Air Pollution, Physical Exercise, and Physical Health: An Analysis Based on Data from the China General Social Survey

Abstract

This study aims to investigate the influence of air pollution and physical exercise on residents’ health using data from the Chinese General Social Survey (CGSS) 2018. The research is based on the Grossman Health Production Function and employs the ordered logit model to analyze the relationship between air pollution, physical exercise, and residents’ health. We find that among the Chinese population aged 18 and above, air pollution negatively affects residents’ physical health, whereas physical exercise has a significantly positive effect. Introducing gender-stratified and urban–rural stratified models, the research reveals that the negative effects of air pollution on residents’ health vary among different groups. On one hand, men and residents living in urban areas will bear heavier health damages due to air pollution. On the other hand, physical exercise positively affect the health of both men and women, whether in rural or urban areas. Additionally, age and higher income are associated with better physical health, whereas education does not show a significant impact. Based on these findings, a series of policy recommendations have been proposed to comprehensively and systematically promote ecological governance and sustainable development. The government should strengthen environmental regulations, reduce residents’ exposure to air pollution risks, improve the equality of basic public services, invest in rural sports infrastructure, promote healthy lifestyles, and provide personalized health guidance through health education and mobile health services. Moreover, when formulating policies, the government should consider the specific needs of different groups to ensure the fairness and sustainability of the policies.

1. Introduction

We are in an era of rapid development, where technological advancements have brought unprecedented convenience and comfort. However, with the acceleration of urbanization, factors such as traffic emissions and industrial pollutants have continuously increased the concentrations of harmful substances in the air. This directly impacts the quality of life and physical health of individuals. Globally, the burden of diseases associated with exposure to air pollution poses significant losses to health. According to estimates by the WHO, annual exposure to air pollution leads to millions of premature deaths and health losses. The current estimates indicate that the impact of air pollution on disease burden rivals those of other prominent global health risks, such as poor diet and smoking, thereby positioning it as one of the foremost environmental threats to human health. The severe impact of air pollution on human life and longevity underscores the urgent need to take action to mitigate this global health threat, prompting individuals to reconsider their lifestyles and explore better ways to protect their health.

In 2018, the WHO released a report stating that air pollution levels are very severe in many parts of the world, causing 7 million deaths annually due to air pollution [1]. And it is projected to become the seventh leading cause of global mortality in the next 20 years. Luechinger (2009) [2] points out that air pollution poses a serious threat to health, and a healthy natural ecosystem can significantly alleviate negative emotions such as anxiety, despair, and depression among residents. There is an urgent need to adopt sustainable development measures to maintain natural balance. Warburton (2016) [3] suggests that there may be a link between air pollution and diseases such as diabetes and neurodegenerative disorders. Yang (2018) [4] highlights that worsening air pollution also imposes significant health costs on residents, leading to rapidly escalating medical expenses and an increasing economic burden from diseases. In March 2021, China experienced the most intense dust storm in nearly a decade. The environmental damage caused by air pollution poses significant health risks to the well-being of Chinese residents. Therefore, it is essential to take note of how air pollution affects the health of residents due to its practical significance.

Air pollution is closely related to various health issues, including respiratory and cardiovascular diseases, prompting individuals to reconsider their lifestyles and explore better ways to protect their health. Meanwhile, as a crucial means of maintaining physical health, physical exercise receives considerable attention. Whether for sports enthusiasts or the general public, physical exercise is regarded as an effective way to improve cardiovascular function, enhance immunity, and alleviate stress. Exercise has been proven to be beneficial to health by experimental research and epidemiological surveys. Giles et al. (2014) [5] assert that long-term regular exercise can counteract the bodily harm caused by air pollution, indicating that the positive effects of exercise on health may outweigh the negative effects of air pollution. Warburton (2016) [3] and others found that regular and systematic physical exercise has enormous health benefits, with almost everyone gaining advantages from active participation in physical activities. Regular physical exercise can reduce the risk of premature death and over 25 chronic diseases by 20–30%, with the potential for an even more significant reduction of up to 50%. Engaging in physical exercise has become one of the healthy lifestyle choices.

Even though scholars have delved deeply into the connection between air pollution, physical exercise, and physical health, there remain certain gaps that warrant additional exploration of their inter-relationships. Firstly, existing empirical evidence regarding the impact of air pollution, physical exercise, and physical health primarily comes from developed countries, with limited evidence from developing nations. Secondly, most existing studies on the correlation among air pollution, physical exercise, and physical health focus on pairwise associations, and relatively less research integrates all three factors. Thirdly, there is heterogeneity in how air pollution and physical exercise affect residents’ health. However, studies in this regard are scarce. While China is speeding up the construction of a beautiful and healthy country, improving air quality is considered one of its essential tasks, and it has achieved particular accomplishments. Therefore, based on the Chinese General Social Survey (CGSS 2018) data, our research investigates how air pollution and physical exercise affect residents’ health, exploring regional and gender variations in their impact. This research aims to provide valuable insights into how air pollution and physical exercise affect resident health, facilitating the scientific prevention and mitigation of health risks caused by air pollution.

Our research makes the following key contributions: Firstly, we extend previous scholars’ related research, establishing an intimate connection between air pollution, physical exercise, and their impacts on resident health. Secondly, this study reveals that the effects of air pollution and physical exercise on health are interactive processes, with variations in the effects of physical exercise depending on the severity of air pollution’s impacts on health. Thirdly, under the combined influence of air pollution and physical exercise, residents’ health is heterogeneous across different regions and genders. Fourthly, current research by Chinese scholars on the link between air pollution, physical exercise, and physical health relies heavily on specific small-scale survey samples. We use data from the Chinese General Social Survey (CGSS 2018), which is reliable and representative. Our findings offer empirical support for the Chinese government’s initiatives in advancing the development of a beautiful and healthy nation.

The structure of the paper is as follows. Section 2 presents a literature review, focusing on air pollution and physical health, as well as sport exercise and physical health, and puts forward research hypotheses. Section 3 describes the source of the survey data, the design of variables, and the methods. Section 4 introduces and analyzes specific empirical results. Section 5 is the discussion section. Section 6 presents the conclusions and recommendations.

2. Literature Review

2.1. Air Pollution and Physical Health

Mainstream research on the impact of air pollution on health is guided by the health production function introduced by Grossman (1972) [6]. Grossman pioneered the concept of health as both a consumable good that provides utility and an investment that generates income. He incorporated health into the consumer utility function, suggesting that the health stock improves with increased exercise frequency and good dietary habits. However, the health stock decreases with the growth of the health depreciation rate, where age plays a significant role in influencing this rate. Based on this, he explored the optimal health demand quantity for consumers to maximize utility. Cropper (1981) [7] introduced air pollution variables within the framework of Grossman’s theoretical model, emphasizing air pollution as a significant factor influencing the health depreciation rate. This study developed an analytical framework to assess the potential effects of air pollution on health, providing solid theoretical support for understanding how air pollution affects health. Subsequently, researchers like Gerking and Stanley (1986) [8] incorporated environmental factors into the health production function, delving into the influence of these factors on the rate of health depreciation. This expanded our understanding of ecological variables in health economics and highlighted the undeniable connection between the environment and health.

Air pollution is one of the primary sources of environmental pollution, and research on the health effects of air pollution is expanding into fields such as health economics, epidemiology, and environmental sociology. Currie et al. (2009) [9] tested the relationship between environmental pollution and infant mortality based on fixed-effects models using pollution data from New Jersey and the individual data of local residents. They found that carbon monoxide concentration significantly negatively impacts infant mortality during pregnancy and post-birth periods. Studies by scholars such as Nevalainen (1998) [10], Correia (2013) [11], and Pascal (2014) [12] indicate a significant negative correlation between different types of air pollution and life expectancy. As air pollution levels rise, adverse effects on life expectancy and health conditions become evident. Cromar and Gladson (2024) [13] analyzed data from the U.S. Environmental Protection Agency for 2018–2020, revealing that air pollution leads to health problems nationwide, including preterm birth, low birth weight, preventable deaths, lung cancer, cardiovascular and respiratory diseases, and affected days. The study also suggests that outdoor wildfires are a crucial factor directly impacting health due to air pollution. Citizens should adhere to healthier air quality standards for overall physical health. Zhu, J et al. (2023) [14] utilized data from the China General Social Survey (CGSS) 2017, conducting an analysis on 4132 valid samples across 28 provincial-level administrative regions in China. The research uncovered a noteworthy adverse effect of perceived air pollution on individuals’ self-assessed health.

Furthermore, scholars have highlighted heterogeneity in how air pollution affects residents’ health. Miao (2010) [15] and Li (2019) [16] argue that differences in physiological structure, socioeconomic status, and perceived quality of life lead to variations in the impact of air pollution on male and female residents. Research by Li (2018) [17] found significant differences among residents with different income levels in social labor division, environmental risk avoidance ability, and self-health awareness. Lower-income residents typically bear greater losses in health benefits, while the negative impact of air pollution on the health of residents gradually diminishes with increasing income levels. Sun, M. et al. (2019) [18] employed CGSS data to explore how air pollution and socioeconomic status collectively influence residents’ health. Their findings revealed that individuals of higher socioeconomic status exhibit superior self-rated health, and the health disparities between rural residents and urban are primarily influenced by income and social class factors. Air pollution contributes significantly to health inequality among residents. Based on the literature reviewed, air pollution negatively affects human health. In China, relatively few scholars have conducted in-depth studies linking air pollution and health, and their research is mainly descriptive and statistical. Therefore, using empirical data from China to analyze air pollution’s impact on health thoroughly becomes particularly necessary. In this study, self-rated health is the dependent variable, being a commonly used measure in studying residents’ physical health. Based on the literature analysis, we propose the following three research hypotheses:

H1. 

The higher the level of air pollution, the poorer self-rated health.

H2. 

The effects of air pollution on the self-rated health of males and females differ.

H3. 

The effects of air pollution on the self-rated health of urban and rural residents differ.

2.2. Sports Exercise and Physical Health

Engaging in sports activities not only helps with exercising the body and improves physical fitness but also fosters social interactions, contributing to the establishment of positive interpersonal relationships. Undoubtedly, sports offer humanity the most economical, practical, and convenient benefits, and their role in promoting health has been widely recognized globally. The positive relationship between sports exercise and self-rated health has been empirically supported by numerous experts and scholars [19]. Researchers such as Abu-Oma (2004) [20] studied 15 European Union countries, finding a significant positive relationship between sports activity levels and self-rated health. Additionally, self-rated health shows associations with factors like gender, age, school attainment, and income. Leisure-time physical activity was positively correlated with self-rated health [21], while alcohol consumption showed no association with poor self-rated health. Wang Dian’e (2009) [22] and other researchers conducted surveys and analyses on the exercise frequency and duration of 1193 residents in 14 provinces and cities in China, identifying a noteworthy positive connection between engaging in sports activities and satisfaction with health.

In addition to the general population, there is a growing body of research by domestic and international scholars on the health effects of sports exercise for the elderly. Developed countries have started formulating policies such as “Healthy Citizen Plans” and “Public Health” to advocate for elderly sports fitness. Beyer (2015) [23] conducted a six-month experimental study on older people aged 65 to 85 with two or more chronic diseases living in German communities, and the findings demonstrated that higher levels of physical activity significantly improved their self-rated health. Researchers like Gao Liang (2015) [24], through a survey of older Chinese people, found that the duration, frequency, and time of sports exercise significantly influenced the self-rated health scores of older people. Long-term participation, a frequency exceeding five times a week, with each session lasting more than 30 min of sports exercise, was found to have significant benefits for maintaining the physical and mental health of older people. Both domestic and international research results consistently point to the promoting effect of sports exercise on self-rated health and emphasize the importance of the mode of sports exercise.

Through experimental research and epidemiological investigations, the benefits of exercise on health have been confirmed. Regular and long-term exercise can counteract the damage to the body caused by air pollution, with the positive effects of exercise on health potentially outweighing the negative effects of air pollution [25,26]. However, some studies suggest that particulate matter exposure can harm the respiratory systems of exercisers [27,28]. Matt et al.’s research (2016) [29] discovered that, during 2 h of moderate-intensity intermittent exercise (15 min of exercise + 15 min of rest, four sets, 50% to 70% of maximum heart rate) in an environment with air pollution caused by traffic, exercise could delay or reduce the damage to the respiratory system caused by air pollution. Scholars such as Xu (2021) [30] found in their research that in situations where air quality is relatively good, the health benefits of physical activity take precedence. However, in conditions of poor air quality, the health benefits derived from physical activity cannot counterbalance the harm caused by exposure to air pollution, and the more activity undertaken, the higher the health risk. These studies suggest that exercise has a specific protective effect against health damage related to air pollution, and this effect is correlated with the concentration of air pollutants. Based on the literature analysis above, we propose the following hypotheses:

H4. 

The more engaged in physical exercise, the self-rated healthier the physique.

H5. 

The effects of physical exercise on the self-rated health of urban and rural residents differ.

H6. 

The effects of physical exercise on the self-rated health of males and females differ.

H7. 

The effects of physical exercise on the residents’ self-rated health are moderated by air pollution.

3. Materials and Methods

3.1. Data

This study primarily utilized data from CGSS2018. The China General Social Survey (CGSS), initiated in 2003 by Renmin University of China and the Hong Kong University of Science and Technology, is a nationwide, comprehensive, and longitudinal academic survey project. It has become an important source of micro-level research data in China. CGSS2018 commenced its national household visits in early July 2018, employing a multi-stage stratified sampling procedure to cover the adult population. The survey was conducted through face-to-face interviews, and the collected data and materials were cleaned, processed, archived, standardized, and internationalized according to international standards. The survey data were published in 2022. After excluding missing values and extreme outliers for the dependent variable, core independent variables, and control variables, in addition, all questions with options 98 indicating ‘do not know’ and 99 indicating ‘refuse to answer’ are removed. A total of 3238 valid samples were obtained.

3.2. Variable Selection and Assignment

(1)

Dependent Variable: The health level in this study primarily refers to physical health. Data on physical health levels were obtained from CGSS 2018, based on the following relevant question: “In your opinion, how would you rate your current physical health?” Respondents’ answers, namely “Very unhealthy”, “Somewhat unhealthy”, “Average”, “Somewhat healthy”, and “Very healthy”, were sequentially assigned values of 1 to 5, respectively.

(2)

Core Independent Variables: The primary factors under investigation in our research include air pollution (Outdoor) and physical exercise (Outdoor). The objective pollution indicators frequently encompass extensive geographical areas, leading to reduced compatibility with micro-level individual data. Moreover, objective pollution indicators are often correlated with macroeconomic indicators, rendering them prone to the influence of the Environmental Kuznets Curve (EKC) hypothesis. This makes it challenging to isolate negative effects accurately. Therefore, this study employs subjective perception scales from the survey questionnaire to evaluate individuals’ levels of exposure to air pollution. The superiority of this method lies in its ability to more intuitively illustrate the connection between pollution and individuals’ perceptions of health. Combining the questions from the CGSS 2018 survey questionnaire, the option “I feel the air quality in my residential area is very good”, is interpreted as follows: responses of “strongly agree” and “agree” are defined as “no pollution” (assigned a value of 0), while responses of “neutral”, “disagree”, and “strongly disagree” are defined as “polluted” (assigned a value of 1).

For physical exercise, based on the question from the CGSS2018 survey questionnaire, “In the past year, have you often engaged in the following activities in your spare time?—Engaging in physical exercise.”, choices of “every day” and “several times a week” were defined as “participating in physical exercise” (assigned a value of 1), while choices of “several times a month”, “several times a year or less”, and “never” were defined as “not participating in physical exercise” (assigned a value of 0). The characteristic distribution of core variables is presented in

Table 1. Distribution characteristics of core variables.

Variable	Distribution Characteristics
Health	“In your opinion, how would you rate your current physical health?” 1 = Very unhealthy (3.64%); 2 = Somewhat unhealthy (15.01%); 3 = Average (23.47%); 4 = Somewhat healthy (39.99%); 5 = Very healthy (17.88%).
Physical exercise	“In the past year, have you often engaged in the following activities during your leisure time?—Engaging in physical exercise”. 1 = Engaged in physical exercise (38.54%); 0 = Not engaged in physical exercise (61.46%).
Air pollution	“I feel the air quality in my residential area is very good”. 0 = No air pollution (65.53%); 1 = air pollution (34.47%).
Gender	0 = Male (50.28%); 1 = Female (49.72%).
Community type	1 = Urban (70.94%); 0 = Rural (29.06%).
Education	“Years of education”. 0 = Not educated (13.03%); 1 = Private school literacy class (0.77%); 5 = Primary school (21.59%); 9 = Junior high school (28.38%); 12 = High school (vocational school) (19.12%); 15 = College (Associate degree) (7.41%); 16 = Bachelor’s degree (8.65%); 19 = Master’s and above (1.05%).

.

(3)

Control Variables: To avoid endogeneity issues arising from omitted variables, reciprocal causation, and the like, this study incorporates control variables into the mathematical model. These variables include age, age squared, gender (female = 1, male = 0), years of education, community type (urban = 1, rural = 0), and the natural logarithm of personal annual income (derived from the following question in the survey: “What was your total income for the entire year last year (2017)?”).

3.3. Methods

This study utilized Stata 17.0 for data analysis. Stata is a statistical software package developed by StataCorp LLC, College Station, TX, USA, providing a wide range of statistical and data analysis capabilities. Because the dependent variable in this study, residents’ health, was represented by values ranging from 1 to 5, it fell into the category of ordered response variables. The distances between the health values were not comparable, and using ordinary least squares (OLS) regression may lead to some degree of bias in the results. Therefore, this study opted for an ordered logit model for analysis. Unlike typical binary or multinomial discrete choice models, the ordered logit model is suitable when the dependent variable has a particular order or level. In the context of this study, where the dependent variable represents residents’ health with values 1, 2, 3, 4, and 5, the ordered response model was well suited, making it the primary regression method employed.

Utilizing Grossman’s health production function, we constructed a baseline model to identify predictive factors, such as air pollution and physical exercise, for residents’ health. The specific settings are as follows:

$${\mathrm{H}\mathrm{e}\mathrm{a}\mathrm{l}\mathrm{t}\mathrm{h}}_{\mathrm{i}}={\mathsf{\beta }}_1\ast {\mathrm{A}\mathrm{i}\mathrm{r} \mathrm{P}\mathrm{o}\mathrm{l}\mathrm{l}\mathrm{u}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}_{\mathrm{i}}+{\mathsf{\beta }}_2\ast {\mathrm{P}\mathrm{h}\mathrm{y}\mathrm{s}\mathrm{i}\mathrm{c}\mathrm{a}\mathrm{l} \mathrm{e}\mathrm{x}\mathrm{e}\mathrm{r}\mathrm{c}\mathrm{i}\mathrm{s}\mathrm{e}}_{\mathrm{i}}+{\mathsf{\beta }}_3\ast {\mathrm{X}}_{\mathrm{i}}+\mathsf{\epsilon }$$

(1)

where health represents an individual’s self-rated health condition, pollution indicates the air quality in the individual’s residential area, exercise indicates physical activity status, X indicates a series of control variables affecting residents’ health, ε indicates the error term, and β indicates the partial effect parameters of the relevant variables. To further analyze the impacts of air pollution and physical exercise on physical health, the following regression model was further constructed:

$${\mathrm{H}\mathrm{e}\mathrm{a}\mathrm{l}\mathrm{t}\mathrm{h}}_{\mathrm{i}}={\mathsf{\beta }}_1\ast \mathrm{A}\mathrm{i}\mathrm{r} {\mathrm{P}\mathrm{o}\mathrm{l}\mathrm{l}\mathrm{u}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}_{\mathrm{i}}+{\mathsf{\beta }}_2\ast {\mathrm{P}\mathrm{h}\mathrm{y}\mathrm{s}\mathrm{i}\mathrm{c}\mathrm{a}\mathrm{l} \mathrm{e}\mathrm{x}\mathrm{e}\mathrm{r}\mathrm{c}\mathrm{i}\mathrm{s}\mathrm{e}}_{\mathrm{i}}+{\mathsf{\beta }}_3{\mathrm{A}\mathrm{i}\mathrm{r} \mathrm{P}\mathrm{o}\mathrm{l}\mathrm{l}\mathrm{u}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}_{\mathrm{i}}\ast {\mathrm{P}\mathrm{h}\mathrm{y}\mathrm{s}\mathrm{i}\mathrm{c}\mathrm{a}\mathrm{l} \mathrm{e}\mathrm{x}\mathrm{e}\mathrm{r}\mathrm{c}\mathrm{i}\mathrm{s}\mathrm{e}}_{\mathrm{i}}+{\mathsf{\beta }}_4\ast {\mathrm{X}}_{\mathrm{i}}+\mathsf{\epsilon }$$

(2)

4. Results

4.1. Descriptive Analysis

Table 1 and

Table 2. Descriptive statistics of core variables.

Variable	Obs	Mean	SD	Min	Max
Health	3238	3.535	1.061	1	5
Physical exercise	3238	0.385	0.487	0	1
Air pollution	3238	0.345	0.475	0	1
Gender	3238	0.497	0.5	0	1
Age	3238	52.537	16.132	18	94
Age squared	3238	3020.293	1712.51	324	8836
Annual income	3238	9.912	1.344	4.605	16.056
Community type	3238	0.709	0.454	0	1
Education	3238	8.63	4.904	0	19

Note. The annual income in the table represents values that have undergone logarithmic transformation.

present the model’s statistical description and distribution of variables, respectively. Upon observing the distribution range and characteristics of the entire sample of 3238 respondents, it is evident that 17.88% and 39.99% of individuals perceive themselves as “very healthy” and “somewhat healthy”, comprising more than half of the sample. The proportion of those feeling “average” is 23.47%, while the percentages for “somewhat unhealthy” and “very unhealthy” are 15.01% and 3.64%, respectively, indicating a relatively substantial presence of such responses. This variation may stem from the individual heterogeneity of respondents and the differing macroeconomic variables across regions. The overall mean of residents’ health for the entire sample is 3.535, falling between “average” (health = 3) and “Somewhat healthy” (health = 4), aligning with similar findings in other surveys on residents’ health.

Additionally, Table 1 informs us that over the past year, 38.54% of respondents engaged in physical exercise, while 61.46% did not; moreover, 34.47% felt there was air pollution, whereas 65.53% did not perceive any air pollution. Males accounted for 50.28% of the sample, and females 49.72%; urban residents made up 70.94%, while rural residents comprised 29.06%. Those with elementary education or less represented 35.39%, those with a secondary education accounted for 47.5%, and those with a university degree or higher made up 17.11%.

4.2. Baseline Model Analysis

Model 1 in

Table 3. Baseline regression estimation results.

Variable	Model 1	Model 2	Model 3	Model 4
Air pollution	−0.203 ** (0.071)	−0.261 ** (0.092)	−0.313 *** (0.077)	−0.178 (0.095)
Physical exercise	0.326 *** (0.070)	0.275 ** (0.086)	0.473 *** (0.076)	0.545 *** (0.094)
Gender (Female = 1)	−0.205 ** (0.066)	−0.206 ** (0.066)	0.264 *** (0.072)	−0.166 (0.086)
Age	−0.062 *** (0.012)	−0.063 *** (0.012)	−0.056 *** (0.013)	−0.066 *** (0.018)
Age squared	0.000 * (0.000)	0.000 * (0.000)	0.000 *** (0.000)	0.000 * (0.000)
Logarithm of annual income	0.189 *** (0.033)	0.189 *** (0.033)	0.198 *** (0.035)	0.271 *** (0.041)
Community type (Urban = 1)	0.134 * (0.086)	0.138 (0.087)	−0.151 (0.093)	0.228 * (0.106)
Education	0.010 (0.009)	0.010 (0.009)	0.009 (0.009)	0.025 ** (0.011)
Air pollution × Physical exercise		0.138 (0.138)
N	3238	3238	3238	3238
Pseudo R2	0.0629	0.063	0.0218	0.1236

Note. *** p < 0.001, ** p < 0.01, and * p < 0.05. And the standard error is given in parentheses.

is the baseline model with added control variables. Upon including a series of control variables, the results show significant associations between air pollution (β = −0.203, p < 0.01), physical exercise (β = 0.326, p < 0.001), community type (urban = 1) (β = 0.179, p < 0.05), age (β = −0.062, p < 0.001), the logarithm of annual income (β = −0.189, p < 0.001), and self-rated health. Residents’ self-rated health is adversely affected by air pollution, indicating that as air pollution levels increase, residents’ health deteriorates, supporting Hypothesis 1. Compared to residents with good air quality, those living in regions with more severe air pollution may experience a certain degree of health decline. This could be explained in multiple ways: firstly, a good air environment contributes to residents’ well-being, satisfaction, and happiness, while severe air pollution may disrupt the ecological balance, leading to irreversible consequences such as decreased air quality and increased health risks, especially leading to various non-communicable diseases like heart disease, stroke, cancer, and chronic pneumonia, which pose significant risks to residents’ physical well-being. Secondly, air pollution may increase residents’ perception of health risks, resulting in specific mental stress. This perception of risk and mental stress may, to some extent, impact mental health. Additionally, air pollution may interfere with residents’ daily lives, social networks, and willingness to engage in social interactions, leading to adverse mental states such as depression, anxiety, and loneliness, negatively affecting residents’ psychological well-being.

The positive correlation between physical exercise (β = 0.326, p < 0.001) and self-rated health is highly significant, indicating that physical exercise positively impacts residents’ self-rated health. This supports Hypothesis 4, suggesting that through physical exercise, residents’ levels of physical health may be effectively improved. The reasons for the health-promoting effects of physical exercise include its ability to enhance cardiorespiratory function, increase muscle strength and bone density, improve metabolism and blood circulation, boost immunity, and positively influence mental health. Through these mechanisms, physical exercise can effectively enhance physical function, prevent the occurrence of chronic diseases, strengthen the body’s resistance and adaptability, and contribute to stress relief, mood enhancement, and overall physical and mental well-being.

To further examine the impacts of air pollution and physical exercise on physical health, we introduced an interaction term in model 2, Air pollution × Physical exercise (β = 0.138, p > 0.05). Hypothesis 7 was not supported, suggesting that the interaction between air pollution and physical exercise may not significantly affect physical health.

Furthermore, in Model 1, age (β = −0.062, p < 0.001) and age squared (β = −0.000, p < 0.05) are considered. This indicates that age significantly influences self-rated health, with a negative correlation. Additionally, the coefficient for age squared is also significant, albeit to a relatively small extent. In this scenario, we can interpret that as age increases, self-rated health tends to decrease, with a possible nonlinear relationship where the rate of decline in self-rated health slows down as age increases. Gender (female) = 1 (β = −0.205, p < 0.01) shows that the health status of males is relatively better than that of females, possibly due to the dual pressures that modern women face in their family lives and careers.

Moreover, personal income positively impact residents’ health, with higher incomes correlating with higher levels of health. This is because higher-income individuals find it easier to access better healthcare resources, including medical facilities, doctors, and medical technologies, while enjoying better living conditions such as housing, diet, and environmental hygiene. Financial stability and economic security alleviate psychological stress, positively influencing health. Furthermore, urban residents are healthier compared to rural residents. Urban residents have access to more abundant medical resources and advanced healthcare facilities, contributing to timely medical services and improved sanitary conditions and promoting healthier lifestyles. The social support systems and rich living facilities in cities also help to alleviate psychological stress, promoting an overall improvement in health.

4.3. Robustness Analysis

To further substantiate the negative effects of air pollution on residents’ well-being and the favorable impact of physical exercise on health, we conducted robustness tests using alternative indicators to measure residents’ health. Pierewan (2015) [31] and other scholars, analyzing 2008 European Values Study data sourced from 47 European countries, found that the determinants of happiness and health are similar, with joy and health showing a positive correlation, aligning with the perspectives of Subramanian (2005) [32] and scholars like Eikemo et al. (2008) [33]. Therefore, in this study, we substitute happiness for health in robustness testing.

In Model 3 of Table 3, we use the alternative indicator of happiness (“In general, do you feel happy with your life?”) as the dependent variable instead of the health level. Assigning “Very unhappy”, “Somewhat unhappy”, “Average”, “Somewhat happy”, and “Very happy” were sequentially assigned values of 1 to 5, respectively. They are ordinal variables, so an ordered logit model is used for analysis. The results indicate a significant negative correlation between air pollution (β = −0.313, p < 0.001) and residents’ happiness. In other words, higher levels of air pollution are linked to a decrease in residents’ happiness, indicating a negative correlation between increased air pollution and happiness levels among residents. Physical exercise (β = 0.473, p < 0.001) demonstrates a significant positive correlation between physical exercise and residents’ happiness, indicating that physical exercise enhances residents’ happiness.

In Model 4 of Table 3, we use another alternative indicator, health2 (“In the past four weeks, how often has your health affected your ability to work or carry out other daily activities?”), as the dependent variable to substitute for health status. Responses of “always”, “often,” and “sometimes” are assigned a value of 0, indicating poor health. Responses of “rarely” and “never” are assigned a value of 1, indicating good health. This indicator is a binary variable, so a logit regression is employed. The results show that sport (β = 0.545, p < 0.001) continues to exhibit a highly significant positive correlation with physical health, confirming that physical exercise effectively promotes health. Air pollution (β = −0.178, p < 0.1) still negatively impacts health.

The above analysis confirms the stability of the positive health effects of physical exercise across various model specifications, providing robust evidence for the conclusion that physical exercise positively influences residents’ health. The negative effect of air pollution on health also demonstrates a certain degree of stability.

4.4. Heterogeneity Analysis

(1)

Community type Heterogeneity Analysis of Health Effects

Table 4. Urban–rural heterogeneity analysis results.

Variable	Model 1	Model 5 (Urban)	Model 6 (Rural)
Air pollution	−0.203 ** (0.071)	−0.237 ** (0.080)	−0.029 (0.164)
Physical exercise	0.326 *** (0.070)	0.449 *** (0.081)	−0.033 (0.143)
Gender (Female = 1)	−0.205 ** (0.066)	−0.160 * (0.078)	−0.213 (0.127)
Age	−0.062 *** (0.012)	−0.050 *** (0.014)	−0.095 *** (0.025)
Age squared	0.000 * (0.000)	0.000 (0.000)	0.000 ** (0.000)
Logarithm of annual income	0.189 *** (0.033)	0.171 *** (0.041)	0.218 *** (0.056)
Education	0.010 (0.009)	−0.004 (0.010)	0.054 ** (0.018)
Community type (Urban = 1)	0.134 * (0.086)
N	3238	2297	941
Pseudo R2	0.0629	0.0589	0.0621

Note. *** p < 0.001, ** p < 0.01, and * p < 0.05. And the standard error is given in parentheses.

shows: community type in the overall sample model have highly significant estimated results. It indicates that urban residents (β = 0.134, p < 0.05) have more prominent health advantages than rural residents. The specific mechanism mainly stems from differences in the impact of other related factors, as evident from the heterogeneity in the estimation results of urban–rural models. Model 5 and Model 6 exhibit a high degree of consistency in the direction and significance of most parameters compared to the overall sample model, yet there is noteworthy variation in the significance of certain parameters. In terms of age and annual income, the health effect patterns between urban and rural residents are not significantly different. Males generally exhibit better health than females, health issues become more pronounced with age, and individuals with higher socioeconomic status have better physical health. Although the health effect of education did not pass the significance test in the overall sample baseline model, this may be due to the influence of urban–rural differences on parameter estimation. Notably, the health effect of education (β = −0.054, p < 0.01) in rural areas is significantly better than in urban areas. Therefore, the government needs to strengthen educational equity, promote health knowledge, and enhance the access of residents with lower education levels in rural areas to healthcare knowledge and related resources.

The core independent variables, air pollution and physical exercise, exhibit significantly different effects on the health of urban and rural residents. The health effect of air pollution for urban residents (β = −0.237, p < 0.01) is very harmful. In contrast, the impact on rural residents is relatively weaker (β = −0.029, p > 0.1), with non-significant model test results supporting Hypothesis 2. The main reason for this estimation result is that rural areas typically lack the dense industrial and traffic activities in urban areas, reducing emissions of exhaust gases and industrial pollutants. The lower population density also contributes to air circulation. A simpler lifestyle requires less energy and resource consumption, thereby reducing the potential risks of environmental pollution. Additionally, rural areas, closer to natural environments with more vegetation coverage, aid in air purification and the absorption of harmful gases. The relatively dispersed distribution of pollution sources in rural areas helps to maintain better local air quality.

The health effect of physical exercise for rural residents (β = −0.033, p > 0.1) did not pass the significance test, while for urban residents, it is significantly positive (β = 0.449, p < 0.001), supporting Hypothesis 5. This difference can be attributed to the more physically demanding labor in rural life, relatively fewer sports facilities and opportunities, lower levels of health education and awareness, and factors like transportation inconvenience. These factors collectively make it more challenging for rural residents to engage in organized sports activities, and their level of physical exercise may be lower compared to urban residents. From a policy perspective, several insights emerge: improving sports infrastructure in rural areas, conducting health education activities, raising awareness of health in rural residents, organizing various rural fitness activities, employing mobile health services and applications to provide personalized health guidance and exercise plans for rural residents, formulating policies to support sports in rural areas, promoting traditional sports activities, and implementing fitness programs for rural youth, among others.

(2)

Gender Heterogeneity Analysis of Health Effects

Due to differences in physiological structure and environmental attitudes, the impact of air pollution on residents’ health may exhibit gender disparities. Blocker and Eckberg (1997) [34] pointed out that gender attributes play a significant role in environmental attitudes. Compared to male residents, female residents may be more concerned about ecological recommendations due to their socialization in caregiver roles and relative structural positions outside the labor market and within the family. Therefore, based on the gender variable, the sample is divided into two sub-samples: females (Model 7) and males (Model 8). The regression results are presented in

Table 5. Gender heterogeneity analysis results.

Variable	Model 1	Model 7 (Female)	Model 8 (Male)
Air pollution	−0.203 ** (0.071)	−0.102 (0.100)	−0.303 ** (0.101)
Physical exercise	0.326 *** (0.070)	0.306 ** (0.099)	0.359 *** (0.099)
Community type (Urban = 1)	0.134 * (0.086)	0.251 * (0.127)	0.034 (0.119)
Age	−0.062 *** (0.012)	−0.066 *** (0.018)	−0.059 *** (0.017)
Age squared	0.000 * (0.000)	0.000 (0.000)	0.000 (0.000)
Logarithm of annual income	0.189 *** (0.033)	0.165 *** (0.047)	0.217 *** (0.046)
Education	0.010 (0.009)	0.011 (0.013)	0.003 (0.119)
Gender (Female = 1)	−0.205 ** (0.066)
N	3238	1610	1628
Pseudo R2	0.0712	0.0666	0.0571

Note. *** p < 0.001, ** p < 0.01, and * p < 0.05. And the standard error is given in parentheses.

.

Table 5 shows that the health effect patterns of physical exercise, age, income, and education do not differ significantly between males and females. Participation in physical exercise shows an apparent health-promoting effect, rejecting Hypothesis 6. As age increases, health issues become more prominent, and higher income is associated with better physical health. However, the impact of education on health is not significant. Female residents in urban areas generally exhibit better health, possibly due to the more developed medical resources and healthcare facilities in cities. This facilitates females’ access to timely, high-quality medical services and a healthier urban lifestyle. On the other hand, the health disparity between urban and rural areas is not significant for males, possibly linked to similar challenges they face regarding occupation, lifestyle, and social pressures. Air pollution (β = −0.303, p < 0.01) significantly affects the health of males negatively, while the impact on females is not significant (β = −0.102, p > 0.1), supporting Hypothesis 3. This may be attributed to the physiological structure of males, making them more susceptible to inhaled pollutants. Additionally, males often engage in outdoor activities and work, increasing their exposure to pollution. Occupational exposure is also more prevalent among males, elevating their risk of health issues related to air pollution.

Considering all these factors, we can notice that the negative impacts of air pollution tend to affect males more significantly. Therefore, policies should focus on enhancing health and awareness among rural females. This could involve strengthening healthcare infrastructure in rural areas, supporting female vocational training and employment opportunities, improving sanitation conditions, and establishing community support systems. Simultaneously, gender equality policies should be formulated to ensure that women have equal access to high-quality healthcare services, thereby comprehensively elevating their health standards.

5. Discussion

Globally, air pollution has emerged as a critical public health concern with extensive and profound effects on human health. As urbanization and industrialization accelerate, emissions of air pollutants continue to increase, leading to worsening air quality and posing a serious threat to residents’ health. According to the World Health Organization, air pollution causes millions of deaths annually due to diseases such as respiratory issues, cardiovascular conditions, and cancer, with its impact continually escalating. Understanding the impact of air pollution on human health and the potential for mitigation measures is crucial, particularly within the framework of sustainable development, which emphasizes the balancing of environmental, social, and economic factors to achieve long-term health and well-being.

Over the past few decades, researchers have extensively and deeply explored the relationship between air pollution and health. A substantial body of literature has confirmed the adverse effects of air pollution on human health, especially on the respiratory and cardiovascular systems. At the same time, physical exercise, as an important way of promoting health, has also received significant attention for its positive impact on human health. Numerous studies indicate that regular participation in physical exercise can improve cardiopulmonary function, strengthen the immune system, and lower blood pressure and blood sugar levels, thereby reducing the risk of chronic diseases. Additionally, physical exercise can enhance individuals’ mental health levels, alleviating psychological issues such as anxiety and depression. Therefore, promoting physical exercise has become an essential component of public health policies in many countries and regions.

This study provides a deep insight into the relationship between air pollution, physical exercise, and residents’ health and, by comparing it with previous research, further deepens our understanding. Previous studies have confirmed the negative impact of air pollution on residents’ health, with Beren (2016) [35] and Salvi et al. (2017) [36] pointing out that particulate matter (PM_2.5 and PM₁₀) is one of the primary factors causing respiratory and cardiovascular diseases. Other pollutants such as carbon monoxide (CO), sulfur dioxide (SO₂), and nitrogen oxides (NO_x) are also closely related to health issues. However, these studies often focus on specific regions or pollutants, and the understanding of relationships across different regions, populations, and pollutants remains insufficient. In contrast, our study not only considers the overall impact of air pollution but also extends this association to a broader population and region.

Furthermore, the literature highlights that when studying the impact of air pollution on residents’ health, it is important not to overlook heterogeneity. Miao (2010) [15], Li (2019) [16], and Li (2021) [37] believe that the impacts of air pollution on the health of different groups may vary. Specifically, due to differences in socioeconomic status, physiological structure, and perceived quality of life between men and women, air pollution tends to have a more adverse effect on the health of female residents. This contradicts our research findings, which indicate that air pollution has a more negative impact on urban males.

Through experimental research and epidemiological surveys, the benefits of exercise for health have been well established. Our study also highlights the positive impact of physical exercise on health. Cutrufello et al. (2011) [27] and Rundell et al. (2008) [28] suggest that exposure to particulate matter poses a risk to active exercisers’ respiratory systems, but studies by Zhao (2014a) [25] and Giles et al. (2014) [26] suggest that long-term regular exercise can counteract the physical damage caused by air pollution. In other words, the positive effects of exercise on health may outweigh the negative impacts of air pollution. In our study, we considered the overall impact of air pollution and focused on the potential impact of physical exercise on health, exploring the interaction between the two factors. Unfortunately, the interaction effect was not significant. However, we also found that physical exercise significantly affects the health of urban residents but not rural residents. This result emphasizes the importance of promoting physical exercise, especially as urban residents increasingly tend toward sedentary and inactive lifestyles.

Notably, our study also explores the differences between different groups. We found that factors such as age, gender, educational level, income level, and living environment have significant differences regarding the impact of air pollution and physical exercise on health. This suggests the need for personalized policies and intervention measures to maximize overall health outcomes. Based on the comparison of previous study results, our research further expands the understanding of the relationship between air pollution, physical exercise, and residents’ health. By comprehensively considering environmental factors, lifestyle, and socioeconomic factors, we provide important scientific bases for formulating more effective public health policies.

Certainly, our study has some limitations. Firstly, the data used in this study are derived from a sample of the Chinese General Social Survey (CGSS) in 2018, which may not fully represent the entire population of China. Secondly, this study mainly focuses on the impact of air pollution and physical exercise on health, without considering other potential influencing factors such as dietary habits and lifestyle. Additionally, this study uses cross-sectional data, which cannot observe long-term effects. Lastly, this study did not conduct in-depth comparisons and analyses of residents from different regions and genders, which may limit our understanding of the variations in their health status. Therefore, caution should be exercised when interpreting the results, and further analysis should be conducted in conjunction with other studies.

6. Conclusions

Using data from the 2018 Chinese General Social Survey (CGSS 2018), our research investigated how air pollution and physical exercise influence the health levels of residents. The findings indicate that air pollution has notable negative impacts on the health of residents, highlighting the urgent need for environmental governance and a profound reflection on environmental quality. By scientifically evaluating the specific health effects of air pollution, we can better understand the health costs that it imposes. Furthermore, this underscores the importance of natural ecological governance as it relates to improving the natural environment and impacts people’s physical health. In contrast to air pollution, physical exercise positively impacts residents’ health, providing scientific evidence for promoting a healthy lifestyle. However, disparities in physical exercise participation among different demographic groups may exist, warranting further research to formulate more targeted policies. By introducing gender-stratified and urban–rural-stratified models, we observed substantial variations among different groups in the detrimental impacts of air pollution on residents’ health. Male and urban residents bear greater health welfare losses due to air pollution, indicating the need to consider the unique needs of different groups when formulating ecological environment policies, ensuring the fairness and sustainability of policies. As residents age, health issues become increasingly prominent. Meanwhile, higher income is associated with better physical health, although the impact of education on health is not significant. This finding reminds us that health issues are complex systemic challenges that require comprehensive consideration of social, economic, and individual factors. The government should consider the specific needs of different groups when formulating policies to ensure their fairness and sustainability. Ecological governance is closely related to health, and the government should strengthen support for ecological civilization construction, including protecting the ecological environment and promoting sustainable development.

To further deepen the understanding of the complex relationship between air pollution, physical exercise, and health, future research can be carried out in the following areas: first, researchers could conduct more long-term cohort studies to track the health statuses of different populations and analyze their associations with environmental factors and lifestyle. Second, researchers could strengthen research on the impacts of emerging environmental pollutants and new forms of exercise on health and explore more effective health intervention measures. Finally, researchers could establish interdisciplinary cooperation mechanisms to promote the integration of environmental science, public health, and sports medicine to jointly address global health challenges.