Objective and perceived proximity to parks in relation to physical activity within parks

Abstract

Objectives

Urban parks serve as important venues for promoting physical activity. This study aimed to examine the associations between both perceived and objective proximity to parks and within-park physical activity (PA) across different intensity levels among urban park users.

Methods

This cross-sectional study was conducted in four urban parks in Beijing, China. A total of 399 park users completed questionnaires assessing perceived proximity, demographic characteristics, and within-park physical activity. Objective proximity was measured as Euclidean distance from participants’ residences to park centers using the Geographic Information Systems. Multilevel linear regression models were used to examine the associations between perceived and objective proximity and different intensities of within-park PA while controlling for sociodemographic covariates, with bootstrap methods employed for non-normally distributed outcomes.

Results

Moderate correspondence was found between perceived and objective proximity measures. Both proximity measures demonstrated significant negative associations with moderate-intensity PA and walking within parks, with exploratory spline analyses indicating that these associations gradually weakened at greater distances beyond approximately 1 mile. No significant associations were found between either proximity measure and vigorous-intensity PA within parks. Although both perceived and objective proximity were associated with within-park PA, these relationships varied across different intensity levels.

Conclusions

These findings suggest that proximity may play a more important role in promoting moderate-intensity activities and walking within parks, while for vigorous-intensity activities, factors beyond park distance may be more relevant considerations.

1. Introduction

Engaging in regular physical activity (PA) is a key health behavior that contributes to a wide range of physical, mental, and social benefits, including reduced risks of cardiovascular diseases, obesity, type 2 diabetes, certain cancers, and depression, as well as improved quality of life and life expectancy.¹ Despite these well-established benefits, physical inactivity remains a significant global health challenge, with approximately one in three adults worldwide failing to meet recommended PA guidelines.² The widespread prevalence of physical inactivity contributes to a substantial burden on global healthcare systems, both in terms of preventable disease and economic costs.³ Rapid urbanization and industrialization, along with the continuous increase in urban population density, have further intensified this challenge, transforming living environments and lifestyle patterns, reducing opportunities for PA, while promoting increasingly sedentary lifestyles.⁴

In response to this public health challenge, there has been increasing recognition of the role that built environment factors play in promoting PA participation and health, as emphasized within the social ecological model framework.^5,6 Among various public spaces and facilities, urban parks have emerged as particularly important environmental resources for health promotion, with the World Health Organization (WHO) recognizing their critical role in facilitating PA and supporting community well-being in urban settings.⁷ Urban parks provide accessible, cost-free venues for diverse forms of PA while offering environmental benefits such as air quality improvement and urban heat island mitigation, and supporting not only individual health behaviors but also community social cohesion and environmental sustainability.^8,9

A substantial body of research has explored the relationship between parks and PA. Among the various park characteristics examined, proximity has been among the most extensively studied factors influencing park utilization and PA participation.¹⁰ The research assumes that reduced travel distance and time barriers will facilitate more frequent park visitation and subsequent PA engagement. However, despite general support for the notion that closer proximity encourages park use and PA, findings across studies have been inconsistent.¹¹ For instance, several studies have found no significant associations between proximity measures to parks and PA outcomes. A study using Geographic Information System (GIS)-based distance measurements found no correlations with PA.¹² Rodriguez et al.¹³ found that objective distance to parks was not significantly associated with moderate-to-vigorous PA in San Diego.

Some scholars argue that objective distance alone may be insufficient to capture individuals’ actual or perceived accessibility, and found poor correspondence between perceived and objective distance measures among general community residents.¹⁴ However, that sample comprised general residents rather than active park users, and the extent to which similar mismatch holds among current park users remains largely unexplored. Similar patterns of low correspondence between objective measures and subjective perceptions have been observed for other PA-related environmental attributes.^15,16 Therefore, better understanding whether perceived proximity, objective proximity, or both are more important for predicting park-based PA use may help identify where efforts to promote PA through parks should be focused.

In addition, PA is a complex health behavior that encompasses multiple dimensions and characteristics,¹⁷ yet previous research on parks and PA has predominantly examined PA as a unitary construct. Intensity is a critical parameter of PA that warrants further examination, as different intensity levels of PA are associated with distinct physical and mental health benefits.¹⁸ From a motivational perspective, individuals may have different reasons and preferences for engaging in different intensity levels of PA.¹⁹ Lower-intensity, habitual activities such as walking tend to be utilitarian and routinized, such that travel distance acts as a salient cost constraining their occurrence,²⁰ whereas higher-intensity activities are more goal-directed and structured, and may depend less on proximity than on the facilities and equipment available within a park.²¹ Consistent with this distinction, park proximity has been associated more consistently with walking and lower-intensity activity than with vigorous activity,²² suggesting that the role of proximity may be contingent on activity intensity. In addition to differentiating PA by intensity, the current study also focuses on PA that occurs specifically within park settings. This allows for a more targeted understanding of how park characteristics influence health behaviors, as research has shown that proximity to parks may not be associated with total PA but more directly related to PA conducted within parks.²³

Taken together, the present study aimed to address these research gaps by examining the relationship between park proximity and within-park PA through the following research questions: (1) What is the agreement between perceived and objective proximity measures among urban park users? (2) How do both proximity measures relate to different intensities of within-park PA? By examining both perceived and objective proximity together with within-park PA at different intensity levels, the present study tests whether proximity's association with park-based PA differs systematically across activity intensities, and whether such differences hold among individuals who already use parks. This knowledge will contribute to evidence-based urban planning policies and public health interventions designed to optimize park accessibility and promote PA.

2. Methods

2.1. Participants

This study occurred in October 2023 within four urban parks located in different neighborhoods in Beijing, China. The four parks were purposefully selected to vary in size (ranging from about 13 to 50 ha), facilities, and surrounding neighborhood characteristics. This variation was intended to capture a range of park exposures and within-park activity opportunities, contributing to variability in both proximity and within-park physical activity across participants. Detailed park profiles and satellite imagery are provided in Supplementary File S1. Researchers conducted data collection over a continuous 7-day period at each park, covering both weekdays and weekends during daytime hours.

Park users were invited to participate in the research through information booths and distributed flyers. Potential participants were informed about the study purpose, their rights, and confidentiality assurances. After providing informed consent, participants completed paper-based questionnaires with the assistance of researchers. Upon completion of the survey, each participant received a small stationery gift valued at approximately $1 USD as a token of appreciation. This study received ethical approval from the Peking University Institutional Review Board, and written informed consent was obtained from all participants. A total of 428 questionnaires were initially collected. After excluding invalid and incomplete responses, the final sample consisted of 399 participants. The sample included 208 females (52.1%) and 191 males (47.9%), with a mean age of 46.2 years (SD = 15.4).

2.2. Measures

2.2.1. Within-park physical activity

Within-park PA was assessed using a modified version of the short-form International Physical Activity Questionnaire (IPAQ),²⁴ which was originally designed to measure general PA levels but was modified to specifically focus on PA occurring within park settings. Participants reported the frequency and duration of walking, moderate-intensity PA, and vigorous-intensity PA within the park over the past seven days. Weekly total minutes spent in each intensity level of PA within the park were calculated. A similar approach has been previously used to measure park-based PA among Asian populations,²⁵ and a neighborhood-specific adaptation of the IPAQ has demonstrated test-retest reliability across activity intensities.²⁶ Nonetheless, as a self-report instrument, the modified IPAQ provides only limited resolution for distinguishing activity intensities, particularly between walking and moderate-intensity activity, a constraint we revisit in the Limitations.

2.2.2. Perceived proximity to parks

Perceived proximity to parks was measured using a face-validated single-item question: “I live close to this park.” Participants responded using a 5-point Likert scale ranging from 1 (strongly agree) to 5 (strongly disagree), with lower scores indicating greater perceived proximity to the respective park. This measure captures participants’ intuitive sense of closeness without requiring them to estimate precise distances, which may be subject to recall bias and individual differences in spatial cognition. As a single-item indicator, however, it provides only a coarse measure of perceived accessibility and may capture more than residential distance alone, such as convenience or familiarity, as we note in the Limitations.

2.2.3. Objective proximity to parks

Objective proximity to parks was assessed based on the Euclidean distance from participants' residential addresses to the centroid of the specific park. Participants provided their home addresses through self-report. Then, these addresses were geocoded using Geographic Information System (GIS) software, and distances were calculated from each participant's home to the center of the park. To facilitate interpretation of results and avoid non-normal distribution issues, objective proximity to parks was further categorized into three distance groups: within 0.5 mile, 0.5 to 1 mile, and greater than 1 mile from the respective park. These cut points reflect distances commonly used in environmental health research to define neighborhood park accessibility, with 1 mile representing a widely applied threshold for walkable access.²⁷

2.2.4. Demographic covariates

In addition to PA and proximity measures, participants self-reported several demographic characteristics including age, gender, height, weight, educational level, and marital status. Age was divided into two groups: younger adults (under 60 years) and older adults (60 years and above), as per Chinese national guidelines for age-based health and social welfare classifications.²⁸ Body Mass Index (BMI) was calculated from self-reported height and weight measurements using the standard formula (weight in kg/height in m²), and subsequently categorized into non-obese (BMI <28.0 kg/m²) and obese (BMI ≥28.0 kg/m²) groups based on the Chinese obesity criteria recommended by the Working Group on Obesity in China, which use a lower threshold than the WHO standard (≥30 kg/m²).²⁹ Education level was grouped into three levels: “less than high school,” “high school diploma,” and “bachelor's degree or higher.” Marital status was classified as married or not married.

2.3. Statistical analysis

Descriptive statistics were used to summarize participant characteristics and key study variables. Skewness and kurtosis were examined for all numerical variables to assess normality assumptions. Variables were considered normally distributed if skewness fell within −2 to +2 and kurtosis fell within −7 to +7.³⁰ Correspondence between subjective and objective proximity to parks was assessed using Kendall rank correlation between perceived proximity and continuous objective distance, as it is appropriate for non-normally distributed data and accounts for tied ranks.³¹ We also conducted one-way analysis of variance (ANOVA) and Tukey post-hoc tests to examine differences in perceived proximity across different objective distance categories.

Multilevel linear regression models with random intercepts at the park level were carried out to predict within-park PA outcomes. Within-park vigorous-intensity PA, moderate-intensity PA, and walking served as the three outcome variables in this study. Since within-park moderate-intensity and vigorous-intensity PA showed non-normal distributions, bootstrapping methods (with 1000 resamples) were used to obtain robust estimates and confidence intervals. For each outcome, objective proximity and perceived proximity were examined in separate models, given their moderate correlation, to assess their independent associations with within-park PA levels. All analyses controlled for age, gender, BMI, educational level, and marital status as covariates. Multicollinearity issues were not detected in any of the models, as the computed variance inflation factors (VIFs) were all well below 5.³² As a sensitivity analysis, restricted cubic splines (4 degrees of freedom) were fit using continuous distance to assess potential nonlinearity in the association with each within-park PA outcome (see Supplementary File S2). We additionally fitted joint models entering both proximity measures simultaneously and compared their fit with the single-predictor models (see Supplementary File S2).All data were initially entered into Microsoft Excel and subsequently analyzed using R (version 4.3). Statistical significance was set at α = .05 for all analyses.

3. Results

3.1. Sample characteristics

Table 1 presents the demographic characteristics, proximity measures, and within-park PA levels of the participants. Participant numbers were similar across the four study parks, ranging from 22.81% to 25.81% per park. The participants had a mean age of 46.18 ± 15.43 years, with most being young adults under 60 years (72.68%). Gender distribution was relatively balanced, with a little more females (52.13%) than males (47.87%). More than half of participants had a bachelor's degree or higher (56.64%), over two-thirds of participants were married (68.67%), and the vast majority of participants were classified as non-obese (93.73%). For proximity measures, participants reported high perceived proximity to parks (1.67 out of 5). Objective proximity to parks was 1.62 ± 2.91 miles on average, with 23.31% of participants living within 0.5 miles, 37.34% between 0.5 and 1 mile, and 39.35% more than 1 mile from their respective parks. Regarding within-park PA levels, walking was the most frequently reported activity (198.15 min/week), followed by moderate-intensity PA (100.30 min/week), and vigorous-intensity PA (32.13 min/week). In addition, within-park vigorous-intensity PA and moderate-intensity PA exhibited non-normal, right-skewed distributions, indicating that most participants engaged in relatively low levels of these activities.

Table 1.

Sample Characteristics (Mean ± SD or n (%)).

Variable	Sample n = 399
Age	46.18 ± 15.43
Young adults (<60 years)	290 (72.68%)
Older adults ( ≥ 60 years)	109 (27.32%)
Gender
Male	191 (47.87%)
Female	208 (52.13%)
Education
Less than high school	33 (8.27%)
High school diploma	140 (35.09%)
Bachelor's degree or above	226 (56.64%)
Marital status
Married	274 (68.67%)
Not married	125 (31.33%)
BMI	23.19 ± 3.22
Non-obese	374 (93.73%)
Obese	25 (6.27%)
Perceived proximity	1.67 ± 0.79
Objective proximity (miles)	1.62 ± 2.91
Within 0.5 mile	93 (23.31%)
0.5-1 mile	149 (37.34%)
Greater than 1 mile	157 (39.35%)
Within-park vigorous-intensity PA (min/week)	32.13 ± 74.28
Within-park moderate-intensity PA (min/week)	100.30 ± 152.29
Within-park walking (min/week)	198.15 ± 197.74
Parks
Dinghui Park	91 (22.81%)
Haidian Park	103 (25.81%)
Huilongguan Park	102 (25.56%)
Shucun Park	91 (22.81%)

3.2. Correspondence between perceived and objective proximity to parks

Kendall's rank correlation analysis revealed a significant moderate positive correlation between perceived proximity to parks and objective distance (τ = 0.474, p < .001), indicating that participants who objectively reside at shorter distances were also more likely to perceive themselves as living closer to the park. One-way ANOVA also revealed significant differences in perceived proximity across the three objective distance categories (F_{(2, 396)} = 90.62, p < .001), indicating that participants' subjective perceptions of park proximity varied according to their actual residential distance from parks. Specifically, Tukey post-hoc tests showed that participants living within 0.5 miles of parks felt significantly closer than those living 0.5-1 mile away (p = .007), and the latter felt significantly closer than those living more than 1 mile away (p < .001). These findings suggest that perceived and objective proximity to parks were moderately consistent among the study sample. The moderate magnitude indicates that residents' subjective sense of proximity overlapped with, but did not fully reproduce, their actual distance, leaving room for meaningful discrepancies between perceived and objective closeness.

3.3. Associations of perceived and objective park proximity with within-park PA

Table 2 presents the associations between perceived park proximity and different intensities of within-park PA after controlling for sociodemographic covariates. The results revealed differential associations across PA intensity levels. Perceived proximity to parks was not significantly associated with vigorous-intensity PA (β = −0.01, 95% confidence interval (CI) [−0.11, 0.09]). However, participants who perceived themselves as living farther from parks reported significantly reduced moderate-intensity PA (β = −0.14, 95% CI [−0.25, −0.04]) and reduced walking (β = −0.24, 95% CI [−0.34, −0.15]) within parks. Similar patterns were observed when examining the associations between objective park proximity and different intensities of within-park PA, as shown in Table 3. Specifically, participants living 0.5-1 mile from parks showed no significant differences in any PA outcomes compared to those living within 0.5 miles. However, those residing greater than 1 mile from parks demonstrated significantly reduced moderate-intensity PA (β = −0.39, 95% CI [−0.67, −0.11]) and walking activity (β = −0.58, 95% CI [−0.84, −0.24]) within parks, while no significant associations were found for within-park vigorous-intensity PA.

Table 2.

Association of perceived park proximity with within-park physical activity.

	Vigorous-intensity PA		Moderate-intensity PA		Walking
	β	95% CI	β	95% CI	β	95% CI
Perceived proximity	−0.01	(-0.11, 0.09)	−0.14	(-0.25, -0.04)	−0.24	(-0.34, -0.15)
Age (older adults)	−0.29	(-0.52, -0.05)	0.42	(0.18, 0.67)	0.72	(0.48, 0.93)
Gender (female)	−0.41	(-0.59, -0.23)	−0.02	(-0.19, 0.18)	−0.05	(-0.22, 0.13)
Marital Status (married)	0.00	(-0.21, 0.19)	0.08	(-0.14, 0.28)	−0.01	(-0.21, 0.16)
Education (high school diploma)	0.12	(-0.21, 0.48)	0.39	(0.05, 0.78)	−0.35	(-0.68, -0.03)
Education (bachelor degree or higher)	0.36	(0.04, 0.74)	0.22	(-0.12, 0.61)	−0.56	(-0.88, -0.22)
BMI (obese)	−0.42	(-0.79, -0.05)	−0.35	(-0.74, 0.03)	−0.08	(-0.44, 0.26)

Note. Values represent standardized coefficients and 95% bootstrap confidence intervals of fixed effects. Bold values indicate statistical significance at p < .05 level.

Table 3.

Association of objective park proximity with within-park physical activity.

	Vigorous-intensity PA		Moderate-intensity PA		Walking
	β	95% CI	β	95% CI	β	95% CI
Objective proximity (0.5-1 mile)	−0.24	(-0.50, 0.06)	−0.17	(-0.45, 0.10)	−0.18	(-0.41, 0.13)
Objective proximity (>1 mile)	−0.15	(-0.43, 0.16)	−0.39	(-0.67, -0.11)	−0.58	(-0.84, -0.24)
Age (older adults)	−0.33	(-0.57, -0.07)	0.37	(0.12, 0.64)	0.68	(0.44, 0.91)
Gender (female)	−0.42	(-0.59, -0.23)	−0.04	(-0.22, 0.16)	−0.07	(-0.24, 0.10)
Marital Status (married)	−0.01	(-0.22, 0.18)	0.09	(-0.13, 0.29)	0.04	(-0.17, 0.21)
Education (high school diploma)	0.11	(-0.22, 0.47)	0.41	(0.07, 0.78)	−0.32	(-0.64, 0.01)
Education (bachelor degree or higher)	0.35	(0.02, 0.72)	0.25	(-0.10, 0.64)	−0.53	(-0.86, -0.18)
BMI (obese)	−0.42	(-0.79, -0.05)	−0.35	(-0.74, 0.03)	−0.07	(-0.44, 0.27)

Note. Values represent standardized coefficients and 95% bootstrap confidence intervals of fixed effects. Bold values indicate statistical significance at p < .05 level.

Regarding the covariates, both sets of analyses revealed that older adults, females, individuals with lower educational attainment, and obese participants engaged in significantly less within-park vigorous-intensity PA than their counterparts. In contrast, younger adults and participants with higher education levels reported engaging in less walking within parks compared with their counterparts. As a sensitivity analysis, restricted cubic spline models with continuous distance indicated significant nonlinear associations for walking and moderate-intensity PA but not vigorous-intensity PA, with steep declines below approximately 1 mile that flattened beyond it; full results are presented in the Supplementary File S2. Given the standard deviations of weekly walking and moderate-intensity PA (≈198 and 152 min), the β of −0.58 for walking (Table 3) corresponds to about 114 fewer minutes per week among participants living more than 1 mile from parks compared with those within 0.5 miles. Applying the same approach, the β of −0.39 for moderate-intensity PA corresponds to about 59 fewer minutes per week among those living more than 1 mile away, while for perceived proximity each standard-deviation increase in perceived distance corresponds to about 47 fewer minutes of walking per week.

4. Discussion

Proximity to parks has been widely recognized as an important correlate of residents’ PA, particularly in urban settings. The current study aimed to examine the correspondence between perceived proximity and objective proximity, and investigate the associations between both perceived and objective proximity to parks and within-park PA across different intensity levels among urban park users. Understanding these relationships is crucial for informing urban planning policies and public health interventions, as insight into how individuals perceive their access to parks may help identify perception gaps that could be related to park use.

We found moderate correspondence between perceived and objective proximity measures, indicating that self-reported proximity measures, even when based on intuitive agreement statements, were moderately aligned with actual distance. In other words, when park users resided closer to parks, they also subjectively perceived themselves as living closer to the parks. Our findings appear somewhat contradictory to previous research findings at first glance. For instance, Lackey and Kaczynski¹⁴ found that the agreement between perceived and objective measures was poor, with only 18% of residents reporting perceived and objective distances to parks that matched. However, it is important to note that Lackey and Kaczynski's study sample consisted of general community residents who were not necessarily actual park users, and some participants may not have been aware of park existence in their neighborhoods. In contrast, our study sample comprised active park users who likely possessed more accurate perceptions of park distance due to their regular engagement with the parks. Therefore, our results do not conflict with Lackey and Kaczynski's findings but rather provide further support for their conclusions that interventions should focus on increasing awareness and use of available recreational parks.

Regarding the relationship between park proximity and within-park PA, we found that both perceived and objective proximity demonstrated significant associations with moderate-intensity PA and walking within parks. These results align with several previous studies examining proximity to public green spaces and PA. For instance, a systematic review by Ramezanghorbani³³ found that proximity to parks was identified as a significant predictor of PA participation. Cohen and colleagues³⁴ demonstrated that residents within a 1-mile radius of parks showed four times higher weekly visitation rates and experienced 38% more exercise sessions compared to those residing at greater distances. Our study extends these previous findings by demonstrating that proximity is also an important correlate of within-park PA.

It is of note that participants living within 0.5 miles and those living 0.5-1 mile from parks showed no significant differences in moderate-intensity PA and walking in our study. The threshold-like pattern observed in our study is consistent with emerging research suggesting non-linear relationships between environmental accessibility and PA, and 1 mile represents a most commonly used buffer distance in environmental health research for defining neighborhood accessibility. This pattern was further supported by a supplementary spline analysis (see Supplementary File S2). It is possible that distances up to 1 mile are perceived as within reasonable walking or cycling distance for our study sample, while distances exceeding 1 mile may introduce barriers such as time and convenience that may be associated with less regular park utilization and within-park PA. This pattern is consistent with biomechanical perspectives on locomotion, in which walking is the most economical mode for shorter distances while higher-intensity activities are typically chosen for facility-related reasons rather than for distance-related energy economy.35, 36, 37

It should also be noted that neither perceived nor objective proximity to parks was significantly associated with within-park vigorous-intensity PA. It is plausible that vigorous-intensity activities often require more than just accessible locations, with factors such as specialized facilities, equipment availability, and dedicated spaces likely playing a role.³⁸ Previous research has demonstrated that park content and design, as well as the quantity of facilities within parks, may be more strongly associated with park-based PA than proximity.^39,40 For instance, Tucker et al.⁴¹ found that over half of parents would choose parks with specific amenities rather than the nearest park to their homes. In addition, the low and right-skewed distribution of within-park vigorous-intensity PA in our sample may have limited the statistical power to detect associations. This is reflected in the descriptive results, where within-park vigorous-intensity PA averaged only 32.13 min per week, compared with 198.15 min for walking and 100.30 min for moderate-intensity PA.

Collectively, these findings indicate that park proximity is one of several factors associated with within-park PA, and may not be sufficient on its own for all intensities of within-park PA. This pattern is consistent with the framework outlined in the Introduction, whereby proximity constrains activities of lower intensity that are sensitive to distance more than activities of higher intensity that depend on the facilities a park affords.^21,22,40 The relevance of proximity for physical activity within parks therefore appears to be conditional on activity intensity.

The unexpected covariate patterns, in which younger and more highly educated adults reported less within-park walking, may be due to these subgroups engaging in physical activity in other settings such as gyms, sports facilities, or organized programs, as previous research has shown that parks in China are predominantly used by older adults for walking.⁴²

The findings from this study offer several practical implications for urban planning and public health practice. First, the moderate correspondence between perceived and objective proximity measures in our study, compared with low correspondence found in previous research among general community residents, suggests that public health initiatives could emphasize increasing park awareness and initial engagement to support subsequent within-park PA behaviors. Second, park access within a 1-mile radius may be a useful planning consideration for low-to-moderate intensity physical activities in urban neighborhoods. This observation should be regarded as hypothesis-generating and specific to the present sample of current park users, rather than as a broadly applicable planning guideline. Third, the lack of association between proximity and vigorous-intensity activity suggests that proximity alone may not be sufficient for more demanding forms of PA. Park quality, amenities, and functional features may play more critical roles in supporting vigorous-intensity activities within park settings.

4.1. Limitations and future directions

Several important limitations should be noted when interpreting our findings. First, the cross-sectional design of our study prevented establishing causal relationships between proximity and within-park PA. Because participants were active park users, reverse causation and self-selection cannot be ruled out, given that people who already value park-based activity may have chosen to live closer to parks or may perceive their distance to parks more favorably. Future studies could use longitudinal designs following cohorts who change residence or examine the impact of new park developments on nearby communities to understand causal mechanisms. Second, objective proximity was measured as straight-line (Euclidean) distance from residence to the park centroid and categorized into three bands. This proxy does not capture street-network distance, walkability, or entrance location. Third, perceived proximity was assessed by a single face-validated item. While this approach is practical in field surveys, it may capture more than residential distance alone, including convenience, route quality, familiarity, and travel mode. Multi-item measures of subjective accessibility would allow finer separation of these components in future research.

Fourth, our sample included existing park users from four urban parks in Beijing, possibly limiting the generalizability to different urban settings and populations. The findings should therefore be interpreted as applying to individuals who already use parks, rather than to the general urban population, non-users, or residents with low park exposure. In addition, on-site recruitment of voluntary participants may introduce further selection bias, as our sample reflects park users who were present during data collection and willing to complete the survey. Future research should include parks with diverse types, sizes, characteristics, and surrounding environments to determine whether our findings can be generalized to different geographic and social contexts. Furthermore, the use of self-reported residential addresses combined with geocoding may have introduced error into the calculated objective distances.

Finally, within-park PA was measured by a context-modified short-form IPAQ, which, like other self-report instruments, is subject to recall bias and social desirability effects, and provides only limited resolution for distinguishing intensity levels, particularly between walking and moderate-intensity activity. Because such misclassification may be differential across intensities, being most likely near the boundary between walking and moderate-intensity activity, it could bias the observed associations in either direction and may partly contribute to the intensity-specific pattern we report. Device-based measures such as accelerometers or GPS tracking should be employed in future research to capture more precise activity patterns and intensity classifications.

Despite these limitations, our study has several important strengths. First, the simultaneous examination of both perceived and objective proximity measures provides a more comprehensive understanding of how physical and psychological access to parks may influence within-park PA. Second, we distinguished between different intensities of within-park PA, which provided a more nuanced perspective on how park proximity may differentially relate to specific types of PA. Third, the use of multilevel modeling at the park level appropriately accounts for the nested structure of participants within parks. Collectively, these strengths contribute to a more detailed understanding of the relationship between park proximity and PA.

Future research could also build conceptually on these findings. Examining which specific park attributes most support higher-intensity activity would help explain why proximity appears more relevant for walking and moderate-intensity use than for vigorous activity. Investigating the sources and consequences of mismatches between perceived and objective proximity could further illuminate how physical and psychological access jointly shape park-based behavior. More broadly, testing the mechanisms underlying the intensity-dependent role of proximity would move beyond the descriptive associations reported here.

5. Conclusion

This study examined the associations between both perceived and objective proximity to parks and within-park PA across different intensity levels among urban park users. Our findings revealed moderate correspondence between perceived and objective proximity measures, with both demonstrating significant associations with moderate-intensity PA and walking within parks, with exploratory analyses suggesting these associations gradually weakened at greater distances beyond approximately 1 mile. In contrast, no relationship was found between both perceived and objective proximity and within-park vigorous-intensity PA. These results suggest that proximity-PA associations vary across different types of PA among current park users, with implications for urban planning policies and public health interventions. In practical terms, distance may be more relevant for lower-intensity activities such as walking, whereas promoting higher-intensity activity within parks may require considering factors beyond proximity alone, such as the facilities parks provide. Future research should employ longitudinal or natural experimental designs to establish causal relationships and inform evidence-based interventions.

Consent to participate

Informed consents were obtained from all participants.

Authors’ contributions

Conceptualization: Zhanjia Zhang. Data curation: Zhengqing Zhou and Junwei Qian. Investigation: Zhanjia Zhang and Xiao Hou. Writing - original draft: Zhanjia Zhang.

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Written informed consent was obtained from all participants. The study protocol was approved by the Peking University Institutional Review Board.

Consent for publication

Not applicable.

Data availability

The dataset analyzed during the current study can be retrieved from the public data repository OSF (https://osf.io/sfxqz/)

Funding

There was no funding source to state.

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: