Table 2 shows the estimated results of participation in physical activity on the multidimensional health of urban and rural residents. Model 1, Model 3, and Model 5 include independent variables and control variables for logit regression, and Model 2, Model 4, and Model 6 are logit regression results after adding mediating variables.
After including independent and control variables, the results of Model 1 indicate that participation in physical exercise can significantly improve the mental health of urban and rural residents (p = 0.000 < 0.01), and the OR (odds ratio = exp0.159) is 1.172, which means that individuals who participate in physical exercise have a 17.2% higher probability of being mentally healthy than those who do not participate in physical exercise. The results of Model 3 indicate that participation in physical activity significantly improves the physical health of individuals (p = 0.000 < 0.01), and the OR (odds ratio = exp0.239) is 1.270, which means that individuals who participate in physical activity have a 27% higher probability of being physically healthy than those who do not participate in physical activity. The results of Model 5 indicate that participation in physical activity significantly improved the social health of individuals (p = 0.000 < 0.01), and the OR (odds ratio = exp0.305) is 1.357, which means that individuals who participated in physical activity had a 35.7% higher probability of being socially healthy than those who did not participate in physical activity. Thus, Hypothesis 1 was verified. From the regression results of the control variables, gender, marriage, subjective well-being, family size, region, and family income had significant effects on mental health in Model 1. Males, those who were married, those with higher subjective well-being, those in economically underdeveloped regions, and those with higher household income had higher levels of mental health. Moreover, the regression coefficient of family size on individual mental health is negative, indicating that too many family members are more likely to face complex family member relationships, leading to their low level of mental health. In Model 3, gender, household registration, party membership, education level, labor participation status, subjective well-being, medication, region, household income, and age have significant effects on individual physiological health, and marriage and family size do not have significant effects on physiological health. The physiological health level was higher among males, those with nonagricultural household registration, party members, those with high education levels, those with labor participation, those with higher subjective well-being, those in economically underdeveloped regions, and those with higher household income, and the regression coefficients of participation in basic medical insurance and age on physiological health were negative. In Model 5, gender, household registration, party members, education level, subjective well-being, family size, region, and household income had significant effects on individual social health, and marriage, labor participation status, medication, and age did not have significant effects on social health. Female sex, agricultural household registration, party member status, low education, higher subjective well-being, larger family size, economically underdeveloped regions, and higher individual income indicated higher levels of social health.
Table 2
Estimated results of participation in physical activity on multidimensional health of urban and rural residents
| mental health | mental health | physiology health | physiology health | social health | social health |
| Model 1 | model 2 | model 3 | model 4 | model 5 | model 6 |
Participation in physical activity | 0.159*** | 0.141*** | 0.239*** | 0.215*** | 0.305*** | 0.295*** |
| (0.0467) | (0.0469) | (0.0518) | (0.0521) | (0.0495) | (0.0496) |
Gender | 0.294*** | 0.308*** | 0.257*** | 0.277*** | -0.194*** | -0.188*** |
| (0.0438) | (0.0440) | (0.0487) | (0.0489) | (0.0467) | (0.0467) |
Household registration | 0.0262 | 0.0178 | 0.244*** | 0.235*** | -0.351*** | -0.356*** |
| (0.0545) | (0.0547) | (0.0614) | (0.0616) | (0.0587) | (0.0587) |
Marriage | 0.118** | 0.107** | -0.0722 | -0.0865 | -0.0635 | -0.0697 |
| (0.0521) | (0.0523) | (0.0593) | (0.0595) | (0.0562) | (0.0563) |
Party Members | 0.120 | 0.0985 | 0.238*** | 0.203** | 0.134* | 0.125 |
| (0.0766) | (0.0768) | (0.0858) | (0.0861) | (0.0764) | (0.0765) |
Education level | 0.0205 | 0.0126 | 0.0777*** | 0.0650** | -0.104*** | -0.108*** |
| (0.0242) | (0.0243) | (0.0278) | (0.0280) | (0.0257) | (0.0257) |
Labour participation status | 0.0528 | 0.0428 | 0.354*** | 0.343*** | -0.0155 | -0.0208 |
| (0.0482) | (0.0483) | (0.0530) | (0.0532) | (0.0517) | (0.0517) |
Subjective well-being | 0.600*** | 0.560*** | 0.394*** | 0.335*** | 0.234*** | 0.212*** |
| (0.0266) | (0.0271) | (0.0275) | (0.0283) | (0.0291) | (0.0298) |
Family size | -0.107** | -0.109** | 0.0108 | 0.0122 | 0.0836* | 0.0836* |
| (0.0455) | (0.0456) | (0.0509) | (0.0512) | (0.0482) | (0.0482) |
Medicare | -0.0252 | -0.0411 | -0.287*** | -0.315*** | 0.108 | 0.0999 |
| (0.0812) | (0.0814) | (0.0937) | (0.0942) | (0.0897) | (0.0898) |
Area | 0.219*** | 0.223*** | 0.285*** | 0.289*** | 0.108*** | 0.109*** |
| (0.0290) | (0.0291) | (0.0323) | (0.0324) | (0.0316) | (0.0316) |
Household income | 0.186*** | 0.152*** | 0.267*** | 0.222*** | 0.0623*** | 0.0439* |
| (0.0221) | (0.0226) | (0.0241) | (0.0246) | (0.0234) | (0.0240) |
Age | -0.00146 | -0.00217 | -0.0385*** | -0.0398*** | 0.00132 | 0.000923 |
| (0.00173) | (0.00173) | (0.00197) | (0.00199) | (0.00184) | (0.00185) |
SES | | 0.193*** | | 0.273*** | | 0.101*** |
| | (0.0270) | | (0.0297) | | (0.0284) |
Prob > chi2 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
Pseudo R2 | 0.0832 | 0.0868 | 0.1699 | 0.1762 | 0.0180 | 0.0190 |
N | 10879 | 10879 | 10879 | 10879 | 10879 | 10879 |
Standard errors in parentheses |
* p < 0.1, ** p < 0.05, *** p < 0.01 |
After adding the mediating variable socioeconomic status, the regression results of Model 2, Model 4 and Model 6 show that participation in physical exercise still has a significant effect on individuals' mental health, physical health, and social health, but the regression coefficients gradually decrease, and the regression coefficients of the mediating variables are all positive, indicating that socioeconomic status has a significant role in promoting individuals' multidimensional health. Further analysis concluded that the promotion effect of participation in physical exercise on individuals' multidimensional health may be achieved by improving individuals' socioeconomic status.
Heterogeneity Test
The regression results are shown in Table 3-Table 5 to further test the promotion effects of physical activity participation on the mental health, physical health, and social health of individuals with different characteristics.
Table 3
Heterogeneity test results: effect of participation in physical activity on individual mental health
| mental health |
| Non-rural | rural | Divorce or else | married | Female | Male | West | central | east |
Participation in physical activity | 0.0898 | 0.205*** | 0.190* | 0.143*** | 0.145** | 0.183*** | 0.00663 | 0.214** | 0.184*** |
| (0.0657) | (0.0670) | (0.0979) | (0.0535) | (0.0631) | (0.0698) | (0.0958) | (0.0853) | (0.0698) |
Controlled variables | Control | Control | Control | Control | Control | Control | Control | Control | Control |
Prob > chi2 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
Pseudo R2 | 0.0696 | 0.0814 | 0.0845 | 0.0825 | 0.0858 | 0.0759 | 0.0611 | 0.0892 | 0.0598 |
N | 5849 | 5030 | 2437 | 8442 | 5682 | 5197 | 2570 | 3448 | 4861 |
Standard errors in parentheses |
* p < 0.1, ** p < 0.05, *** p < 0.01 |
Table 4
Heterogeneity test results: effect of participation in physical activity on individual physical health
| physiology health |
| Non-rural | rural | Divorce or else | married | Female | Male | West | central | east |
Participation in physical activity | 0.447*** | 0.0299 | 0.148 | 0.274*** | 0.238*** | 0.253*** | 0.0707 | 0.191** | 0.359*** |
| (0.0749) | (0.0717) | (0.112) | (0.0588) | (0.0698) | (0.0777) | (0.104) | (0.0934) | (0.0790) |
Controlled variables | Control | Control | Control | Control | Control | Control | Control | Control | Control |
Prob > chi2 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
Pseudo R2 | 0.1415 | 0.1675 | 0.2123 | 0.1593 | 0.1724 | 0.1653 | 0.1383 | 0.1913 | 0.1270 |
N | 5849 | 5030 | 2437 | 8442 | 5682 | 5197 | 2570 | 3448 | 4861 |
Standard errors in parentheses |
** p < 0.05, *** p < 0.01 |
Table 5
Heterogeneity test results: effect of participation in physical activity on individual social health
| social health |
| Non-rural | rural | Divorce or else | married | Female | Male | West | central | east |
Participation in physical activity | 0.391*** | 0.234*** | 0.400*** | 0.297*** | 0.358*** | 0.233*** | 0.334*** | 0.106 | 0.440*** |
| (0.0739) | (0.0679) | (0.108) | (0.0560) | (0.0663) | (0.0747) | (0.110) | (0.0892) | (0.0724) |
Controlled variables | Control | Control | Control | Control | Control | Control | Control | Control | Control |
Prob > chi2 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
Pseudo R2 | 0.0219 | 0.0189 | 0.0306 | 0.0168 | 0.0202 | 0.0146 | 0.0213 | 0.0171 | 0.0235 |
N | 5849 | 5030 | 2437 | 8442 | 5682 | 5197 | 2570 | 3448 | 4861 |
Standard errors in parentheses |
*** p < 0.01 |
The results in Table 3 show that participation in physical exercise significantly improved the psychological health of individuals in agricultural households, unmarried individuals, married individuals, females, males, individuals in the central region, and individuals in the eastern region, but the enhancement in the psychological health of individuals in nonagricultural households and individuals in the western region was not significant. With the development of the urban economy, individuals in nonagricultural households can more easily choose professional counseling or other ways to relieve psychological stress and solve psychological problems. The promotion of psychological health did not differ by marriage status or sex, but comparing the regression coefficients of the two groups, we find that participating in physical exercise promotes psychological health more strongly among unmarried individuals than among married individuals. Similarly, physical exercise promotes mental health more strongly in males than in females, and it significantly improves the mental health of individuals in the central and eastern regions, but this effect is not significant among individuals in the western region.
The results in Table 4 show that participation in physical exercise significantly improved the physiological health of individuals in nonagricultural households, married individuals, male and female individuals, and individuals in the central and eastern regions, but the physiological health of individuals in agricultural households, unmarried individuals, and individuals in the western region was not significantly affected. At this stage, farming remains the main way for most rural individuals to participate in productive life, and the effect of participation in physical exercise on their physiological health is not significant because they often participate in physical labor. Participation in physical exercise can significantly improve the physiological health of both males and females, but the regression coefficient shows that the promotion effect on physiological health is more significant for males.
The results in Table 5 show that participation in physical exercise significantly improved the social health of individuals in nonagricultural households, individuals in agricultural households, unmarried individuals, married individuals, females, males, individuals in the western region, and individuals in the eastern region and was not significant for the promotion of social health of individuals in the central region.
Mediation Effect Test
Table 6
Tests for mediating effects of socioeconomic status
| SES model 1 | mental health model 2 | mental health model 3 | physiology health model 4 | physiology health model 5 | social health model 6 | social health model 7 |
Participation in physical activity | 0.221*** | 0.159*** | 0.141*** | 0.239*** | 0.215*** | 0.305*** | 0.295*** |
| (0.0398) | (0.0467) | (0.0469) | (0.0518) | (0.0521) | (0.0495) | (0.0496) |
SES | | | 0.193*** | | 0.273*** | | 0.101*** |
| | | (0.0270) | | (0.0297) | | (0.0284) |
Controlled variables | Control | Control | Control | Control | Control | Control | Control |
Prob > chi2 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
Pseudo R2 | 0.0838 | 0.0832 | 0.0868 | 0.1699 | 0.1762 | 0.0180 | 0.0190 |
N | 10879 | 10879 | 10879 | 10879 | 10879 | 10879 | 10879 |
Standard errors in parentheses |
*** p < 0.01 |
Table 6 shows the mediating effect test model established to test whether socioeconomic status plays a mediating role in participation in physical activity to promote the mental health, physical health and social health of individuals. The regression results in the table show that the regression coefficient β1 of participation in physical exercise on socioeconomic status was 0.2221 and was significant at the 1% confidence interval. This finding indicates that participation in physical exercise can significantly improve the socioeconomic status of individuals. The results of Model 3, Model 5, and Model 7 show that δ1 and δ2 are significant in all three models after the inclusion of socioeconomic status as a mediating variable. In Model 3, Model 5 and Model 7, the regression coefficients of physical activity participation and socioeconomic status on mental health were 0.141 and 0.193, respectively; the regression coefficients of physical activity participation and socioeconomic status on physical health were 0.215 and 0.273, respectively; and the regression coefficients of physical activity participation and socioeconomic status on social health were 0.295 and 0.101. All three regression results are significant at the 1% confidence interval, and the regression coefficients of 1 physical exercise in Model 3, Model 5 and Model 7 are smaller than those in Model 2, Model 4 and Model 6, respectively, that is, δ1 < α1, indicating that participation in physical exercise improves individuals' multidimensional health by increasing their socioeconomic status. Socioeconomic status partially mediates participation in physical exercise to improve individual multidimensional health, and Hypothesis 2 is verified. The total effects of participation in physical exercise to improve individual mental health level, physical health level, and social health level were 0.159, 0.239, and 0.305, respectively, and the mediating effects of socioeconomic status were 0.043 (0.221x0.193), 0.060 (0.221x0.273), and 0.022 (0.221x0.101), respectively. Physical exercise improved individual mental health by 26.8% (0.04265/0.159), individual physical health by 25.2% (0.0603/0.239), and individual social health by 7.3% (0.0223/0.305), and all three effects were mediated by improvements in socioeconomic status. The analysis of the above results shows that physical exercise can not only promote individuals' physical and psychological health but also improve individuals' social health by increasing social interaction and can play a role in improving individuals' socioeconomic status.
Robustness test
To ensure the robustness of the research results, this paper draws on the research results of existing scholars and recodes the indicators related to the independent variable. (“never” coded as 1, “several times a year or less” coded as 2, “several times a month” coded as 3, “several times a week” coded as 4, and “everyday” coded as 5). The logit model was used again to include independent variables, control variables, and mediating variables, and the results are shown in Table 7.
Table 7
Results of robustness tests-the effects of physical activity participation on the multidimensional health of urban and rural residents
| mental health model 1 | mental health model 2 | physiology health model 3 | physiology health model 4 | social health model 5 | social health model 6 |
Participation in physical activity | 0.0684*** | 0.0628*** | 0.101*** | 0.0940*** | 0.110*** | 0.107*** |
| (0.0147) | (0.0148) | (0.0160) | (0.0160) | (0.0153) | (0.0154) |
Controlled variables | Control | Control | Control | Control | Control | Control |
SES | | 0.192*** | | 0.270*** | | 0.100*** |
| | (0.0270) | | (0.0297) | | (0.0285) |
Prob > chi2 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
Pseudo R2 | 0.0839 | 0.0875 | 0.1713 | 0.1775 | 0.0191 | 0.0201 |
N | 10879 | 10879 | 10879 | 10879 | 10879 | 10879 |
Standard errors in parentheses |
*** p < 0.01 |
The results of the robustness test indicate that participation in physical exercise can significantly improve individuals' mental health level, physical health level and social health level, and the results of the robustness test of the mediating effect also indicate that participation in physical exercise has a significant positive effect on socioeconomic status, and socioeconomic status plays a mediating role in physical exercise to improve individuals' multidimensional health level, with a partial mediating effect. The results of the robustness test proved that the results of the previous empirical analysis were reliable.