Gender Differences: The Role of Built Environment and Commute in Subjective Well-Being

Abstract

The literature has shown extensive interest in exploring the factors of subjective well-being (SWB). However, most research has conducted cross-sectional analysis of the built environment (BE), commute, and SWB, and little is known about gender differences in their connections. Based on two periods of survey data of 4297 respondents from China, the study performs a cross-sectional and longitudinal examination of whether the BE and commute have effects on SWB, and how the effects differ between men and women. The results reveal that BE features, including destination accessibility and residential density, significantly affect SWB, with stronger impacts observed among men. Men benefit more from greater accessibility and are more negatively affected by higher residential density than women. In contrast, commute mode and duration influence SWB in similar ways for both genders. A shift from nonactive to active commuting improves well-being for men and women alike. Furthermore, certain life events produce gender-specific effects. For instance, childbirth increases SWB for men but decreases it for women. These findings highlight the importance of gender-sensitive planning in building inclusive urban and transportation environments that enhance population well-being.

1. Introduction

The relationships between the built environment (BE), commute characteristics, and subjective well-being (SWB) have attracted increasing academic attention. Understanding whether these connections differ between men and women is essential for enhancing quality of life through urban planning and policy making [1,2]. Although there is a growing body of research on this topic in Western countries, it is important to note that findings from these contexts may not be directly applicable to developing countries like China and India, due to differences in travel habits and social characteristics. As one of the largest developing countries, China is undergoing rapid urbanization, characterized by urban sprawl and changing social features [3]. This presents a valuable opportunity to examine the connections between BE features, commute patterns, and SWB in a developing country context.

Despite the extensive research on the connections between the BE, commute, and SWB, there are still several significant gaps in the literature. First, the literature primarily focuses on the cross-sectional connections between commute and SWB, which has confirmed the role of commute features in affecting SWB [4,5]. However, the cross-sectional findings fail to explain the causal effects of commute features on SWB. Disentangling the causal connections between commute and SWB is essential for promoting targeted strategies to enhance SWB. Some research has offered evidence of changes in commute modes and commuters’ emotions [6,7]. In this regard, it is plausible to hypothesize that changes in commute modes and duration may result in SWB changes. The causal effects remain, as far as we know, predominantly unexplored. Second, researchers have explored the role of the BE in shaping SWB in the literature, and they find that the BE not only directly affects SWB but also indirectly affects SWB through some mediators [5,8,9]. However, it remains unclear whether the BE changes result in SWB changes from a longitudinal perspective. As the BE is associated with daily life choices (e.g., travel and shopping), changes in the BE may result in changes in satisfaction with different life domains. Moreover, some studies have found that changes in SWB can occur as a result of BE changes due to residential relocation [10]. Thus, more longitudinal explorations are needed to further observe the causal connections between the BE and SWB. Third, the literature indicates significant gender differences in perceptions of the surrounding environment and travel behavior between men and women [11,12]. The differences in these factors may affect SWB both through differences in themselves and through differential satisfaction with specific sub-domains [4,13]. However, there has been limited research examining how these changes in factors reshape SWB in a gender-specific manner. Understanding the distinct impacts of the BE and commute on SWB for men and women has significant policy implications for enhancing SWB.

Drawing on two periods of survey data collected from 4297 residents across China, the study contributes to our current knowledge by (1) performing a cross-sectional and longitudinal examination of the connection between the BE, commute, and SWB and (2) examining potential gender differences in this connection. By addressing these research gaps, the study aims to provide insights for policymakers and urban planners in identifying areas that require prioritized attention and efforts for enhancing SWB.

This paper is structured as follows. Section 2 presents a review of prior studies on the BE, commute, and SWB and gender differences in their connections. Section 3 details the data and model used in this study. Section 4 presents the model results. Section 5 summarizes the important findings. Finally, the study is concluded.

2. Literature Review

2.1. Connection Between the BE and SWB

In recent decades, both theoretical and empirical studies have increasingly focused on the relationships between the BE, commuting behavior, and SWB [14].Existing research has highlighted that the BE is closely associated with car dependence and vehicle kilometers traveled [15,16]. Furthermore, even within the same travel environment, travelers with different behavioral or attitudinal profiles exhibit substantial differences in travel satisfaction [17]. Given that enhancing SWB is a core objective of urban planning [9,18,19], it is essential to deepen our understanding of the population-specific and policy-sensitive mechanisms through which the BE affects mobility and well-being. It is widely recognized that both individual and external features contribute to SWB. In addition to SWB, which reflects individuals’ self-reported life satisfaction, objective well-being, represented by measurable factors such as health and income, also plays a key role in evaluating quality of life. Integrating these perspectives provides a more comprehensive understanding of how the BE and commuting patterns affect residents’ experiences in urban and rural areas [20]. The literature has highlighted the significant role of the BE in affecting SWB, although it is generally considered to be less important than individual factors [19]. For example, Cao explored the effects of neighborhood BE on SWB using Campbell et al.’s framework [10]. The results of structural equation models indicate that neighborhood connectivity and density have significant effects on SWB. Yin et al. employed a machine learning approach to explore neighborhood BE and SWB and found that the BE features account for 35.50% of respondents’ SWB, which is lower than the contribution of individual factors [19]. Li et al. explored the role that metro proximity plays in affecting SWB based on survey data from Shanghai and found that living closer to a metro station significantly enhances respondents’ SWB [21]. However, the majority of research on the BE and SWB has been conducted using cross-sectional analyses. These cross-sectional findings merely describe part of the portrait, as they limit the potential to explore the causality between the BE and SWB.

Some researchers have attempted to explore the connection between the BE and various life choices using longitudinal or quasi-longitudinal designs [22,23]. For example, Wang et al. analyzed the effects of the BE on commute mode choice based on longitudinal survey data from China and found that changes in the BE trigger shifts in commute modes [11]. In another example, Aditjandra et al. analyzed the effects of BE changes on changes in travel behavior by employing structural equation models and found that BE changes lead to changes in driving behavior [8]. Aside from travel behavior, the literature suggests that BE changes also contribute to changes in health and crime risk [24]. Given that SWB encompasses satisfaction with various life domains, changes in life choices resulting from BE changes may also impact SWB. Although several studies have paid attention to the relationship between the BE and SWB [2], there is limited longitudinal analysis of the BE and SWB, particularly with a focus on gender differences.

2.2. Connection Between Commute and SWB

The literature on commute and SWB mainly focuses on two key features: commute duration and commute modes. Most prior studies on commute duration and SWB have consistently observed a negative relationship between them [19,25,26]. For example, Sun et al. used structural equation models to analyze the commute duration and SWB link [25]. They found that commute duration negatively affects SWB through several mediators, despite lacking a significant direct effect. Yin et al. employed a machine learning method to explore the nonlinear effects of commute duration on SWB [19]. The findings indicate that although the specific nonlinear patterns vary across modes, the effects are all negative in general. However, despite this prevailing body of research, a consensus has yet to be reached concerning the relationship between commute duration and SWB. Some studies have failed to uncover evidence supporting this negative association [1,27]. For example, Morris and Zhou, using data from the American Time Use Survey, found no correlation between commute duration and SWB. This result may be explained by the fact that people who have long commutes tend to choose their jobs prudently, maximizing personal utility [27]. Aside from the aforementioned cross-sectional analyses, a limited number of researchers have analyzed the effects of commute duration on SWB based on longitudinal data [2,28]. Similar to cross-sectional studies, the findings of longitudinal analyses are also mixed. Several researchers have found that commute duration has a negative effect on SWB [2], whereas some others have observed no significant correlation between them [29,30]. Choi et al. used four periods of survey data from the United States to analyze the relationship between commute duration and SWB and confirmed the significant role of commute duration in affecting SWB [28]. By contrast, Clark et al. analyzed six periods of survey data from England and found no evidence supporting the negative effects of commute duration on SWB [29]. However, it is worth noting that most previous studies have treated individuals as the unit of analysis, with limited attention given to examining gender differences in the relationship between commute duration and SWB.

In addition to commute duration, there has been increasing interest in investigating the association between commute modes and SWB [31,32,33]. It is commonly believed that commuting is a stressful activity, but different commute modes are known to impact SWB differently. Active modes are generally found to be less stressful, and many studies have provided evidence supporting this relationship [34,35]. Using a multiple linear regression model, Smith found that active commuters exhibit the highest SWB based on an analysis of survey data from the United States [33]. Similarly, St-Louis et al. compared the SWB of commuters using different modes and found that pedestrians, train commuters, and cyclists tend to have higher SWB compared to other commuters [34]. Some researchers have provided conflicting evidence regarding the link between commute modes and SWB. For example, Yin et al. found that active commuters have a lower SWB compared to transit commuters [19]. However, these cross-sectional studies limit the potential to infer the causal effects of commute modes on SWB. Only a limited number of studies have explored the longitudinal relationship between commute modes and SWB [2,11]. Martin et al. analyzed 18 periods of survey data from Great Britain and found that shifting from driving to walking or cycling improves respondents’ SWB [11]. Based on two periods of survey data from China, Wang et al. found that shifting from motorized modes to walking or cycling leads to losses in SWB, whereas shifting in the opposite direction has no significant effect [2]. However, the literature rarely focuses on gender differences in the longitudinal relationship between commute modes and SWB.

2.3. Gender Differences in SWB

The literature indicates significant gender differences in travel behavior, needs, opportunities, and perceptions of the surrounding environment [1,11,36]. These differences are likely rooted in the distinct social roles assigned to women and men. Throughout history, women have carried the primary responsibility for managing and accommodating the travel requirements associated with both work and household maintenance. Given their increased family responsibilities, such as childcare and shopping, women tend to exhibit more complex mobility patterns [37]. Consequently, women often make travel-related decisions under more pronounced spatial and temporal constraints and are generally less influenced by the surrounding environment [11]. Several studies have offered empirical support for these observations [1,38]. For example, Sweet and Kanaroglou found significant gender differences in the links between travel time, activity participation, and SWB based on survey data from Canada. The findings indicate that reductions in travel time contribute to higher SWB by promoting activity participation among women [1]. Joshanloo and Jovanović, analyzing survey data from over 100 countries, identified variations in gender differences in SWB across different spatial contexts [38]. These gender differences in travel behavior and perceptions of the surrounding environment not only affect SWB directly but also affect SWB through differential access to opportunities as an indirect function of mobility [10,39]. However, there remains a scarcity of studies that specifically investigate gender differences in the links between the BE, commute, and SWB based on longitudinal or quasi-longitudinal designs, which is important for promoting targeted strategies.

Despite substantial advances in understanding the independent effects of the BE and commuting patterns on SWB and increasing recognition of gender differences, relatively few studies have systematically integrated all three elements into a unified analytical framework. Gender fundamentally shapes time allocation, mobility behavior, and social roles, thereby mediating how individuals experience and are affected by their residential environment [20]. For example, women often show heightened sensitivity to neighborhood safety and access to daily amenities, leading to distinct well-being outcomes compared to men [18]. Furthermore, associations between multiscale BE characteristics and health-related outcomes are highly heterogeneous across urban and rural settings and are further moderated by mobility patterns and socio-demographic factors such as gender [40]. These findings underscore the need to explicitly integrate gender and travel behavior into analyses of the BE–well-being relationship to reveal nuanced, context-specific mechanisms and support more equitable urban policy.

In summary, while many studies have explored the connections among the BE, commute, and SWB using cross-sectional data, few have examined their causal relationships through longitudinal analysis. This gap limits our understanding of whether interventions targeting the BE and commute can enhance SWB, and leaves the role of gender differences underexplored. Addressing these gaps, the present study investigates both cross-sectional and longitudinal links between the BE, commute, and SWB, with special attention to gender differences.

3. Data and Methods

3.1. Data and Variables

To explore the heterogeneous relationship between the BE, commute, and SWB among men and women, the study utilizes data from the Labor-force Dynamics Survey (CLDS). It is performed nationally by Sun Yat-sen University in collaboration with over 100 universities every two years. It has become one of the most representative surveys focusing on labor, the living environment, health, and SWB in recent years in China. Sun Yat-sen University has been gradually releasing the survey data, and the survey data collected in 2012, 2014, and 2016 have been publicized (see https://www.sysu.edu.cn/ for details, accessed on 1 August 2025). The site primarily provides content in Chinese, and non-Chinese readers can request access and support by directly contacting the CLDS project team via email (cssda-ta@mail.sysu.edu.cn). The CLDS has been widely used in previous research, with numerous studies providing detailed descriptions of its sampling design, variables, and longitudinal structure [2]. Given the absence of the BE variables in the 2012 dataset, this analysis solely employs survey data from participants in the 2014 and 2016 surveys. To ensure the validity of the analysis, respondents who were unemployed, did not participate in either survey wave, or had missing data for any study variable were excluded from the sample. After these exclusions, the final analytical sample comprised 2219 men and 2078 women. The geographic distribution of the survey participants and the study area are illustrated in Figure 1.

The dependent variable, SWB, was measured on a five-point scale, with higher values indicating greater life satisfaction. BE variables, such as residential density and destination accessibility, were treated as continuous variables. Community type was coded as one for urban and zero for rural. Commute mode was coded as one for active commuting, including walking and cycling, and zero for non-active commuting. Commute duration was recorded as a continuous variable in minutes. All control variables, including individual and household features and life events, were coded according to standard social survey conventions.

The dataset records respondents’ SWB by asking them to report their SWB using a five-point term ranging from very dissatisfied (1) to very satisfied (5). Among the exogenous variables, individual and household features are directly extracted from the dataset. Respondents are asked to report their average commute duration and main commute modes at each period. The change in commute duration is determined by subtracting the duration of the first period from that of the second period. If a respondent walks or cycles to work in the first period and commutes by non-active modes in the second period, the value of shifting from active to non-active is equal to 1; otherwise, it is equal to 0. If a respondent uses non-active modes to work in the first period and walks or cycles to work in the second period, the value of shifting from non-active to active is equal to 1; otherwise, it is equal to 0. All life events, with the exception of income changes, are directly sourced from the dataset. The variable of change in income is derived by calculating the difference in income between the two periods. Among the four BE features, the community type and distance to central business district (CBD) are directly extracted from the dataset. The dataset records the population and area of each neighborhood, which are used to obtain the residential density. The dataset records the presence or absence of seven key amenities (i.e., hospitals, banks, playgrounds, squares, reading rooms, schools, and sports amenities) within each neighborhood. The value of destination accessibility is determined as the sum of these seven categories of key amenities. The descriptive statistics of the variables are presented in

Table 1. Sample characteristics.

Indicator	Description	Period One		Period Two
		Mean/Percentage		Mean/Percentage
		Men	Women	Men	Women
Commute features
Active commuting	Using active modes to work	58.99%	54.33%	52.05%	45.52%
Commute duration	Average commute duration (min)	23.99	19.37	30.86	29.29
Shifting from active to non-active	1, using active modes for commute at period 1 and non-active modes for commute at period 2; 0, otherwise			15.77%	20.93%
Shifting from non-active to active	1, using non-active modes for commute at period 1 and active modes for commute at period 2; 0, otherwise			13.34%	12.13%
Change of commute duration	Difference in commute duration between two periods			6.87	9.92
BE features
Destination accessibility	Accessibility to seven categories of key amenities	3.85	3.79	3.94	3.90
Community type	1, urban neighborhood; 0, rural neighborhood	34.29%	32.29%	34.88%	33.49%
Residential density	Number of population/neighborhood area (10,000 persons/km2)	1.28	1.29	1.31	1.32
Distance to CBD	Distance from home to CBD (km)	3.88	3.74	4.01	3.88
Change of destination accessibility	Difference in destination accessibility between two periods			0.09	0.11
Change of community type	Difference in community types between two periods			0.59%	1.20%
Change of residential density	Difference in residential density between two periods (10,000 persons/km2)			0.03	0.03
Change of distance to CBD	Difference in distance to CBD between two periods (km)			0.13	0.14
Individual and household features and life events
Age	Age in years	48.68	46.71	50.68	48.71
Education	1, college degree and above; 0, otherwise	11.97%	13.12%	12.21%	13.47%
Family size	Number of family members	4.75	4.79	4.70	4.65
Income	Family yearly income (10,000 RMB)	3.93	3.80	5.74	4.85
Switching employers	1, the survey taker changes jobs between two periods; 0, otherwise			8.84%	9.34%
Childbirth	1, the survey taker has a child between two periods; 0, otherwise			9.26%	8.13%
Car acquired	1, the survey taker purchases a car between two periods; 0, otherwise			13.47%	14.59%
Electric-bike acquired	1, the survey taker purchases an electric bike between two periods; 0, otherwise			54.09%	53.29%
Marriage	1, the survey taker gets married between two periods; 0, otherwise			25.80%	26.34%
Divorce	1, the survey taker gets divorced between two periods; 0, otherwise			6.89%	5.25%
Change in income	Difference in family income between two periods (10,000 RMB)			1.81	1.05
Dependent variables
SWB	A five-point term ranging from very dissatisfied (1) to very satisfied (5)	3.62	3.65	4.05	3.75
Change of SWB	1, changing from a situation of satisfaction (or neutral) to a situation of dissatisfaction; 2, changing from a situation of dissatisfaction (or neutral) to a situation of satisfaction; 3, otherwise			1.60	1.53

Note: Period 1 corresponds to the 2014 wave; Period 2 corresponds to the 2016 wave of the CLDS.

.

3.2. Method

The conceptual framework for this study is depicted in Figure 2. Solid arrows represent the associations explicitly examined in this study. Dashed arrows indicate potential relationships not investigated in the current study, but acknowledged as possible influences. To examine gender differences in the relationship between the BE, commute, and SWB, we employed multilevel ordered logit regression models for cross-sectional analysis [41]. Models were estimated separately for men and women, with the outcome variable, SWB, measured on a five-point scale ranging from very dissatisfied to very satisfied. For the longitudinal analysis, we used multilevel multinomial logit models to assess how changes in commute and BE features influence changes in SWB across the two survey periods. Both modeling strategies account for the hierarchical structure of the data, allowing for random effects at the neighborhood and individual levels, and enable us to capture within-person variations, as well as unobserved heterogeneity across neighborhoods. The ordered variable $y_{ni}^{\ast }$ is transformed into a latent continuous variable $y_{ni}^{\ast }$.

$$y_{ni}=\left\{\begin{array}{l}1, if y_{ni}^{\ast }\le {\sigma }_1\\ 2, if {\sigma }_1<y_{ni}^{\ast }\le {\sigma }_2\\ ⋮\\ 5, if y_{ni}^{\ast }>{\sigma }_4\end{array}\right.$$

(1)

where $y_{ni}$ represents the SWB of respondent $i$ in neighborhood $n$ in period one. ${\sigma }_1$, ${\sigma }_2$…${\sigma }_4$ represent the cuts.

Then, $y_{ni}^{\ast }$ is estimated as follows:

$$\begin{array}{l}\mathrm{Level} 1: y_{ni}^{*}={\eta }_0+\alpha X_{ni}^{AC}+\lambda X_{ni}^{CD}+\gamma X_{ni}^{IH}+{\epsilon }_{ni}\\ \mathrm{Level} 2: {\eta }_0=\beta X_n^{BE}+{\varsigma }_n\end{array}$$

(2)

where ${\eta }_0$ represents a varying intercept; $X_{ni}^{AC}$ and $X_{ni}^{CD}$ represent the commute features (i.e., mode and duration) of respondent $i$ in period one; $X_{ni}^{IH}$ represents the individual and household features of respondent $i$ in period one; $X_n^{BE}$ represents the BE of the neighborhood in period one; ${\alpha }_{ni}$, ${\lambda }_{ni}$, ${\gamma }_{ni}$, and ${\beta }_n$ represent corresponding parameters; and ${\epsilon }_{ni}$ and ${\varsigma }_n$ are errors at two levels, respectively.

Second, a multilevel multinomial logit approach is used to explore the effects of the BE and commute features on SWB from a longitudinal perspective. In this model, changes in the BE and commute features are included. As some respondents relocate their residences between the two periods, two potential approaches are considered to account for the varying intercept: the neighborhood in the first period and the neighborhood in the second period. After conducting tests on both models, it is determined that the model utilizing the neighborhood in the first period as the basis for the varying intercept yields better results. Consequently, the final model can be represented as follows:

$$\begin{array}{l}\mathrm{Level} 1: U_{nim}={\mu }_0+\phi X_{nim}^{AC}+\kappa X_{nim}^{CD}+\rho X_{nim}^{IH}+\varphi \mathrm{\Delta }{X}_{nim}^{AC} +\psi \mathrm{\Delta }{X}_{nim}^{CD}+\tau \mathrm{\Delta }{X}_{nim}^{BE}+\chi X_{nim}^{LE}+{\zeta }_{nim}\\ \mathrm{Level} 2: {\mu }_0=\omega X_n^{BE}+{\delta }_{nm}\end{array}$$

(3)

where $U_{nim}$ represents the utility associated with the $mth$ category of transition in SWB; ${\mu }_0$ represents a varying intercept; $\mathrm{\Delta }{X}_{ni}^{AC}$ and $\mathrm{\Delta }{X}_{ni}^{CD}$ represent changes in commute modes and duration, respectively; $\mathrm{\Delta }{X}_{ni}^{BE}$ and $X_{ni}^{LE}$ represent changes in the BE and life events, respectively; ${\phi }_{ni}$, ${\kappa }_{ni}$, ${\rho }_{ni}$, ${\varphi }_{ni}$, ${\psi }_{ni}$, ${\tau }_{ni}$, ${\chi }_{ni}$, and ${\omega }_n$ represent corresponding parameters; and ${\delta }_n$ and ${\zeta }_{ni}$ represent errors at two levels, respectively.

$$P_{ihwm}={\displaystyle \frac{\exp (\omega X_n^{BE}+\phi X_{nim}^{AC}+\kappa X_{nim}^{CD}+\rho X_{nim}^{IH}+\varphi \mathrm{\Delta }{X}_{nim}^{AC}+\psi \mathrm{\Delta }{X}_{nim}^{CD}+\tau \mathrm{\Delta }{X}_{nim}^{BE}+\chi X_{nim}^{LE}+{\delta }_{nm})}{{\displaystyle {\sum }_{m=1}^M\exp (\omega X_n^{BE}+\phi X_{nim}^{AC}+\kappa X_{nim}^{CD}+\rho X_{nim}^{IH}+\varphi \mathrm{\Delta }{X}_{nim}^{AC}+\psi \mathrm{\Delta }{X}_{nim}^{CD}+\tau \mathrm{\Delta }{X}_{nim}^{BE}+\chi X_{nim}^{LE}+{\delta }_{nm})}}}$$

(4)

where $P_{ihwm}$ is the probability of respondent $i$ having the $mth$ category of transition in SWB, and $M$ represents the number of categories of transitions.

In the above methods, the total variance is composed of variances at different levels. A term, namely intra-class correlation (ICC), is a commonly utilized index to assess spatial heterogeneity, and it is determined by calculating the ratio of the neighborhood variance to the total variance. The ICC value can be obtained as follows:

$$ICC={\displaystyle \frac{{\sigma }_n^2}{{\sigma }_{ni}^2+{\sigma }_n^2}}$$

(5)

where ${\sigma }_n^2$ and ${\sigma }_{ni}^2$ are variances on the neighborhood and individual scales, respectively. The variance at the individual scale is $\sqrt{{\pi }^2/3}$ in logit models [42].

4. Results

4.1. Cross-Sectional Analysis of Gender Differences in SWB

Figure 3 presents the cross-sectional analysis results of gender differences in the relationship between the BE, commute, and SWB. In the two models, the ICC values indicate that the neighborhood variances account for 15.35% and 12.32% of the total variances in SWB, respectively. The results highlight the importance of addressing spatial heterogeneity in SWB.

With regard to commute features and individual and household features, no significant differences are observed in their effects on SWB. Specifically, engagement in active commute is found to have a positive effect on SWB, consistent with previous evidence that active commuting is associated with greater psychological well-being, including reduced stress levels and improved mood [36]. An increase in commute duration is significantly associated with lower levels of SWB. Regardless of gender, older and well-educated respondents from high-income families tend to report higher levels of SWB. This suggests that promoting active commuting may enhance well-being by increasing physical activity and social contact, while long commutes might reduce available leisure or family time, thus diminishing overall satisfaction.

As for BE features, the results reveal significant differences in their effects among women and men, and they exhibit stronger effects on men’s SWB. Destination accessibility and residential density are significant contributors to SWB in both gender groups, indicating that improved access to various amenities and residing in low-density communities positively impact SWB for both women and men. These findings align with the findings in previous studies [43]. Conversely, the effect of community type on women’s SWB is not found to be significant, while for men, living in urban areas is associated with higher SWB. Furthermore, gender differences are observed in the effects of distance to CBD, with men living farther from CBD tending to report higher levels of SWB. These gender differences may be explained by the fact that men are more likely to benefit from diverse urban amenities and employment opportunities, while women’s SWB may depend more on local resources and neighborhood-level factors, such as perceived safety and daily convenience.

4.2. Longitudinal Analysis of Gender Differences in SWB

Figure 4 and Figure 5 visualize the longitudinal analysis results of gender differences in the relationship between the BE, commute, and SWB. The ICC values of the model for men indicate that neighborhood variances account for 18.75% and 14.51% of the total variances for the two transitions in SWB, respectively. By comparison, the ICC values of the model for women are 10.90% and 12.17%, respectively. The results indicate the necessity of addressing spatial heterogeneity in SWB changes, as residents in different urban and rural contexts, especially those in marginal urban areas, experience markedly different commuting constraints and life satisfaction trajectories. Ignoring such spatial variations could lead to generalized planning policies that fail to capture localized needs.

Regarding commute features, the results indicate no significant gender differences in their effects on the propensity of two transitions (i.e., going from dissatisfied to satisfied and from satisfied to dissatisfied). Specifically, an increase in commute duration significantly contributes to the transition from satisfied to dissatisfied while also playing a significant role in deterring going from dissatisfied to satisfied. In the relationship between commute duration and SWB, the role of gender is limited. In addition, regardless of gender, respondents who switch their commute modes from non-active modes to walking or cycling between two periods have a propensity to go from dissatisfied to satisfied. These effects of changes in commute features partially support the causal relationship between commute features and SWB. Significant gender differences are observed in the relationship between active commuting and the two transitions in SWB. Men active commuters at period one are less likely to go from dissatisfied to satisfied, whereas women active commuters at period one are more likely to go from satisfied to dissatisfied. Long-duration commuters in period one are less likely to go from dissatisfied to satisfied. This longitudinal evidence underscores the potential for promoting active commuting and reducing commute durations as effective strategies for sustaining and improving well-being over time.

The findings highlight the greater sensitivity of women’s SWB to life events. Specifically, switching employers increases the propensity to go from satisfied to dissatisfied for both men and women. Childbirth significantly drives a transition from dissatisfied to satisfied for men, whereas for women, it not only reduces the propensity to transition from dissatisfied to satisfied but also increases the likelihood of transitioning from satisfied to dissatisfied. The gender differences in the effects of childbirth may be attributed to the unequal distribution of childcare responsibilities, with women typically shouldering a greater burden [44]. Purchasing a car and the change in income show similar effects on the two transitions. They significantly drive a transition from dissatisfied to satisfied and deter a transition in the opposite direction. In contrast, purchasing an electric bike only significantly affects two transitions for women. Divorce only reduces the propensity to go from satisfied to dissatisfied for women. These findings highlight the role of social context and gendered expectations in shaping how life events affect individual well-being.

Similar to the cross-sectional results of BE features, BE changes exert stronger effects on men’s SWB changes. Specifically, the increase in destination accessibility encourages a transition from dissatisfied to satisfied and deters a transition in the opposite direction for men. In contrast, its effects on women’s SWB changes are not significant. Moving from rural communities to urban communities increases the propensity to go from dissatisfied to satisfied for both men and women. The increase in residential density deters a transition from dissatisfied to satisfied and encourages a transition in the opposite direction for men, respectively. In contrast, it only shows a significantly negative effect on the transition from dissatisfied to satisfied for women. The increase in distance to CBD only improves the propensity to go from dissatisfied to satisfied for men.

Among the baseline contexts, the effects of community type, residential density, and age show significant gender differences. The community type only shows a significantly negative effect on the transition from dissatisfied to satisfied for men. The residential density only encourages men to go from satisfied to dissatisfied. Age is a significantly positive and negative contributor to the two transitions, respectively.

Compared with previous longitudinal studies focusing separately on commuting duration or mode shifts [28,29], this study provides an integrated longitudinal analysis by concurrently examining changes in both commuting characteristics and BE features, explicitly highlighting gender-specific transitions in SWB, thereby offering a more nuanced understanding of causal mechanisms.

5. Discussion

This study focuses on gender differences in the relationship between the BE, commute, and SWB. It addresses the following underexplored issues: (1) it explores the relationship between the BE, commute, and SWB from both cross-sectional and longitudinal perspectives; and (2) it distinguishes whether the relationship differs between men and women. This section discusses the key findings and implications that emerge from the paper.

First, the study revisits commute and SWB from both cross-sectional and causal perspectives. Our findings are consistent with previous studies regarding the positive and negative effects of active commuting and commute duration on SWB [45,46]. Moreover, we present additional evidence of these relationships through a longitudinal analysis. The results show that an increase in commute duration not only limits the likelihood of dissatisfied individuals regaining satisfaction but also increases the risk of a decline in well-being among those who were previously satisfied. Combining these findings with the negative effect of commute duration from a cross-sectional perspective, we infer a causal relationship between commute duration and SWB. Regarding commute modes, our study reveals that a shift from non-active modes to active modes increases the chance of dissatisfied individuals achieving greater satisfaction. This emphasizes the importance of promoting active modes in commute trips, particularly in rapidly urbanizing areas, where residents are often compelled to switch to motorized travel due to urban sprawl. To achieve this, it is crucial to consider implementing planning strategies and public policies such as affordable housing projects that offer better proximity to workplaces and the creation of cyclist-friendly environments. This emphasizes the importance of promoting active commuting, particularly in rapidly urbanizing areas where motorization is increasing. Our longitudinal findings support prior research: Knott et al. showed that shifts toward active commuting reduce depressive symptoms [6], while Stutzer and Frey found that longer commute times undermine life satisfaction [47]. Our study extends this evidence by showing that switching from non-active to active commuting improves SWB, especially among previously dissatisfied individuals.

Second, the study performs a cross-sectional and longitudinal examination of the relationships between the BE and SWB. The results indicate that destination accessibility and residential density are significantly correlated with SWB from a cross-sectional perspective. These results agree with most prior studies [48]. Moreover, the results indicate that changes in BE features also result in changes in SWB from a longitudinal perspective. The results indicate that a decrease in residential density and moving from rural to urban areas promote positive changes in well-being among those who were dissatisfied. Additionally, changes in destination accessibility and distance to CBD have the potential to improve SWB for those initially dissatisfied. These results not only provide evidence of a causal relationship between the BE and SWB but also underscore the possibility of improving SWB by optimizing the BE features. Furthermore, recent evidence from Beijing shows that residents in marginal rural areas of megacities experience substantial mobility constraints due to limited transport and job access, which significantly impact their well-being and commuting patterns [49].

Third, the results enrich the literature by exploring gender differences in the effects of the BE on SWB. The results reveal significant gender differences in both cross-sectional and longitudinal links. In the cross-sectional analysis, the community type and distance to CBD are found to be significant contributors to men’s SWB, while they do not show the same influence on women’s SWB. In the longitudinal analysis, it is observed that changes in destination accessibility and distance to CBD only trigger transitions of men’s SWB. These results indicate that the BE features have stronger effects on men’s SWB. This phenomenon may be explained by the fact that women often undertake additional family responsibilities, such as childcare and shopping [50]. Consequently, their mobility patterns are subject to more spatial and temporal constraints, and the residential BE may have a weaker impact on their daily travel patterns. In light of these findings, it is imperative for urban planners and policymakers to recognize and account for these gender differences to develop more targeted and refined strategies. In addition to these structural and societal factors, behavioral differences between men and women may also contribute to the observed disparity. Previous research has shown that men are more likely to engage in longer commutes, explore a wider range of urban spaces, and make greater use of available urban amenities and recreational facilities [12]. This broader spatial behavior could make men more sensitive to changes and improvements in the BE, thereby amplifying its impact on their SWB. Conversely, women’s daily routines may remain more localized due to caregiving roles or safety concerns, reducing their exposure to or benefit from changes in neighborhood or city-scale BE features.

Fourth, significant gender differences in the links between life events and changes in SWB are recognized. Specifically, while the effects of switching employers, car acquired, marriage, and changes in income do not significantly differ between genders, other variables are more influential on women’s SWB. Childbirth emerges as a negative contributor to women’s SWB, likely due to increased responsibilities and role conflicts, as women often bear a disproportionate share of childcare and domestic duties following childbirth. This added burden may limit their mobility, leisure time, and personal autonomy, leading to declines in SWB. By comparison, it only shows a positive effect on achieving satisfaction for men who were previously dissatisfied. This difference is probably because women usually spend more time on childcare compared to men [51]. Divorce is found to significantly trigger a transition from dissatisfied to satisfied only among women. This unexpected pattern may reflect a release from stressful marital relationships or an increase in perceived autonomy post-divorce, which for some women could outweigh the negative emotional and social costs typically associated with marital dissolution. Additionally, the purchase of an electric bike shows positive and negative effects on two transitions for women, respectively, whereas its effects on men’s SWB changes are not significant. Thus, encouraging the ownership and usage of electric bikes may be an effective way to enhance women’s SWB. Moreover, it is worth noting that SWB in China may be influenced by broader contextual factors beyond commuting and the BE. For instance, Zhang et al. found that air pollution significantly decreases individual well-being [52]. Although environmental factors were not included in the current analysis, future research could incorporate such variables to provide a more comprehensive and integrated understanding of well-being determinants in rapidly urbanizing Chinese cities.

6. Conclusions

Although the literature has extensively examined the connection between the BE, commute, and SWB based on cross-sectional data, there is a lack of consideration for causality. Moreover, little is known about whether gender differences exist in this connection. The study provides additional evidence of gender differences in the connection from both cross-sectional and causal perspectives.

The findings highlight that men’s well-being is more strongly shaped by BE characteristics, such as residential density and access to destinations. This suggests a heightened sensitivity among men to the structural and functional attributes of their neighborhoods, potentially due to differentiated patterns of spatial interaction. These gendered sensitivities underscore the importance of nuanced urban design that accounts for varying behavioral responses to environmental change.

The influence of commuting mode and duration on well-being does not significantly differ between men and women. Both genders experience benefits from active travel and stress from extended commute durations. This convergence suggests that interventions promoting active transportation can yield universal well-being benefits irrespective of gender.

The influence of life transitions, such as childbirth, reveals substantial gender asymmetry. While childbirth enhances well-being among men, it reduces it among women, reflecting persistent inequalities in caregiving roles. Similarly, other life events, such as divorce and changes in mobility resources, exhibit gender-specific effects, pointing to broader societal patterns embedded in daily experience.

These findings emphasize notable gender differences in how the BE and commuting influence SWB, highlighting the need for gender-sensitive considerations in urban planning and transportation system design. However, the study is not without limitations. First, self-selection effects were not considered due to data unavailability. Future studies could explore the role of self-selection in shaping individuals’ residential and commuting choices. Second, we did not examine the pathways through which the BE and commuting affect SWB, such as travel satisfaction or residential satisfaction. If data on these mediating variables become available, future research should investigate the mechanisms underlying these relationships.