a cross-sectional study, which limits our ability to infer causal relationships regarding the long-term development of depression levels. Future research should employ longitudinal study designs to track changes in depressive symptoms among older adults over time, enabling a more robust establishment of causal relationships over time. At the same time, due to the research objectives, this study did not utilize the most recent 2020 CHARLS data. This choice may limit the timeliness and accuracy of the findings. Future research will consider using the latest data sets to enhance the comprehensiveness and precision of the study. Secondly, older adults may experience recall bias when recalling childhood adversity, potentially impacting the accuracy of study outcomes. Cognitive decline in older adults may hinder accurate recollection of childhood experiences, leading to memory omissions or errors. Such recall bias could either underestimate or overestimate the impact of childhood adversity on depression in older adults. To mitigate this issue, the CHARLS questionnaire was designed to be as detailed and specific as possible, aiming to minimize recall bias among respondents. However, this measure cannot completely eliminate the influence of bias. Therefore, future studies could consider using more objective measurement tools and methods, such as corroborating self-reports with family members’ recollections or community records. Thirdly, while this study’s sample represents the middle-aged and older adults in China, its findings may be influenced by cultural backgrounds. Future research should delve deeper into the cultural specificity of study results and consider conducting cross-cultural comparative studies to better understand the impact of childhood adversity on psychological health among older adults across different cultural contexts. Finally, although this study addressed selection bias through propensity score matching, endogeneity issues due to omitted variables may still exist given our focus on the impact of childhood adversity on current depressive symptoms in older adults. Future research should further explore and control for other potential factors that may influence both childhood adversity and depressive symptoms in later life to enhance the robustness and credibility of the results.