A Study on the Heterogeneity of Social Engagement among Older Adults: An Empirical Analysis based on the 2018 CLASS Data

doi:10.21203/rs.3.rs-3099913/v1

Download PDF

Research Article

A Study on the Heterogeneity of Social Engagement among Older Adults: An Empirical Analysis based on the 2018 CLASS Data

https://doi.org/10.21203/rs.3.rs-3099913/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

Social participation, as an important part of actively responding to population aging, plays a key role in promoting the utilization of the elderly population dividend in the context of China's longevity society. This article addresses the flaw in social participation research that ignores the heterogeneity of social participation and explores the heterogeneity of social participation among older people from the perspective of socioeconomic status.

Methods

Using latent class analysis to generate a comprehensive index of socioeconomic status, and employing multilevel linear models to examine the differences in social participation outcomes among different socioeconomic status groups.

Results

The results show that the higher the socioeconomic status of the elderly group, the higher their level of social participation. The improvement of the socioeconomic level in the region has a significant positive effect on the social participation of the elderly group with lower socioeconomic status.

Conclusions

Therefore, this article reveals the heterogeneity of social participation among older people in terms of socioeconomic dimensions, highlighting the need for policy intervention to further improve the regional socioeconomic development level to promote the social participation of older people and advance social participation research.

Social Participation

Socioeconomic Status

Older Adults

Heterogeneity

Due to population aging, social costs such as healthcare burdens are expected to increase, leading to a growing concern for factors that contribute to maintaining independence and health among older adults (Gee & Gutman, 2000). The aim is to uphold the value of older adults while ensuring their well-being. Fortunately, compared to previous generations, older adults today have experienced significant improvements in physical fitness, economic status, and cultural levels. In the context of China's transition into an aging society (Yang Juhua, 2022; Zhu Hui, 2022), numerous scholars have advocated for leveraging the dividend of an aging population to alleviate social burdens, emphasizing the need to promote and harness the value of older adults through their active participation (Lu Jiehua, 2022; Xie Lili, 2021). Many researchers believe that active social engagement facilitates successful aging and enhances the value of older adults, encompassing "activities and social roles that enable individuals to survive and thrive in society throughout their lives" (Levasseur et al., 2004; Chapman, 2005).

Existing research has already confirmed the positive correlation between social engagement and various aspects of physical and cognitive health among older adults (Bukov, Maas & Lampert, 2002; Ertel, Glymour & Berkman, 2009). Most scholars advocate for increasing social engagement among older adults as an effective means to achieve better health in their later years. For instance, Herzog et al. (2002) stated, "Maintaining active social networks is beneficial for older adults, and this behavior can be self-initiated, relatively easy to execute, does not require high-tech demands, and is not expensive." This statement supports the hypothesis that the power to promote the health of older adults lies in individuals choosing to engage in social activities (Holstein & Minkler, 2003).

While the benefits of social engagement have been widely acknowledged by researchers, differing opinions still exist. Tracing back to the origins of gerontological studies, the Disengagement Theory (Cumming & Henry, 1961) and the Activity Theory (Havighurst, 1966) represent two contrasting perspectives on whether social engagement among older adults should be encouraged. It is evident that not all older adults are encouraged to engage socially. The divergence of these two viewpoints can be attributed to the social demands corresponding to different historical periods at the macro level, and it mainly stems from a crucial issue at the micro level, namely, the autonomy of older adults. For example, some studies have shown that the happiness of late life is not associated with social engagement itself but rather with the individual's free choice and discovery of meaningful activities (Chapman, 2005; Katz, 2000). This argument is based on earlier research, which found that individuals perform best in environments they perceive as voluntary, as it gives them a sense of control, satisfaction, and evokes positive emotions (Solomon, Sneed & Serow, 1979). For instance, some older adults achieve higher levels of happiness in later life by opting for less rather than more engagement (Chapman, 2005). In other words, the ability to choose one's behavior freely and in accordance with personal will is crucial for achieving desired outcomes. Therefore, in the realm of social engagement among older adults, the maximization of health benefits and social value depends on voluntary and self-directed participation.

However, it should be noted that not all older adults have the freedom to choose or the opportunity for unrestricted choice. We must recognize the barriers that exist in older adults' social engagement. Existing research has explored various factors influencing social engagement among older adults, such as human capital (e.g., education, income, wealth), social capital (e.g., social relationships, membership in social groups), health capital (e.g., physical health, disability status), and cultural capital (e.g., religion) (Butrica et al., 2009; Egerton & Mullan, 2008; Li & Ferraro, 2006; Mutchler et al., 2003; Taniguchi, 2006; van Groenou & van Tilburg, 2012). These studies have examined the relationships between these factors and social engagement but have rarely delved into their patterns within socioeconomic groups and the impact of the individual's socioeconomic environment.

In our study, we aim to further expand the influence of socioeconomic factors on social engagement among older adults. Compared to cultural influences on older adults' behavior, socioeconomic factors, including individual socioeconomic status and the socioeconomic level of the region, are more direct factors. This aspect of research on the influencing factors of social engagement among older adults is relatively scarce internationally and domestically, with most studies focusing on the impact of individual socioeconomic status on older adults' physical activity (Hyyppa, 2003; Lindström, Hanson & Östergren, 2001). We believe that physical activity falls within the realm of social engagement, and it is necessary to further explore the relationship between socioeconomic factors and a more comprehensive range of social engagement.

Therefore, in order to advance the argument that the choice of social engagement is important for late-life health, our study focuses on the external factors influencing social engagement among older adults and seeks to further explore the characteristics of social engagement among older adults in different socioeconomic environments. The research on this issue will contribute to proposing more targeted interventions or policy approaches to promote social engagement among older adults and facilitate a happy and healthy late life.

2.1 Social Engagement among Older Adults

Social engagement among older adults can be defined as the participation in activities involving interaction with others in the community or society (Levasseur et al., 2010). These activities encompass various forms of engagement, including volunteering, physical exercise, leisure activities, religious involvement, and political participation (Maier & Klumb, 2005). In our study, we primarily focus on these five forms of social engagement.

Current research on social engagement among older adults predominantly centers around two key questions. The first question pertains to the relationship between social engagement and the well-being of older adults in their later years. Studies indicate that social engagement and connections with others can enhance physical and cognitive health in later life (Mitka, 2001). For instance, social engagement in the form of community involvement, participation in neighborhood associations, church groups, or non-governmental organizations, has been found to have beneficial effects on maintaining and restoring daily activities and preventing disabilities (Mendes de Leon, Glass & Berkman, 2003).

Given the positive impact of social engagement on the health of older adults, actively promoting social engagement has been considered an important component of addressing the challenges posed by population aging, aiming to maintain the health and value of older adults (Tong Yufen, 2020). However, the increased demand for contributions from older adults does not necessarily translate into increased opportunities for social engagement. Various barriers hinder the social engagement of older adults, which can be classified as either "internal" or "external" obstacles. "Internal" obstacles, such as lack of motivation or limited leisure time, are more common among higher socioeconomic groups, while "external" barriers, such as lack of financial resources, transportation, or illness/disabilities, are more prevalent among lower socioeconomic groups (Chinn, White, Harland, Drinkwater & Raybould, 1999). This highlights the significant heterogeneity in social engagement among subgroups of older adults.

The second question in the research on social engagement among older adults concerns the differences in social participation among individuals with different characteristics. For example, retired older men from white-collar professions with higher wealth levels tend to engage more actively in volunteer work, social clubs, and service organizations compared to those from blue-collar occupations with lower wealth levels (Harwood, Pound & Ebrahim, 2000; Warburton & Stirling, 2007). Older adults with higher levels of education and better social skills are more likely to volunteer in a wider range of religious, educational, political, senior, and other organizations and contribute more time (Tang, 2008). While these studies provide evidence of the relationship between individual characteristics, socioeconomic status, and social engagement, they do not explore how regional socioeconomic characteristics shape these patterns of social engagement among older adults.

Therefore, it is necessary to further investigate the factors that limit social engagement among older adults and examine to what extent individual and regional socioeconomic conditions constrain their freedom to engage in social activities. Empirical research addressing this question, particularly in the context of China, is still needed.

2.2 socioeconomic status and social participation

From a theoretical perspective and previous research findings, we establish the arguments and hypotheses of this study. Firstly, from a theoretical perspective, we draw insights from critical gerontology and Bourdieu's theory of habitus and capital to explore the constraints on social participation choices in later life. These perspectives are often used in the study of inequality (Bezanson, 2006; Bourdieu, 1999). They provide a critical view on why people act in certain ways and the underlying reasons for these behaviors.

Critical gerontology, rooted in a political-economic perspective (Quadagno & Reid, 1999), emphasizes the sensitivity to social, political, and economic factors in understanding the lived experiences and needs of older adults (Ray, 1996). We utilize this perspective to understand the material and cultural foundations of the existing age-based social order and explore whether certain individuals within the older population have greater freedom of choice (Holstein & Minkler, 2003). In studying older adults' social participation, it is important to examine not only the health-related benefits experienced by those engaged in social participation but also the constraints faced by those who do not participate. This helps to examine the structural limitations of socioeconomic conditions underlying older adults' social participation behaviors, which may go beyond individual factors.

Bourdieu's theory of habitus and capital highlights how the structural inequality of class, which constitutes material disadvantage, can potentially limit older adults' choices in social participation. Habitus refers to the internalization of cultural discourse that shapes people's behavioral tendencies, determining what is considered reasonable or unreasonable behavior. Individuals from middle-class and upper-class backgrounds may possess a "cultivated" habitus that inclines them towards more extensive social engagement. Higher socioeconomic status older adults also possess corresponding human, cultural, and social capital. Factors such as education level, income, and social networks contribute to their proactive engagement and ability to participate in social activities (Bourdieu, 1999).

Furthermore, previous research has accumulated relevant findings. It has been found that individuals with higher socioeconomic status, specifically those with higher education levels, tend to participate in groups that emphasize personal development, such as part-time work and continued learning, while older age and being married are associated with lower participation in these activities. Additionally, the likelihood of engaging in any form of social participation is lower among individuals with poor physical health (Minagawa & Saito, 2015). Lower socioeconomic status groups have a lower risk of engaging in physical activities during leisure time (Lippke & Cihlar, 2021), while higher-income individuals are more likely to have better mental and physical health (Sun et al., 2021). The groups chosen by higher socioeconomic status individuals may offer higher levels of activity or better health protection (Nummela et al., 2008; Ashida, Kondo & Kondo, 2016), providing opportunities for building social relationships that positively impact health (Lee et al., 2008). Correspondingly, lower socioeconomic status groups exhibit a smaller proportion of social participation (Ashida, Kondo & Kondo, 2016). Lack of financial resources may limit some individuals' full participation in society (Niedzwiedz et al., 2016). People living in areas with lower social capital are at a higher risk of poor health outcomes (Sundquist et al., 2004). Therefore, social participation has been observed to be influenced by socioeconomic status.

Moreover, compared to research on the impact of individual socioeconomic status on older adults' social participation, there is limited research on the influence of regional socioeconomic levels on social participation, particularly in the context of physical activity. Studies in this area have identified influences of both individual and environmental determinants on social participation in different regions (Lindström, Merlo & Östergren, 2002). Kamphuis found socioeconomic inequalities in sports participation, with adverse perceived neighborhood factors (e.g., feeling unsafe, limited social networks), family factors (material and social deprivation), and individual perceptions of physical activity (e.g., negative outcome expectations, low self-efficacy) significantly associated with physical inactivity, particularly prevalent among lower socioeconomic groups (Kamphuis et al., 2008). The conclusions of this research need to be further explored within the social context of China.

Given that social participation may be influenced by structural inequality and environmental factors, it is important to examine how the assumption that older adults can choose their lifestyles, including their social participation, is established (Gilleard & Higgs, 2007). However, there has been limited discussion on this topic in the academic community. Based on previous research findings, we propose the following research hypotheses for our study on social participation in later life, which are influenced by individual socioeconomic status and the broader socioeconomic environment:

Hypothesis 1

Older adults with higher socioeconomic status have higher levels of social participation.

Hypothesis 2

Regions with higher socioeconomic characteristics are more favorable for older adults' social participation levels.

3.1 Data

The data used in this study is derived from the China Comprehensive Social Survey of the Elderly (CLASS2018), conducted by the Institute of Gerontology at Renmin University of China. The survey aims to comprehensively understand the basic information, health status, and economic conditions of the elderly population. It focuses on capturing information on the elderly's long-term care service needs, long-term care resources, attitudes and plans towards aging, intergenerational relationships, and social adaptation, among other aspects. The survey employed a stratified multistage probability sampling method, with survey samples distributed across 134 counties/districts and 462 villages/residential areas nationwide. A total of 11,511 valid individual questionnaires and 462 community survey questionnaires were collected. As this study specifically focuses on the issue of social participation among the elderly, after variable screening and missing value treatment, a final sample of 11,402 valid respondents aged 60 and above was obtained.

3.2 Variables and Measurements

In this study, the dependent variable is the level of social participation among older adults, primarily measured through five aspects of social participation: physical exercise, leisure activities, religious participation, volunteer participation, and political participation. These aspects were measured by asking respondents whether they engaged in activities such as square dancing, playing mahjong/chess/cards, participating in religious activities, engaging in volunteer service, and voting in local residents' committee/village committee elections in the past year or three years. A score ranging from 0 to 5 was assigned to each aspect, where 1 indicates participation and 0 indicates non-participation. The overall level of social participation was then measured by summing up the scores of these five aspects, with a higher score indicating a higher level of social participation among older adults.

The main explanatory variable in this study is socio-economic status. Socio-economic status was operationalized from an objective perspective. Previous research suggests that the objective aspect of socio-economic status refers to individuals with higher socio-economic status having more money, higher levels of education, and occupational prestige (Goodman et al., 2001). However, it has been argued that such measurements may not be appropriate for measuring the socio-economic status of older adults in developing countries (Zhu & Xie, 2007). Additionally, in the social and cultural context of China, some scholars argue that household registration (hukou) is an important factor in determining socio-economic status (Wu & Treiman, 2004). Therefore, some studies have included consideration of household registration when measuring socio-economic status among older adults (Jiao Kaishan, 2014). In this study, we adopt the conclusions of the aforementioned research. Considering that older adults are typically in retirement, the measurement of occupational prestige was not included. Thus, the operationalization of socio-economic status includes three aspects: education level, income, and household registration. In this study, household registration is divided into two categories: rural household registration and non-rural household registration. Education level is categorized into four levels based on the respondent's highest level of education completed: illiterate, primary school, middle school, and high school or above. Income is divided into four categories based on personal annual income: low income (below 2300 yuan), low-middle income (between 2300 and 5000 yuan), middle-high income (between 5000 and 10000 yuan), and high income (above 10000 yuan). For each individual socio-economic status indicator mentioned above, a latent class analysis (LCA) method is used to construct a comprehensive index of socio-economic status. By comparing the fit indices of the model, the model with the optimal number of latent class variables is selected. In other words, based on the education, income, and household registration variables selected in this study, the respondents can be broadly divided into three groups representing high, middle, and low socio-economic status.

Furthermore, this study also considers the income level of the area in which older adults live as an environmental factor that may influence the relationship between socio-economic status and social participation. The average income categorized by community is calculated as a measurement of the area's socio-economic level, reflecting the affluence of each community. Finally, the control variables in this study include gender, age, self-rated health, marital status, ethnicity, and religious beliefs.

3.3 analysis methods

The purpose of this study is to explore the heterogeneity of elderly individuals' social engagement in the context of social status, taking into consideration the influence of regional factors. Specifically, we aim to investigate whether the relationship between different levels of social and economic status and the level of social engagement among elderly individuals is influenced by the socioeconomic level of the region they reside in. To address this research question, a multilevel modeling approach is employed, which is widely favored for examining the impact of individual and community characteristics on the phenomena of interest. Traditional statistical models often overlook the similarities among individuals in different geographic backgrounds, potentially leading to inaccurate results. In contrast, multilevel modeling techniques not only account for the complex error structure arising from clustering of individuals within geographical regions but also effectively handle biases associated with missing data and measurement errors, enabling a more precise understanding of the underlying driving factors behind the phenomenon (Duncan, Jones, & Moon, 1998).

In this study, a multilevel linear regression model, known as the Multilevel Linear Model, is utilized to fit the data, with social engagement level serving as the dependent variable. Through stepwise nesting of models, relevant explanatory variables are progressively included, starting from a simple model, to construct a complex final explanatory model.

The specific steps for constructing the model are as follows: First, a random intercept model, without any explanatory variables, is built. The variation of social and economic status across different levels is examined, resulting in the null model, Model 1:

$${y}_{ij}={\beta }_{0\text{j} }+{r}_{ij}, {r}_{ij} \sim N\left(0, {\tau }_{00}^{2}\right)$$

$${\beta }_{0\text{j} }={\gamma }_{00 }+{u}_{0\text{j} }, {u}_{0j} \sim N\left(0, {\sigma }_{}^{2}\right)$$

$${y}_{ij}={\gamma }_{00 }+{u}_{0\text{j} }+{r}_{ij}$$

In this context, the notation used is as follows: $\text{i}$ represents the level-one unit, which refers to the individuals being surveyed. $\text{j}$ represents the level-two unit, which corresponds to the community units. ${y}_{ij}$ represents the score of social engagement level. ${\beta }_{0\text{j} }$ represents the fixed effect, which denotes the mean of social engagement level scores. ${u}_{0j}$ and ${r}_{ij}$represent the random effects, representing the between-group variance and within-group variance, respectively. ${\sigma }_{}^{2}$ represents the variance at the individual level. ${\tau }_{00}^{2}$represents the variance at the community level.

Building upon Model 1, Model 2 includes the addition of control variables such as gender, age, self-rated health, marital status, ethnicity, and religious affiliation. On the foundation of Model 2, Model 3 introduces the main explanatory variable of individual-level socioeconomic status, represented by dummy variables indicating middle socioeconomic status and high socioeconomic status. Expanding on Model 3, Model 4 incorporates a high-level independent variable, namely community-level socioeconomic status, to explore its impact on social engagement. Model 4 is then enhanced by introducing random coefficients for socioeconomic status, resulting in a Random Coefficient Model (Model 5). This signifies that the relationship between socioeconomic status and social engagement varies across different regions. Lastly, Model 6, built upon Model 5, includes cross-level interaction terms. These terms account for the interaction between socioeconomic status and the regional income variable, allowing an examination of whether the relationship between socioeconomic status and social engagement is moderated by regional income level. The specifications for Model 6 are as follows:

$${y}_{ij}={\beta }_{0 }+{\beta }_{1}{a}_{nij}+{\beta }_{2}{b}_{ij}+{\beta }_{3}{a}_{nij}*{b}_{ij}+{\sum }_{k=4}^{p}{\beta }_{\text{k}}{x}_{kij}$$

$$+({u}_{0j}+{u}_{1j}{a}_{nij}+{r}_{0ij})$$

In Model 6, the notation is as follows: $a$ represents the socioeconomic status (n = 2/3).$b$ represents the regional socioeconomic level. ${x}_{\text{k}}$ represents a set of control variables. ${a}_{ij}*{b}_{ij}$ represents the cross-level interaction between socioeconomic status and regional socioeconomic level. The intercept term is also considered a random variable.

The estimation of the multilevel linear models is performed using the maximum likelihood estimation method. To compare the fit of the models, this study utilizes two information criteria: AIC (Akaike Information Criterion) and BIC (Bayesian Information Criterion). The model estimation in this study is conducted using the "mixed" command in Stata statistical software.

4.1 Descriptive Study Results

Table 1 presents the descriptive statistics of the sample used in this study, which consisted of 11,402 participants. Firstly, in terms of the overall social participation level of the elderly, it is observed to be relatively low with a mean score of 1.46. When examining each dimension of social participation, it is found that leisure participation and political participation are relatively more prevalent, while physical exercise, volunteer participation, and religious participation show similar proportions, but are lower than leisure and political participation. Secondly, in terms of the regional income level, the total score is 2.45, with a mean value of 1.56, indicating that there is still room for improvement in community efforts to promote social participation among the elderly. Lastly, looking at the statistics of the controlled confounding factors, it is noted that the distribution of male and female participants in the sample is relatively balanced. The average age of the sample is around 71 years, and the self-rated health status is generally reported as average. More than half of the elderly participants have spouses, with the majority being of Han ethnicity, while a minority of the elderly have religious beliefs. Table 2, Table 3, and Table 4 provide the results of latent class analysis for the primary explanatory variable at the individual level, namely social-economic status. In this study, latent class analysis was conducted on three observable socioeconomic variables: education level, income, and household registration, to estimate a comprehensive measurement of social-economic status. Table 2 presents the fitted models for the three latent class models. By comparing the AIC and BIC values, it is observed that Model M3 has the lowest values among the three models, indicating that Model M3 with three latent classes is the optimal choice for measuring social-economic status among the elderly in this study. Table 4 reports the predicted probabilities (latent class marginal means) of high education level, high income, and rural household registration within the three latent classes. It is noted that in latent class 3, the predicted probabilities of high education level and high income are the highest, while the predicted probability of non-rural household registration is the lowest. Based on this, it can be inferred that latent class 3 represents the group with the highest social-economic status. On the other hand, in latent class 1, the predicted probability of rural household registration is the highest, while the predicted probabilities of income level and education level are the lowest, indicating that latent class 1 corresponds to the group with low social-economic status. Latent class 2 represents the group with intermediate characteristics across the categories. Therefore, it can be concluded that latent class 2 represents the group with moderate social-economic status. Based on the posterior probabilities estimated by latent class model M3, the latent class membership for each sample is determined. Specifically, there are 2,835 samples classified into the low social-economic status group, 4,505 samples classified into the moderate social-economic status group, and 4,505 samples classified into the high social-economic status group (refer to Table 4 for details).

Table 1

Descriptive Statistics of Variables
Variable Name	Sample Size	Mean	Standard Deviation	Minimum	Maximum
Dependent Variable
Level of Social Participation	11402	1.46	1.53	0	5
Physical Exercise	11402	0.25	0.43	0	1
Leisure Participation	11402	0.42	0.493	0	1
Volunteer Participation	11402	0.20	0.397	0	1
Religious Participation	11402	0.19	0.389	0	1
Political Participation	11402	0.41	0.492	0	1
Individual-level Explanatory Variables
Education Level	11402	1.16	0.93	0	3
Personal Annual Income	11402	1.56	1.25	0	3
Household Registration	11402	0.53	0.50	0	1
Community-level Explanatory Variable
Regional Income Level	11402	1.56	0.40	1.16	2.45
Control Variables
Gender (Male = 1)	11402	0.50	0.50	0	1
Age	11402	71.45	7.37	60	108
Self-rated Health	11402	2.69	0.90	1	5
Marital Status (Married = 1)	11402	0.69	0.46	0	1
Ethnicity (Han Chinese = 1)	11402	0.95	0.22	0	1
Religious Belief (Yes = 1)	11402	0.07	0.25	0	1

Table 2

Three Latent Class Model Data (N = 11402)
Model Number of Latent Classes	Sample Size	Degrees of Freedom (df)	AIC	BIC
M1: Class 1	11402	3	42967.990	42990.010
M2: Class 2	11402	7	40430.130	40496.210
M3: Class 3	11402	9	40426.130	40477.520

Table 3

Predicted Probabilities of Three Variables in Two Different Latent Classes (N = 11402)
Socioeconomic Status (SES) Indicator	Measurement	Single-Class Predicted Probabilities
Socioeconomic Status (SES) Indicator	Measurement	Latent Class 1 (Low SES)	Latent Class 2 (High SES)	Latent Class 3 (High SES)
Household Registration	Agricultural Hukou	1	.5513103	.0299441
Education Level	Education Level	.5440694	.7015687	1
Income	Annual Income	.2625756	.7605635	.9920393

Table 4

Sample Estimates of Posterior Probabilities of Latent Class Membership
	Frequency	Percentage	Cumulative Percentage
Low Socioeconomic Status	2835	24.86	24.86
Medium Socioeconomic Status	4505	39.51	64.37
High Socioeconomic Status	4505	35.63	100
Total	11402	100

Subsequently, a multilevel linear regression model was employed to estimate the effects of individual-level social-economic status variables, basic control variables, and regional-level community socioeconomic status variables on social participation level. The results of the model estimation are presented in Table 5 and Table 6.

Table 5

Comparison of Fit Indices for Multilevel Models
Model Number of Latent Classes	Sample Size	Degrees of Freedom (df)	AIC	BIC
Model 1	11402	3	41941.100	41963.120
Model 2	11402	9	41775.370	41841.450
Model 3	11402	11	41654.920	41735.680
Model 4	11402	12	41656.910	41745.010
Model 5	11402	14	41654.090	41756.880
Model 6	11402	16	41644.980	41762.440

Table 6

Estimated Results of Multilevel Model on the Relationship between Socioeconomic Status and Social Participation
Parameter	Social Participation Level
Fixed Parameters
Intercept	0.970^*
Medium Socioeconomic Status (Low)	1.199^***
High Socioeconomic Status (Low)	1.237^***
Regional Income Level	0.517
Regional Income Level X Medium Socioeconomic Status	-0.625^***
Regional Income Level X High Socioeconomic Status	-0.581^**
Gender (Male = 1)	0.007
Age	-0.006^***
Self-rated Health	-0.140^***
Marital Status (Married = 1)	0.083^**
Ethnicity (Han Chinese = 1)	0.225^**
Religious Belief (Yes = 1)	0.290^***
Random Parameters
Level 2 (Regional Level)	0.143
Variance (Intercept)	0.023
Covariance (Medium)	2.82e-12
Covariance (High)	7.46e-15
Level 1 (Individual Level)
Variance	2.246
Note: Standard errors are shown in parentheses. *p < 0.01, p < 0.05, *p < 0.1.

Figure 1: Social Participation Levels of Socioeconomic Status Groups Across Different Regional Income Levels

4.2 Model Fit Comparison

Table 5 presents the AIC and BIC values for different models. From Model 1 to Model 4 and Model 5, the AIC and BIC values progressively decrease, indicating an improvement in model fit. Specifically, Model 1 is an intercept-only model that does not consider the effects at the regional level. Models 1 to 5 are two-level linear regression models that incorporate regional effects. Model 4, which includes both individual-level explanatory variables and control variables, as well as regional-level explanatory variables, shows a significant improvement over Model 1, indicating that the added control variables have a significant effect on social participation levels.Model 4 has a higher AIC than Model 5, but a lower BIC than Model 5. Considering our research objectives to examine the role of cross-level interactions, we prioritize the model with a lower AIC. Therefore, Model 5 is chosen as the final model for further interpretation in Table 6.

4.3 Heterogeneity in Social Participation Levels among Different Socioeconomic Status Groups

According to the results of Model 6, compared to older adults in the low socioeconomic status (SES) group, those in the middle SES group have a 1.204-point higher social participation level, and those in the high SES group have a 1.208-point higher social participation level. This indicates that higher socioeconomic status is associated with higher levels of social participation among older adults. Additionally, the cross-level interaction term between SES and regional income is significant (p = 0.003), suggesting that the relationship between socioeconomic status and social participation is moderated by regional income.

Furthermore, to examine whether the relationship between socioeconomic status and social participation varies across different regional socioeconomic characteristics, we introduced an interaction term between individual socioeconomic status and regional income level in the model. Although regional income level did not show a significant effect, which may be due to sampling error, the cross-level interaction term was significant. As it showed in Fig. 1, we observed that as regional income level increases, there is a significant increase in social participation level among older adults in the low socioeconomic group, while the social participation level of older adults in the middle and high socioeconomic status groups remains relatively stable, with a slight decreasing trend. However, this slight trend may be attributed to sampling error.

This study utilized the 2018 "China Longitudinal Aging Social Survey (CLASS)" data to examine the impact of socioeconomic status on social participation among older adults. Three variables related to socioeconomic status, namely education level, income, and household registration (hukou), were selected. Through latent class analysis, a composite measurement index of socioeconomic status was constructed, categorizing older adults into three groups: low socioeconomic status, middle socioeconomic status, and high socioeconomic status. The study analyzed the influence of socioeconomic status on social participation among older adults and further investigated whether there were variations in the heterogeneity of social participation across different income levels in the regions using multilevel modeling

In comparison with previous studies, we further examined the heterogeneity of social participation among older adults and explored the interaction between individual and regional levels using a multilevel data structure. Specifically, we investigated whether the effects of socioeconomic status on social participation are influenced by the socioeconomic characteristics of the individuals' residing regions.

The results of this study support Hypothesis 1, which posits that higher socioeconomic status is associated with higher levels of social participation among older adults. Our findings are consistent with previous research, suggesting that older adults with higher socioeconomic status possess higher levels of social capital and other resources, which provide them with a broader range of social participation choices. Older adults with sufficient economic resources can engage in preferred social and leisure activities, substitute paid work for volunteering when volunteering becomes burdensome, reduce their voluntary contributions, or even consider relocating to another community that better meets their needs (Rozanova, Julia et al., 2012).

Regarding Hypothesis 2, our results partially support the notion that regions with higher socioeconomic characteristics are more conducive to social participation among older adults. The increase in regional socioeconomic levels significantly promotes the social participation of older adults with low socioeconomic status, while the social participation levels of older adults with middle and high socioeconomic status remain relatively stable, with a slight declining trend. Similar findings have been reported in previous studies, suggesting that unfavorable community environments (e.g., feelings of insecurity, low social networks), family factors (e.g., material and social deprivation), and individual factors (e.g., low self-efficacy, perceived negative outcome expectations) are associated with reduced social participation, and are more prevalent among lower socioeconomic groups. One possible explanation for this trend is that communities with lower socioeconomic status have fewer resources related to social participation or a lower proportion of freely accessible resources (Estabrooks, Paul A et al., 2003), which limits the social participation options for older adults in these groups. It is possible that these individuals have the desire to engage in social participation, but their living environment restricts their social participation behaviors. Scholars have further explored the constraints on social participation in communities with lower socioeconomic development, which may be attributed to lower residential security, thus limiting the level of social participation. In addition, higher educational attainment and community economic conditions may promote social participation in leisure activities by providing social, instrumental, and informational resources (Van Lenthe et al., 2005).

Finally, this study has some limitations that warrant further research. Firstly, as this study used cross-sectional data from the 2018 China Health and Retirement Longitudinal Study (CHARLS), it only provides a snapshot of older adults' social participation at a specific time point and does not capture the dynamics or changes in their social participation levels over time. Future research should focus on understanding the factors influencing the stability or changes in older adults' social participation levels. Secondly, this study faces endogeneity issues, as the causal direction between socioeconomic status and social participation cannot be determined based on cross-sectional data alone. Lastly, due to data limitations, we were unable to exclude older adults with physical disabilities. Considering the significant impact of physical disabilities on older adults' social participation levels, future research should collect relevant data to obtain more accurate results.

In conclusion, this study further explores the heterogeneity of social participation among older adults, as well as the potential regional differences in this heterogeneity. We find that low levels of social participation are more prevalent among older adults with lower socioeconomic status, indicating an opportunity to reduce socioeconomic disparities in social participation. To the best of our knowledge, this is the first study conducted in the context of China's social background, which examines the social participation levels of older adults based on their individual socioeconomic status and the socioeconomic development level of their residing communities. Future research should focus on specific approaches at the individual and community levels to better understand how socioeconomic disadvantages contribute to lower levels of social participation. Our findings suggest that policy interventions targeting lower socioeconomic groups should address community factors to maximize social participation among this population and ultimately reduce socioeconomic disparities in social participation.

Ethics approval and consent to participate

The experimental protocol was established, according to the ethical guidelines of the Helsinki Declaration and was approved by the Human Ethics Committee of Minzu University of China . Written informed consent was obtained from individual or guardian participants.

Consent for publication

Not Applicable.

Availability of data and materials

The datasets generated and/or analysed during the current study are available in the CLASS repository, [http://cnsda.ruc.edu.cn/].

Competing interests

The authors declare that they have no competing interests.

Funding

This study was supported in part by grants from by the Graduate Research and Practice Projects of Minzu University of China.

Authors' contributions

Liu Yan analyzed and interpreted data on social engagement of older adults and was a major contributor to writing the manuscript. Final manuscript read and approved by the author.

Acknowledgements

I appreciated the Graduate Research and Practice Projects of Minzu University of China supported my research.

Authors' information

Yan Liu

School of Ethnology and Sociology, Minzu University of China

[email protected]

Gee, E.M., & Gutman, G.M. The overselling of population aging: Apocalyptic demography, intergenerational challenges, and social policy. [M]. New York: Oxford University Press, 2000
Yang, Juhua. "Longevity Era and Longevity Dividend." Journal of Shandong University (Philosophy and Social Sciences Edition), 2022, 253(04): 103-111.
Zhu, Hui. "From Aging Burden to Longevity Dividend: China's Approach under National Strategic Positioning." Journal of Shandong University (Philosophy and Social Sciences Edition), 2002, 253(04): 112-120.
Lu, Jiehua. "Positive Aging Perspectives in the Era of Longevity: Evolution, Essence, and Advantages in Practice." Journal of Shandong University (Philosophy and Social Sciences Edition), 2022, 253(04): 94-102.
Xie, Lili, Wang, Fei, Hu, Kang. "Patterns of Social Participation among Chinese Older Adults and Their Impact on Social Adaptation." Population Research, 2021, 45(05): 49-63.
Tong, Yufen, Liao, Yuhang. "Development of Human Resources for the Aging Population in China under the Silver Wave." Journal of China Institute of Industrial Relations, 2020, 34(02): 27-36.
Levasseur , M. , Desrosiers , J. , & Noreau , L. “Is social participation associated with quality of life of older adults with physical disabilities?” [J]. Disability & Rehabilitation, 2004, 20,1206 – 1213.
Chapman , S.A. 2005. “Theorizing about aging well: Constructing a narrative” [J]. Canadian Journal on Aging, 24 (1), 3 – 18 .
Bukov , A. , Maas , I. , & Lampert , T. 2002. “Social participation in very old age: Cross-sectional and longitudinal fi ndings from BASE” [J].Journal of Gerontology: Social Sciences, 57B , P510 – P517 .
Butrica, B. A., Johnson, R. W., & Zedlewski, S. R. “Volunteer dynamics of older Americans” [J]. Journals of Gerontology, Series B: Psychological Sciences and Social Sciences, 2009, 64, 644–655.
Egerton, M., & Mullan, K. “Being a pretty good citizen: An analysis and monetary valuation of formal and informal voluntary work by gender and educational attainment” [J]. British Journal of Sociology, 2008, 59, 145–164
Ertel , K.A. , Glymour , M.M. , & Berkman , L.F. “Social networks and health: A life course perspective integrating observational and experimental evidence ”[J]. Journal of Social & Personal Relationships, 2009, 26 (1), 73 – 92 .
Herzog , A. , Ofstedal , M. , & Wheeler , L.M. “Social engagement and its relationship to health” [J].Clinics in Geriatric Medicine, 2002, 18 ( 3 ), 593 – 609 .
Holstein , M. , & Minkler , M. “Self, society, and the ‘new gerontology’” . [J].Gerontologist, 2003, 43(1), 787 – 797.
Coleman, P. Cumming E. and Henry W. “Growing Old: The Process of Disengagement” [J]. Ageing & Society, 2009, 11(2), 217-220.
Havighurst RJ. “Successful aging” [J].Gerontologist,1961
Katz , S. “Busy bodies: Activity, aging, and the management of everyday life” [J]. Journal of Aging Studies, 2000, 14(2), 135 – 152.
Solomon , D. , Sneed , M. , & Serow , R. “Some effects on the interaction between locus of control and perceived voluntariness of participation” [J].Journal of Social Psychology, 1979, 108 (1), 129 – 131.
Egerton, M., & Mullan, K. “Being a pretty good citizen: An analysis and monetary valuation of formal and informal voluntary work by gender and educational attainment” [J]. British Journal of Sociology, 2008, 59, 145–164.
Li, Y., & Ferraro, K. F. “Volunteering in middle and later life: Is health a benefit, barrier or both? ” [J]. Social Forces, 2006,85, 497–519.
Mutchler, J. E., Burr, J. A., & Caro, F. G. “From paid worker to volunteer: Leaving the paid workforce and volunteering in later life” [J]. Social Forces, 2003, 81, 1267–1293.
Taniguchi, H. “Men and women’s volunteering: Gender differences in the effects of employment and family characteristics” [J]. Nonprofit and Voluntary Sector Quarterly, 2006, 35, 83–101.
van Groenou, M. B., & van Tilburg, T. “Six-year follow-up on volunteering in later life: A cohort comparison in the Netherlands” [J]. European Sociological Review,2012, 28, 1–11.
Hyyppa, M. T. “Social Participation and Health in a Community Rich in Stock of Social Capital.” [J]. Health Education Research,2003,18 (6): 770–779.
Levasseur, Mélanie, Lucie Richard, Lise Gauvin, and Émilie Raymond. “Inventory and Analysis of Definitions of Social Participation Found in the Aging Literature: Proposed Taxonomy of Social Activities.” [J]. Social Science & Medicine, 2010, 71 (12): 2141–2149.
Lindström, Martin, Bertil S Hanson, and Per-Olof Östergren. “Socioeconomic Differences in Leisure-Time Physical Activity: The Role of Social Participation and Social Capital in Shaping Health Related Behaviour.” [J]. Social Science & Medicine, 2001, 52 (3): 441–451.
Maier , H. , & Klumb , P. “Social participation and survival at older ages: Is the effect driven by activity content or context?” [J]. European Journal on Aging,2005,2, 31 – 39 .
Mitka , M. “Aging patients are advised “Stay active to stay alert ”. [J]. Journal of the American Medical Association,2001, 5 (16), 2437 – 2439 .
Mendes de Leon , C.F. , Glass , T.A. , & Berkman , L.F. “Social engagement and disability in a community population of older adults: the New Haven EPESE” [J]. American Journal of Epidemiology,2003,157 (7), 633 – 642.
Harwood , R.H. , Pound , P. , & Ebrahim , S. “Determinants of social engagement in older men” [J]. Psychology, Health & Medicine ,2000, 5 (1), 75 – 85.
Chinn, D. J., White, M., Harland, J., Drinkwater, C., & Raybould, S. “Barriers to physical activity and socioeconomic position: implications for health promotion” [J]. Journal of Epidemiology and Community Health, 1999, 53, 191–192.
Warburton , J. , & Stirling , C. “Factors affecting volunteering among older rural and city dwelling adults in Australia” [J].Educational Gerontology,2007, 33(1), 23 – 43 .
Bezanson , K. “Gender and the limits of social capital”[J]. Canadian Review of Sociology and Anthropology, 2006,43(4), 427 – 443.
Bourdieu, P. “The weight of the world: Social suffering in contemporary society” [M]. Cambridge, UK : Polity Press, 1999.
Quadagno, J. and Reid, J. “The Political Economy Perspective in Aging” [J]. Handbook of Theories of Aging, 1999.
Ray, R. E. “A Postmodern Perspective on Feminist Gerontology” [J]. The Gerontologist, 1996, 36(5), 674-680.
Ashida, Toyo, Naoki Kondo, and Katsunori Kondo. “Social Participation and the Onset of Functional Disability by Socioeconomic Status and Activity Type: The JAGES Cohort Study.” [J]. Preventive Medicine,2016, 89 (August): 121–128.
Lee, Hyo Young, Soong-Nang Jang, Seonja Lee, Sung-Il Cho, and Eun-Ok Park. “The Relationship between Social Participation and Self-Rated Health by Sex and Age: A Cross-Sectional Survey” [J]. International Journal of Nursing Studies, 2008,45 (7): 1042–1054.
Lippke, Sonia, and Volker Cihlar. “Social Participation during the Transition to Retirement: Findings on Work, Health and Physical Activity beyond Retirement from an Interview Study over the Course of 3 Years” [J]. Activities, Adaptation & Aging ,2021, 45 (2): 135–158.
Minagawa, Yuka, and Yasuhiko Saito. “Active Social Participation and Mortality Risk Among Older People in Japan: Results From a Nationally Representative Sample.” [J]. Research on Aging,2015,37 (5): 481–499.
Niedzwiedz, Claire L., Elizabeth A. Richardson, Helena Tunstall, Niamh K. Shortt, Richard J. Mitchell, and Jamie R. Pearce. “The Relationship between Wealth and Loneliness among Older People across Europe: Is Social Participation Protective?” [J]. Preventive Medicine,2016,91 (October): 24–31.
Nummela, Olli, Tommi Sulander, Ossi Rahkonen, Antti Karisto, and Antti Uutela. “Social Participation, Trust and Self-Rated Health: A Study among Ageing People in Urban, Semi-Urban and Rural Settings” [J]. Health & Place,2008,14 (2): 243–253.
Sun, Jian, Xiaoyin Lyu, Shoujun Lyu, and Rui Zhao. “The Effect of Social Participation on Income-Related Inequality in Health Outcome among Chinese Older Adults” [J]. International Health, 2021,13 (1): 80–88.
Sundquist, Kristina, Martin Lindström, Marianne Malmström, Sven-Erik Johansson, and Jan Sundquist. “Social Participation and Coronary Heart Disease: A Follow-up Study of 6900 Women and Men in Sweden.” [J]. Social Science & Medicine, 2004,58 (3): 615–622.
Lindström, Martin, Juan Merlo, and Per-Olof Östergren.“Individual and Neighbourhood Determinants of Social Participation and Social Capital: A Multilevel Analysis of the City of Malmö, Sweden.” [J]. Social Science & Medicine, 2002 54 (12): 1779–1791.
Kamphuis, Carlijn Barbara Maria, Frank J. Van Lenthe, Katrina Giskes, Martijn Huisman, Johannes Brug, and Johan P. Mackenbach. “Socioeconomic Status, Environmental and Individual Factors, and Sports Participation.” [J]. Medicine & Science in Sports & Exercise,2008, 40 (1): 71–81.
Gilleard , C. , & Higgs , P. “Cultures of ageing” [M]. Edinburgh, UK : Pearson Education, 2007.
Goodman, E et al. “Adolescents' perceptions of social status: development and evaluation of a new indicator” [J]. Pediatrics, 2001, vol. 108,2.
Zhu, Haiyan & Yu Xie. “Socioeconomic Differentials in Mortality among the Oldest Old in China” [J]. Research on Aging,2007, 29(2).
Wu Xiaogang & Donald J. Treiman. “The Household Registration System and Social Stratification in China:1955-1996” [J]. Demography,2004, 41(2).
Jiao, Kaishan. "Factors Affecting Health Inequality: A Study on Sociology." Sociological Studies, 2014, 29(05): 24.
Duncan, Craig, Jones, KelvynMoon, Graham.“Context, Composition and Heterogeneity: Using Multilevel Models in Health Research” [J]. Social Science & Medicine,1998, 46(1):97–117.
Rozanova, Julia et al. “Unequal social engagement for older adults: constraints on choice” [J]. Canadian journal on aging,2012, vol. 31,1: 25-36.
Estabrooks, Paul A et al. “Resources for physical activity participation: does availability and accessibility differ by neighborhood socioeconomic status?” [J]. Annals of behavioral medicine: a publication of the Society of Behavioral Medicine,2012, vol. 25,2: 100-4.
Van Lenthe, F. J., Brug, J., & Mackenbach, J. P. “Neighbourhood inequalities in physical inactivity: the role of neighbourhood attractiveness, proximity to local facilities and safety in the Netherlands” [J]. Social Science & Medicine, 2005,60, 76

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

A Study on the Heterogeneity of Social Engagement among Older Adults: An Empirical Analysis based on the 2018 CLASS Data

Status:

Version 1

Abstract

Background

Methods

Results

Conclusions

Figures

1 Background

2 Literature Review

2.1 Social Engagement among Older Adults

2.2 socioeconomic status and social participation

3 Material and Methods

3.1 Data

3.2 Variables and Measurements

3.3 analysis methods

4 Results

4.1 Descriptive Study Results

4.2 Model Fit Comparison

4.3 Heterogeneity in Social Participation Levels among Different Socioeconomic Status Groups

5 Discussion And Conclusion

Declarations

References

Additional Declarations

Status:

Version 1