Robust
The above OLS regression results show that the influence of social participation on the physical health, mental health and self-rated health of empty nesters has passed the significant test. However, whether empty nesters have social participation or not may be their own choice, or they may be influenced by other factors, and the estimation coefficient may be biased by self-selection. Based on this, this study chooses the tendency score matching method to deal with the endogenous problems caused by self-selection bias. In order to ensure the robustness of the estimation results, this study selects four methods: K nearest neighbor matching, caliper nearest neighbor matching, kernel matching and radius matching.
From Table 8, it can be seen that covariate deviation of empty nesters with social participation (treatment group) and empty nesters without social participation (control group) is greatly reduced after matching compared with that before matching, basically meeting the balance requirement of tendency score matching, and self-selection deviation is largely eliminated. Table 9 shows the estimated results under different methods. The average treatment effect (ATT) of physical health, mental health and self-rated health of empty nesters is greater than 0 and significant. This shows that when the external factors of the empty nesters are similar to those of the control group, the health status of the empty nesters with social participation will be obviously improved. Hypothesis 1 is verified again. According to the regression results of PSM, compared with empty nesters without social participation, empty nesters with social participation have a 0.57 ~ 0.68 percentage point higher probability of physical health, a 0.59 ~ 0.62 percentage point higher probability of mental health and a 0.62 ~ 0.65 percentage point higher probability of self-rated health. After controlling the endogenous problems caused by sample self-selection bias, the influence of social participation on the health status of empty nesters has changed slightly, but it still has a significant impact.
Table 8
Variable | Processing | Control | Standard Deviation (%) | Deviation Deduction (%) | T value | P value |
Gender | Front Matching | 0.507 | 0.554 | -9.4 | 63.6 | -3.19 | 0.001 |
After Matching | 0.508 | 0.491 | 3.4 | 1.17 | 0.242 |
Age | Front Matching | 68.699 | 68.309 | -9.3 | 80.6 | -3.16 | 0.002 |
After Matching | 68.691 | 68.573 | 1.8 | 0.64 | 0.523 |
Married | Front Matching | 0.708 | 0.731 | -5.2 | 77.9 | -1.75 | 0.081 |
After Matching | 0.709 | 0.714 | -1.1 | -0.39 | 0.699 |
Education | Front Matching | 1.524 | 1.309 | 28.8 | 83.1 | 9.72 | 0.000 |
After Matching | 1.518 | 1.482 | 4.9 | 1.53 | 0.127 |
Residence | Front Matching | 0.313 | 0.181 | 31.0 | 93.2 | 10.44 | 0.000 |
After Matching | 0.311 | 0.320 | -2.1 | -0.66 | 0.509 |
Number of Children | Front Matching | 0.022 | 0.014 | 3.7 | -9.4 | 1.24 | 0.216 |
After Matching | 0.021 | 0.013 | 4.0 | 1.48 | 0.139 |
Chronic Disease | Front Matching | 0.504 | 0.456 | 9.7 | 64.0 | 3.29 | 0.001 |
After Matching | 0.504 | 0.521 | -3.5 | -1.20 | 0.231 |
Religious | Front Matching | 0.121 | 0.100 | 6.6 | 67.0 | 2.22 | 0.026 |
After Matching | 0.121 | 0.128 | -2.2 | -0.71 | 0.479 |
Ethnicity | Front Matching | 0.930 | 0.931 | -0.4 | -1279.5 | -0.12 | 0.901 |
After Matching | 0.930 | 0.943 | -5.1 | -1.80 | 0.072 |
Pension Insurance | Front Matching | 0.287 | 0.155 | 32.2 | | 10.85 | 0.000 |
After Matching | 0.286 | 0.283 | 0.7 | 97.7 | 0.23 | 0.821 |
Medical Insurance | Front Matching | 0.979 | 0.969 | 5.8 | 95.3 | 1.97 | 0.049 |
After Matching | 0.979 | 0.978 | 0.3 | 0.10 | 0.919 |
Table 9
Propensity Score Matching Results
Matching Method | Average Treatment Effect | Bootstrap Standard Error | T value |
Physical Health | K nearest Neighbor Matching | 0.057 | 0.014 | 3.85*** |
Radius Matching | 0.068 | 0.012 | 5.59*** |
Closest Neighbor Matching in Caliper | 0.068 | 0.012 | 5.59*** |
Nuclear matching | 0.068 | 0.012 | 5.64*** |
Mental Health | K nearest Neighbor Matching | -0.062 | 0.019 | -3.17*** |
Radius Matching | -0.059 | 0.016 | -3.67*** |
Closest Neighbor Matching in Caliper | -0.059 | 0.016 | -3.67*** |
Nuclear matching | -0.059 | 0.015 | -3.73*** |
Self-Health | K nearest Neighbor Matching | 0.065 | 0.016 | 3.85*** |
Radius Matching | 0.063 | 0.013 | 4.63*** |
Closest Neighbor Matching in Caliper | 0.063 | 0.013 | 4.63*** |
Nuclear matching | 0.062 | 0.013 | 4.58*** |
Note: The control variables are consistent with Table 2. The same below.
Heterogeneity Test
The above analysis results show that social participation can improve the physical health, mental health and self-rated health level of empty nesters. However, due to the differences in life expectancy of Chinese residents due to various reasons and the differences between urban and rural areas caused by the dual structure of urban and rural areas, the impact of social participation on the health status of empty nesters will be different. This part mainly analyzes the heterogeneity of the impact of social participation on the health of empty nesters from the perspectives of age and urban and rural areas.
As can be seen from Table 10, social participation has a significant impact on the physical health, mental health and self-rated health of empty nesters of all ages at 1% statistical level. In terms of physical health, social participation has the greatest influence on the empty nesters aged 80 and above. For every unit of social participation of empty nesters aged 80 and above, their physical health increases by 1.385 units. In terms of mental health, social participation has the greatest influence on the 70-79-year-old empty nesters. For every unit of social participation of 70-79-year-old empty nesters, their mental health rises by 1.622 units. In terms of self-rated health, social participation has the greatest influence on the 70-79-year-old empty nesters. For every unit of social participation of 70-79-year-old empty nesters, their self-rated health increases by 0.228 units. To sum up, social participation has the greatest impact on the health of empty nesters aged 70–79. It can be seen from Table 11 that social participation has a significant impact on empty nesters living in both urban and rural areas. According to the estimated coefficient of social participation in the table, it can be seen that the possibility of social participation in improving the health level of urban empty-nesters is greater than that of rural empty-nesters. This may be because: the proportion of social participation of urban empty-nesters is higher, and a good atmosphere can infect more urban empty-nesters to participate in various social activities, and they are more likely to get more friends, which is conducive to improving the health level of urban empty-nesters; Compared with rural empty nesters, empty nesters in cities have richer and more detailed conditions for social participation, as well as various kinds of participating activities. They can choose their favorite activities, such as dancing, learning painting, taking tutoring classes, etc., so as to improve their health.
Table 10
Effect of Social Participation on The Health Status of Empty Nesters in Different Age Groups
Variable | 60–69 Years Old | 70–79 Years Old | 80 Years Old and Above |
Physical Health | Mental Health | Self-Health | Physical Health | Mental Health | Self-Health | Physical Health | Mental Health | Self-Health |
Social Participation | 0.391*** (0.050) | -1.264*** (0.188) | 0.126*** (0.027) | 0.663*** (0.091) | -1.622*** (0.283) | 0.228*** (0.037) | 1.385*** (0.183) | -1.194** (0.511) | 0.147** (0.063) |
Constant Term | 19.118*** (0.035) | 19.330*** (0.136) | 2.908*** 90.019) | 18.617*** (0.060) | 19.779*** (0.196) | 2.727*** (0.025) | 17.630*** (0.111) | 19.237*** (0.337) | 2.818*** (0.041) |
Control Variable | YES | YES | YES | YES | YES | YES | YES | YES | YES |
R2 | 0.0133 | 0.0090 | 0.0038 | 0.0200 | 0.0144 | 0.0132 | 0.0434 | 0.0085 | 0.0053 |
Table 11
Effect of Social Participation on Health Status of Empty Nesters in Urban and Rural Areas
Variable | City | Village |
Physical Health | Mental Health | Self-Health | Physical Health | Mental Health | Self-Health |
Social Participation | 1.035*** (0.100) | -1.930*** (0.285) | 0.259*** (0.042) | 0.565*** (0.056) | -0.820*** (0.175) | 0.119*** (0.023) |
Constant Term | 18.545*** (0.075) | 18.426*** (0.228) | 2.877*** (0.033) | 18.731*** (0.036) | 19.663*** (0.118) | 2.834*** (0.015) |
Control Variable | YES | YES | YES | YES | YES | YES |
R2 | 0.0534 | 0.0225 | 0.0168 | 0.0156 | 0.0037 | 0.0034 |
This study investigated the relationship between social participation and physical health, mental health, and self-rated health among Chinese empty nesters. The results showed that, first, there were significant positive effects of social participation on physical health, psychological health, and self-rated health among the empty nesters, and the various health effects held after the endogenous problems caused by self-selection of the sample were treated by propensity score matching. The social participation of empty nesters should not be limited to behavioral changes, rather, the increased awareness of social participation of empty nesters can lead to a "win-win" situation, i.e., the realization of active aging and health strategies.
Social participation, as an important element of the concept of active aging, can encourage empty nesters to get out of their homes and participate in various social activities to help them develop good living and exercise habits and reduce the likelihood of depression and suicide[21]. It is evident that social participation plays a very important role in the health of empty nesters. With the increasing diversification of modern means of communication and exchange, the forms of social participation are also diversified, and there is a negative correlation between the degree of social participation and loneliness among empty nesters, i.e., the higher the frequency of participation, the lower the loneliness of empty nesters[22]. Social participation also helps empty nesters to improve their physical fitness, maintain health and prevent diseases, and the health status of empty nesters who actively participate in sports and fitness is better than those who do not regularly participate in sports and fitness[23]. Also learning participation can promote empty nesters' subjective well-being because they can feel the care and love of teachers or classmates during the learning process, and it can achieve the re-socialization of empty nesters, which in turn can improve their subjective well-being[9]. In contrast, socially engaged empty nesters are more likely to have good health status, and therefore are encouraged to take the initiative to get out of the house and actively participate in social activities.
In terms of control variables, the three variables of gender, age, and chronic illness passed significance tests for physical health, mental health, and self-rated health. The results for the gender variable showed that male empty nesters had better physical health, mental health, and self-rated health than female empty nesters. This may be due to two reasons: on the one hand, male empty nesters have less family care responsibilities and less work discrimination pressure compared to female empty nesters, combined with the relatively higher ability of male empty nesters to alleviate bad emotions and the fact that males are more physically active compared to females, resulting in better health status of male empty nesters compared to female empty nesters[24]; on the other hand On the other hand, it may validate the "health-survival" paradox proposed by Archer that men are in better health than women at older ages[25]. The results of the age variable show that age has a negative effect on all types of health, i.e., the physical health, mental health, and self-rated health of empty nesters decreases with age. As age increases, empty nesters' health capital depletion begins to accelerate, and their ability and energy to participate in social activities gradually decreases. Instead of being positively influenced, participation in social activities among elderly empty nesters may be counterproductive and induce the emergence of various hidden symptom problems[26–27]. The results of chronic disease variables show that having a chronic disease leads to poorer physical health, mental health, and self-rated health among empty nesters[28]. Generally speaking, empty nesters with illnesses try to avoid participating in time-consuming and labor-intensive activities to prevent adverse effects on their health, while not participating in social activities makes it difficult to improve the health status of empty nesters. Marriage and ethnicity variables showed a significant contribution to the physical health of the empty nesters. Marriage, literacy, residence, religion and pension insurance showed significant negative effects on the mental health of empty nesters, i.e. the better or higher the above variables, the better the mental health of empty nesters. Literacy, place of residence and medical insurance had a significant positive effect on the self-rated health of empty nesters, and chronic diseases had a negative effect on them.
This study also looked at both heterogeneity and mechanistic effects in order to assess the impact of social engagement on the health of empty nesters. The results showed that there was age and urban-rural heterogeneity in the effect of social participation on the health of empty nesters, i.e., social participation had a greater effect on the health of urban empty nesters; social participation had the greatest effect on the mental health of empty nesters aged 70–79 years, social participation had the greatest effect on the physical health of empty nesters aged 80 years and older, and social participation had the greatest effect on the health of 70 The impact of social participation on the self-rated health of empty nesters aged 70–79 was the greatest. Economic income and cognitive ability had a significant mediating role in the influence of social participation on the health of empty nesters.
There are some limitations of this paper. First, there is a causal relationship between social participation and the health of empty nesters, i.e., empty nesters with poorer health status are less likely to have social participation, and this finding has been demonstrated in studies; second, the health status and social participation of empty nesters are based on assessment scales, which may be biased to some extent.