GENDER DIFFERENCES IN PHYSICAL ACTIVITY ASSOCIATED WITH URBAN NEIGHBORHOOD PARKS

Abstract

Background:

Urban neighborhood parks are designed to provide easy access to recreation and physical activity. We analyzed data from the first National Study of Neighborhood Parks, which assessed the characteristics and use of a representative sample of US urban neighborhood parks. This paper compares factors associated with active commuting (e.g., walking) to parks among men and women and park characteristics associated with observed moderate-to-vigorous physical activity (MVPA) within neighborhood parks.

Methods:

We used systematic direct observation to quantify parks visitors of all ages in 162 US neighborhood parks in 25 cities in 2016 and surveyed a sample of adult visitors: 877 women and 793 men. We used descriptive statistics to identify park facilities, amenities, and park management practices associated with park use. We also conducted multivariate regressions to identify factors most closely associated with observed park-based MVPA among all age groups and with self-reported levels of active commuting to parks among adults.

Results:

Reasons to visit parks varied by gender with women more likely than men to bring children (59% vs 42% for males, p < .001), and men more likely than women to go to parks to relax (38% vs 29% p=.01). Bringing children to parks was associated with more motorized transport among women, but not among men. Active commuting to parks was associated with living closer to parks (β= −.92, p < .0001), greater frequency of park use (β=.99, p <.0001), and the park having a working drinking fountain (β=.62, p=.01), with no significant differences between men and women. Men and boys used park facilities very differently than women and girls. Men and boys engaged in the most MVPA in soccer fields, gyms, and skate parks, while for women and girls the top three sites were pools, playgrounds, and walking paths.

Conclusions:

Significant gender differences in why men and women visit parks and how parks are used likely reflect cultural determinants of gender roles. Our findings suggest that park management practices should be adjusted to mitigate the lower use of parks and lower levels of park-based MVPA among women and girls compared to men and boys.

Introduction

Gender disparities in park use and park-based physical activity have been noted through systematic observation in multiple studies of park use (Evenson, Jones, Holliday, Cohen, & McKenzie, 2016; Joseph & Maddock, 2016). Women and girls have lower levels of park use than men and boys of all ages (Cohen et al., 2016) and also are less likely to engage in moderate-to-vigorous physical activity (MVPA) while in parks (Evenson et al., 2016; Joseph & Maddock, 2016). Women and girls also tend to engage in insufficient physical activity throughout the lifespan (Troiano et al., 2008). Physical inactivity is a primary determinant of many unfavorable health outcomes—including higher levels of osteoporosis and sarcopenia (Piercy et al., 2018), conditions that influence physical function and independence in activities of daily living as people age, and which disproportionately affect women.

Societal conditions provide both barriers and facilitators to physical activity. Today, utilitarian physical activity is limited, given the sedentary nature of most work and the prevalence of motorized transport. Thus, physical activity is largely relegated to leisure time, which is currently dominated by television and other electronic media (Ng & Popkin, 2012). However, the availability of parks and outdoor recreation facilities may serve to inspire leisure time physical activity. In 14 cities across the world, the number of nearby parks was associated with levels of physical activity (Sallis et al., 2016). One study showed that having parks nearby one’s home was associated with a lower rates of overweight /obesity among women (Veitch et al., 2016).

Parks can support leisure time physical activity through engagement in sports or competitive recreational pursuits, but these activities tend to be more appealing to men than to women (Simon & Uddin, 2018). Identifying parks facilities and amenities that support MVPA among women and girls should be considered a priority to mitigate gender disparities. Yet parks can also support MVPA by encouraging active transport, meaning any mode that uses human power, like walking, running, or biking to arrive at the park. Active transport may be especially relevant as a means to engaging in physical activity for women, given their role as caregivers and lower participation in sports (Simon & Uddin, 2018). Indeed one study showed that park users accrue more physical activity getting to the park than they do while in the park (Evenson, Wen, Golinelli, Rodriguez, & Cohen, 2013). Bocarro et al. found that park facilities were associated with boys’ physical activity while girls physical activity was associated with the presence of other children (Bocarro et al., 2015), suggesting that programming and social activities might be more relevant than facilities for girls.

Park management practices comprise the distribution of resources invested by government agencies in maintenance, design, and construction of park facilities and in park programming. How local agencies allocate their resources may thus favor some groups over others, depending on factors like age group and gender. For example, investment in playgrounds favor children, while investment in skateparks favor young adult men and boys. Investments in chess tables, benches, and shady trees might favor older adults, who may get their exercise walking to the park rather than in the park. By examining park features and how individuals interact with them, as well as which groups are most attracted to parks with specific features, can help illuminate how park management practices can support the needs of different population groups or, in contrast, neglect them by failing to include services, programs, or facilities that may be better suited to their preferences.

Most studies on physical activity and active transport have examined neighborhood socio-demographic factors and neighborhood characteristics, such as street design or sidewalk conditions (Saelens & Handy, 2008; Saelens, Sallis, & Frank, 2003; Sallis et al., 2016). This paper adds to the scope of factors that may influence physical activity by capitalizing on the National Study of Neighborhood Parks (NSNP), which measured park characteristics and park management practices that may attract visitors and support their MVPA.

Because the NSNP also collected survey data from adult park users, it enables us to examine associations between park features and active transport. Understanding which park characteristics may draw people to a park, and whether or not they engage in physical activity while there, can guide future park investments that more equitably serve all population groups.

Methods

The data analyzed in this study were derived from the 2016 wave of observations and surveys fielded from the National Study of Neighborhood Parks (Cohen et al., 2016). The study included 177 parks in a representative sample of 25 US cities with a population of greater than 100,000 residents. Within each city we selected a 10–15% sample of all neighborhood parks between 3–20 acres, choosing parks that were at least 1 mile apart and where half the parks were in neighborhoods below the median poverty level and half above. The study was approved by the RAND Human Subjects Protection Committee.

Park observations and surveys were conducted on clement days between April and August, 2016, using the System of Observing Play and Recreation in Communities (SOPARC), a validated observational tool (McKenzie TL, Cohen DA, Sehgal A, Williamson S, & Golinelli D, 2006). Two to four local field staff from each of the selected cities were recruited and trained as data collectors. SOPARC uses momentary time sampling and direct observation methods to count individuals in specific park areas by gender, age group, race/ethnicity, and activity level. It also assesses characteristics of the park areas, including the facility types and their usability, whether there are supervised activities occurring, and the types of activities. Supervised activities were entered as a yes/no response for each target area/facility, with a yes indicating that there appeared to be someone in charge and/or responsible for leading or monitoring the activity, e.g., a coach leading a practice or a lifeguard watching a swimming pool. Caregivers monitoring their own children were not counted as a supervised activity.

We mapped the parks by dividing all the spaces with boundaries that reflected the functions of the varied facilities and activity areas (e.g., playground, basketball court, sports field). Each park had a comprehensive assessment so that every facility in the park was counted and observed. Trained observers rotated through each of the mapped areas and observed them systematically according to the following schedule: Tuesday, 8am, 11am, and 2pm; Thursday, 12pm, 3pm, and 6pm; Saturday, 9am, 12pm, and 3pm; and Sunday, 11am, 2pm, and 5pm. Parks were observed during a single week, except when the weather was inclement; on those occasions, observations were rescheduled for the next week when there was no rain (same day of week and time of day).

During an observation each park user in each specific area was categorized into one of 24 groups defined by gender (male, female), age group (child, teen, adult, senior), and physical activity level (sedentary [e.g., seated, standing], moderate [e.g., walking], or vigorous [e.g., running, climbing]).

In addition, detailed inventories of park facilities were automatically generated based on the park mapping each facility as a target area (e.g., gym, classrooms, walking paths, lawns, ballfields, bleachers). The presence of amenities was also assessed (e.g., bathrooms, water fountains, picnic tables, parking lots). Each target area was assessed for the presence of supervised activities, which, along with the number and type of facilities, served as variables reflecting park management practices. At the end of each day, observers also documented the condition of the parks—including the presence of signage, graffiti, and trash—as a reflection of park maintenance practices.

Intercept Surveys

In addition to counting the park users and documenting the conditions of each park facility area, we conducted intercept surveys of adult park users in-between observations. No incentives were offered for participation. We included only the parks where we were able to recruit people to answer the surveys (162/177). The surveys were self-administered and were available in English and Spanish. Field staff were instructed to follow a quota system with a goal of recruiting between 2 and 30 visitors per park, based on the number of users observed in the park during the National Study of Neighborhood Parks 2014 data collection (which did not include intercept surveys). The quotas identified the number of surveys to be administered by visitor gender (50% male) and the level of physical activity in which they were engaged. In general, 1/3 of respondents were to be engaging in MVPA just prior to being approached and 2/3 were to be sedentary. We asked respondents to report their gender; age; race/ethnicity; their physical activity levels; frequency and duration of park use; mode of travel to the park; whether their home was within a ¼ mile, ½ mile, 1-mile, or more than 1-mile radius of the park; and reasons to visit the park. Respondents were asked to check all reasons for visiting the park that applied: exercise/physical activity, play games/sports, bring children, meet friends, relax, and other. Items to measure physical activity and park use were previously validated (Evenson et al., 2013). For measuring mode of transport to the park, respondents had 5 options: on foot (e.g., walking, jogging, and running), non-motorized vehicle (e.g., bicycle, skateboard); motor vehicle (e.g., car, motorcycle); public transport (e.g., bus, train); and other. Active transport was defined as getting to the park on foot or with a non-motorized vehicle.

Data Analysis

We calculated univariate descriptive statistics for park characteristics, park observations, and park user surveys. We conducted two-sample descriptive statistics by park users’ transportation modes (active commuting vs. motorized transport). We fitted a linear model for aggregated park-based MVPA and a logistic regression for users’ active commuting to parks (yes/no). Both models controlled for park-level variables, including park size in acres, average population density, average poverty rate, park facilities and amenities (e.g., signage promoting physical activity, gym, walking path). Neighborhood demographics and socio-economic data were drawn from the US Census 2010 and the American Community Survey 2012. We applied the generalized estimating equation method to calculate standard errors to account for repeated measures within the same park for park-based MVPA and for user clustering for the transportation model outcome. The model for active vs. non-active transportation further adjusted for person-level covariates, including age, race/ethnicity, gender, frequency and duration of visiting parks, and distance between residences and parks, and the model for park-based MVPA further adjusted for the fixed-effect or observation times (days of a week and hours in a day).

We documented the number of park facilities and distinguished between all facilities and built facilities. The category of built facilities excluded facilities like lawns, gardens, and natural features that were not man-made. We also examined the top ten facilities that generated the most MVPA by gender and age group in parks that had these facilities.

Results

Park observations

Based on the park observations, we observed 28,773 women and girls, and 34.5% were engaged in MVPA. Girls (48.2%) and female teens (40.8%) were more likely to be engaged in MVPA compared to adult women (25.1%) and older adult women (28.0%). In contrast, we observed 43% more men and boys (41,134), with 42.8% engaged in MVPA. Boys and male teens were also more likely to be engaged in MVPA compared to adult men (34.5%) and older adult men (28.6%) (p<.0001). The percentage engaged in MVPA by age group did not differ by gender. We also observed an average of 6.7 supervised activities per park (Table 1).

Table 1.

Descriptive characteristics of survey respondents and number and activity of park users

From intercept surveys	Females	Males	p-value
Number respondents	877	793
Average age	37.9 (range 18–82)	40.0 (range 18–89)	0.002
18–25	15.4%	15.2%	0.002
26–49	66.9%	60.0%
50–64	12.4%	19.0%
≥65	5.2%	5.8%
Self-reported race/ethnicity
Non-Hispanic white	39.4%	36.7%	0.03
Non-Hispanic black	14.5%	20.0%
Hispanic/Latino	32.2%	28.3%
Non-Hispanic Asian/Pacific Islander	4.4%	5.1%
Non-Hispanic other or multiple	9.5%	9.9%
Reason for visiting the park
Exercise/physical activity	32.5%	37.2%	0.05
Play games/sports	32.5%	38.4%	0.01
Bring children	59.4%	42.7%	<.0001
Meet friends	15.7%	17.6%	0.32
Relax	29.0%	38.6%	<.0001
Other	8.9%	7.7%	0.39
From SOPARC Observations	Females Observed (#parks=162) N or %	Males Observed (# parks=162) N or %	p-value
Number of park users	28,773	41,134	0.01
children (< 12 years)	9,254 (32.2%)	13,671 (33.2%)	<.0001
teens (13–19 years)	3,342 (11.6%)	6,124 (14.9%)
adults (20–59 years)	15,017 (52.2%)	19,716 (47.9%)
seniors (>60 years)	1,160 (4.0%)	1,623 (3.9%)
children in MVPA	4,461 (48.2%)	7,129 (52.2%)	0.57
teens in MVPA	1,365 (40.8%)	3,218 (52.6%)	0.09
adults in MVPA	3,775 (25.1%)	6,810 (34.5%)	0.15
seniors in MVPA	325 (28.0%)	464 (28.6%)	0.93

Adult park user surveys

Among the 177 parks included in the study, we obtained surveys from adult users of 162 parks, including 877 women and 793 men. Among these parks the average park size was 9.2 acres, the average number of park facilities was 7.4, the average population density 25,667, and 19.0% of households were below the federal poverty level in a mile radius of the park.

The overall response rate to the survey was 47%. Among women, the average age was 37.9 years and two-thirds were between the ages of 26 and 49. Thirty-nine percent were non-Hispanic White, 32.2% were Hispanic, and 14.5% were non-Hispanic Black (Table 1). On average, men were a little older than women, and the male group of participantss and included a higher percentage of non-Hispanic Black men. Among the 850 women and 758 men who answered the question related to how they came to the park, 44.5% and 48.0%, respectively, said they use active commuting to get to the park, while the remainder relied on motorized transport.

Reasons for visiting the park differed by gender. The top reason for both men and women was bringing children, but the percent among women was 59.4% vs. 42.7% among men (p < .0001). Relaxing was a reason selected by 38.6% of men vs 29.0% of women (p < .0001). Men were also more likely than women to report going to the park for playing games or sports (38.4% vs 32.5%, p < .05) (Table 1).

Table 2 compares active commuting and motor transport by gender and the association of these two transportation modes with individual park use behaviors and with park characteristics. A greater frequency of park visits was associated with active commuting to the park among both men and women. Both men and women tended to stay longer when driving to parks, but women were much more likely than men to stay for more than 3 hours if they drove to the park. The average duration of stay in parks reached by active commuting tended to be lower than visits to parks by motor transport. Overall, when combining duration and frequency of park visits, men who used active commuting also spent more time in the parks (307 vs. 232 minutes/week, p < .002), both compared to participants who drove and to women. Those who lived closer to the parks were the most likely to use active commuting, yet even among those who lived within ¼ mile of the of the park, about 24% of both men and women used motorized transport. Women who reported that their main reason for visiting the park was to bring children were less likely to use active transport to get to the park (47.5% vs 52.5%, p < .03).

Table 2.

Descriptive statistics of factors associated with active and non-active transport to parks among female and male park users

	Women				Men
	Overall % N=850	Use active commuting to park n=378	Use motorized transport to park n=472	Chi-square / t-test p-value	Overall % N=758	Use active commuting to park n=364	Use motorized transport to park n=394	Chi-square t-test p-value
Frequency of park use
Once a week or more	63.8%	52.4%	47.6%	<.0001	65.8%	55.7%	44.3%	<.0001
Less than once a week	35.8%	30.3%	69.7%		29.6%	32.1%	67.9%
Duration of park use
0–30 minutes	13.1%	69.4%	30.6%	<.0001	13.2%	66.0%	34.0%	<.0001
31–60 minutes	31.3%	49.3%	50.8%		30.3%	56.1%	43.9%
More than 1 hour, but less than 2 hours	25.3%	38.6%	61.4%		23.4%	41.2%	58.8%
2–3 hours	22.2%	36.5%	63.5%		23.9%	33.2%	66.9%
More than 3 hours***	7.2%	21.3%	78.7%		7.3%	52.7%	47.3%
Est average duration of park use (minutes)	85.9	72.9	96.2	<.0001	87.3	77.3	96.5	<.0001
Duration x frequency	265	271	260	.6	272	310	237	0.002
Distance from park
0 – 0.25 mile	27.1%	76.1%	23.9%	<.0001	26.8%	75.9%	24.1%	<.0001
0.25 – 0.50 mile	12.0%	65.7%	34.3%		12.4%	53.2%	46.8%
0.5 – 1 mile	14.8%	31.0%	69.1%		16.2%	42.3%	57.7%
More than 1 mile	30.5%	13.1%	86.9%		28.0%	15.1%	84.9%
Reason for visiting the park
Exercise/physical activity	32.1%	49.8%	50.2%	0.03	36.6%	54.7%	45.3%	0.005
Play games/sports	32.5%	25.0%	75.0%	<.0001	37.9%	40.4%	59.6%	0.003
Bring children	59.1%	47.5%	54.5%	0.03	42.2%	44.1%	55.9%	0.11
Meet friends	15.4%	51.2%	48.8%	0.09	16.9%	64.4%	35.6%	<.0001
Relax	29.0%	53.0%	47.0%	0.001	38.1%	61.3%	38.7%	<.0001
Other	9.1%	49.3%	50.7%	0.35	7.9%	52.7%	47.3%	0.43
	Women				Men
	N = 850	Use active commuting to park n=378	Use motorized transport to park n=472	Chi-square/ t-test p-value	N = 758	Use active commuting to park n=364	Use motorized transport to park n=394	Chi-square/ t-test p-value
Supervised activities at park
None	8.9%	55.3%	44.7%	<.0001	8.0%	54.1%	45.9%	0.005
1 to 5	37.4%	51.9%	48.1%		38.9%	54.8%	45.2%
6 or more	53.6%	37.5%	62.5%		56.1%	42.8%	57.2%
Average number of supervised activities	11.9	10.4	13	0.007	12.0	10.3	13.6	0.001
Facilities
Avg number of facilities (max 25)	8.7	8.9	8.5	0.08	9.0	9.2	8.7	0.06
Avg number of sports facilities (max 17)	6.7	6.9	6.5	0.11	7.0	7.2	6.7	0.07
Avg number of elements in playground area	4.2	4.2	4.2	0.98	4.2	4.2	4.2	0.99
Avg number of play structures	4.4	3.6	5	0.0004	4.6	4.1	5	0.048
	% of respondent s from parks with facility
Baseball field	61.6%	38.6%	61.5%	<.0001	61.3%	42.8%	57.2%	0.0003
Basketball court (outdoor)	59.3%	48.0%	52.0%	0.01	62.3%	53.4%	46.6%	0.0001
Classroom	26.2%	54.7%	45.3%	0.0003	32.2%	58.7%	41.3%	0.0001
Dog park	8.8%	37.3%	62.7%	0.19	9.6%	46.6%	53.4%	0.79
Gym	22.6%	51.0%	49.0%	0.04	23.7%	58.3%	41.7%	0.002
Multi-purpose sports field	41.1%	46.4%	53.6%	0.34	44.6%	46.5%	53.6%	0.44
Picnic area	56.0%	40.6%	59.5%	0.009	57.8%	45.2%	54.8%	0.07
Playground	93.6%	44.4%	55.7%	0.78	94.5%	48.6%	51.4%	0.19
Skate park	10.8	59.8%	40.2%	0.002	12.8	52.6%	47.4%	0.34
Splash pad	15.1	49.2%	50.8%	0.23	15.0	50.0%	50.0%	0.75
Swimming pool	21.6	44.6%	55.4%	0.98	2.1	46.7%	53.3%	0.69
Tennis court	39.9	44.5%	55.5%	0.97	44.2	46.6%	53.4%	0.48
Walking path	35.4	48.2%	51.8%	0.11	34.2	48.7%	51.4%	0.8
Water feature	17.9	54.6%	45.4%	0.006	19.3	54.1%	45.9%	0.1
Weight room	7.1	35.0%	65.0%	0.12	8.3	47.6%	52.4%	0.95
Park Amenities
Any buildings	45.1	42.4%	57.6%	0.18	44.7	47.0%	53.0%	0.54
Art, sculptures or monuments	24.0	51.5%	48.5%	0.02	22.7	54.1%	45.9%	0.07
Bike racks	43.0	42.1%	57.9%	0.22	43.9	45.4%	54.7%	0.19
Community garden	2.7	73.9%	26.1%	0.004	4.9	54.1%	46.0%	0.45
Decorative water fountains	10.7	58.2%	41.8%	0.005	10.8	62.2%	37.8%	0.01
Designated park parking lot*	64.4	36.6%	63.4%	<.0001	64.4	42.6%	57.4%	<.0001
Flood lights for sports fields	51.5	43.4%	56.6%	0.51	50.5	49.6%	50.4%	0.38
Grills or fire pits	31.2	28.3%	71.7%	<.0001	32.5	36.2%	63.8%	<.0001
Many shade trees	53.5	48.6%	51.4%	0.01	56.5	52.8%	47.2%	0.003
Natural water features	10.8	47.8%	51.2%	0.49	11.1	46.4%	53.6%	0.76
Ornamental garden	11.6	53.5%	46.5%	0.05	11.7	52.8%	49.2%	0.34
Outdoor shelters (structures with roofs)	50.6	35.8%	64.2%	<.0001	47.9	41.3%	58.7%	0.0004
Pathway lighting	45.5	44.2%	55.8%	0.88	48.2	48.2%	51.8%	0.92
Pet waste bags/receptacles	16.4	33.8%	66.2%	0.006	16.5	38.4%	61.6%	0.02
Restroom*	69.8	41.3%	58.7%	0.005	69.8	48.4%	51.6%	0.76
If restroom, it is accessible	62.7	40.5%	59.5%	0.24	61.7	47.4%	52.6%	0.22
Trash cans	99.4	44.3%	55.7%	0.11	99.3	47.8%	52.2%	0.15
Vending machines*	16.5	44.3%	55.7%	0.96	17.3	56.5%	43.5%	0.03
Working water fountain	69.8	47.2%	52.8%	0.01	74.9	49.8%	50.2%	0.09
Signage
Marketing posters	24.1	42.4%	57.6%	0.5	25.9	50.5%	49.5%	0.42
Marketing banners	43.1	48.6%	51.4%	0.03	44.1	53.6%	46.4%	0.006
Bulletin boards	37.3	51.1%	48.9%	0.003	40.4	53.6%	46.4%	0.01
Signage promoting physical activity	32.4	53.5%	46.6%	0.0003	35.1	58.3%	41.7%	<.0001
Maintenance
Graffiti	68.5	45.4%	54.6%	0.44	71.1	49.5%	50.5%	0.19

“*” in the first column denotes statistically significant differences between men and women.

^*p < .05;

^**p<.01;

^***p<.001

Park facilities more likely to attract visitors who drove to the park included large numbers of play structures in the play area. Park amenities associated with motor transport included the presence of: a restroom, outdoor shelters, a working drinking fountain, pet waste bags/receptacles, grills or fire pits, many shade trees, a parking lot, a baseball field, a picnic area, an outdoor basketball court, marketing banners, and more supervised activities. Facilities and amenities associated with greater active commuting to the parks included: having artwork, a decorative water fountain or water feature, a community garden, a gym, a skate park, a classroom, and marketing bulletin boards or signage promoting physical activity.

Figure 1 shows the top ten park areas responsible for the most MVPA in parks. In most facilities men and boys accrued substantially more MVPA than women and girls. The top three areas differed substantially by gender. For women and girls, pools, playgrounds, and walking paths were associated with the most MVPA, supporting an average of 166, 102, and 94 person-hours of MVPA, respectively. For men and boys, it was the soccer field, the gym, and the skate park supporting an average of 319, 251, and 242 person-hours of MVPA, respectively—more than double that of women and girls. Among children, girls were most active in the pools and water features and boys were most active in the gyms and on the soccer fields. Female teens were also active in pools, followed by soccer fields and multipurpose fields, and male teens engaged in the most MVPA on soccer fields and in skate parks. Adult women were most active on walking paths and adult men were most active on soccer fields. Seniors, both men and women, were most active on a pickleball court. (Pickleball is a relatively new sport that has been described as a hybrid of ping-pong and tennis.) However, very few parks even had these facilities and only one park had a pickleball court. No parks had fitness equipment. Lawns were the most common facility in all the studied parks (95% of all parks), but they supported relatively little MVPA.

Figure 1.

Top 10 areas and facilities where park users engage in moderate-to-vigorous physical activity, by gender (rank of popularity and % of parks with the facility)

Table 3 shows the models indicating which factors were most closely associated with active commuting to the park and MVPA in the park. Being male increased the likelihood of MVPA in the park, but not active commuting. Population density was important for active commuting to the park. For every additional 10,000 population, the odds of walking to the park increased by 12.7%. Park size appeared to play no role in active commuting to the park, but a higher neighborhood poverty level was associated with much higher active commuting rates. Those who visited a park at least once per week had a 2.7-fold increased odds of walking to the park and those living further away were more likely to drive to the park. Active commuting was not associated with age or race. However, every additional minute in the park corresponded to a 1% decreased odds of using active commuting.

Table 3.

Models of predictors of active commuting to parks and MVPA within parks (male & female park users)

Variable	Active commuting to parks reported by survey respondents			Observed MVPA in Parks
	Estimate	SE	Pr > |Z|	Estimate	SE	Pr > |Z|
Female	−0.03	0.16	0.84	−0.38	0.04	<.0001
Male (ref)
Park size (Acres)	−0.02	0.02	0.32	0.02	0.01	0.009
Population density (per 10,000)	0.12	0.03	<.0001	0.10	0.01	<.0001
% households in poverty	2.04	0.80	0.01	−0.83	0.34	0.01
Age (years)	0.00	0.005	1.00
White Only (ref)
Black Only	0.10	0.25	0.70
Hispanic Only	0.27	0.18	0.12
Asian Only	−0.25	0.38	0.51
Other or Multiple	−0.15	0.27	0.59
Uses park 1x/week or more	0.99	0.18	<.0001
Average duration of stay (minutes)	−0.01	0.002	<.0001
Distance from park (mile categories)	−0.92	0.07	<.0001
Park has bulletin boards	−0.07	0.28	0.81	−0.08	0.11	0.49
Park has signage promoting PA	0.48	0.25	0.06	0.09	0.11	0.37
Park has classroom	0.40	0.41	0.33	0.01	0.13	0.96
Park has gym	−0.41	0.43	0.35	−0.04	0.20	0.82
Park has pool	−0.04	0.28	0.89	−0.15	0.15	0.31
Park has walking path	0.003	0.19	0.99	−0.05	0.09	0.58
# park facilities	−0.03	0.05	0.56	0.10	0.03	0.00
Park has restroom	−0.56	0.29	0.05	0.07	0.13	0.57
Park has picnic table	0.54	0.25	0.03	0.04	0.10	0.71
Park has grills or fire pit	−0.70	0.19	0.00	0.05	0.11	0.63
Park has outdoor shelter	0.01	0.18	0.96	0.03	0.09	0.69
Park has working water fountain	0.62	0.23	0.01	0.19	0.10	0.06
# supervised activities	0.002	0.01	0.81	0.24	0.03	<.0001
Park has baseball field	−0.61	0.21	0.00	−0.25	0.10	0.01
Park has multipurpose sports field	−0.12	0.18	0.50	0.10	0.10	0.32
Park has outdoor basketball court	0.13	0.22	0.54	−0.004	0.09	0.96

If the park had signage promoting physical activity, the odds of active commuting increased by 60%. A picnic table and a working water fountain also were associated with greater odds of active commuting. Amenities associated with driving to the park included the presence of a grill or fire pit, a baseball field, and a restroom, the latter of which was associated with increased odds of driving by 43%.

Higher levels of MVPA at the park were associated with larger park size and population density. For every 10% increase in neighborhood poverty level the engagement in MVPA was reduced by 7.0%. Every additional park facility was associated with a 10.5% increase in MVPA and every additional supervised activity offered was associated with a 27% increase. The presence of a baseball field was associated with 22% less MVPA.

Discussion:

Key findings of this study include the stark differences between men and women in reasons for park visits and between women and girls and men and boys in park-based MVPA and in the use of park facilities. Yet there were few gender differences in active commuting to parks.

Relatively few women and girls take advantage of park facilities and spaces that support MVPA, like skateparks, soccer fields, and basketball courts. Although there have been partially successful efforts to increase women’s sport participation (Deaner et al., 2012; Kaestner & Xu, 2007) and there has been a shift in more women joining the work force over the past half century (BureauLaborStatistics, 2018), women still function as the primary caregivers for children, and this appears to constrain their activity in public parks. Even with more progress in gender equality, women still engage in lower levels of physical activity than men (Sturm & Cohen, 2019). A study published more than 20 years ago found similarly higher proportions of women accompanying children in parks as we have observed in 2016 (Scott D, 1994). Caretaking may be the reason the presence of a skatepark was associated with active transport among women, yet no women were seen using that facility.

Most of the organized sports observed were among men and boys. Parks with sports fields and that host tournaments may be more likely to have vending machines, which may explain the relationship between vending machines and active commuting among men. We saw few programs and activities that appeared to cater to women and girls. Women’s lower use of parks and lower physical activity levels have been highlighted as an important issue in several local studies, (Derose, Han, Williamson, & Cohen, 2018; Jilcott, Vu, Morgan, & Keyserling, 2012) as well as in other countries (Veitch et al., 2016). Since the park facilities do not attract women and girls as much as men and boys, it appears that there is a need to determine whether park-based programs and classes might be more appealing.

Proximity of parks has been cited as an important facilitator of park use, especially for women and girls (Cohen et al., 2006; Kaczynski et al., 2014; Mowen, Orsega-Smith, Payne, Ainsworth, & Godbey, 2007), and is confirmed in our study with a clear association between distance and active commuting. Recognizing the role of active transport in leisure time activities may be a more immediate means for boosting MVPA among women and girls. In an international study on walking, the number of parks within a 1 km-radius was the most important correlate of MVPA (McKenzie & van der Mars, 2015). Although parks comprise neighborhood amenities that are one of the most common destinations for active transport (Boone-Heinonen et al., 2009), many people drive to parks. However, they miss out on the physical activity obtained in active transport when they do so, even though absolute duration of park use may be similar. Therefore, the movement to create parks within a half-mile of all urban residences (TPL, 2017) is a worthy goal and is likely to promote more MVPA among women and girls.

This study provides more granular detail as to the most relevant park facilities and features that are associated with active transport and park-based physical activity. Besides distance, the presence of a working water fountain and a picnic table were the two park features most strongly associated with active commuting. This suggests that one of the important roles of a neighborhood park is that it should be a place for respite. If people walk a long distance, they need a place to sit and rest once arriving. The presence of a restroom, grill, and baseball field were associated with motorized transport. These features suggest a longer duration of stay at a park, which, on the one hand, may encourage driving, possibly because after a long stay, people might feel too tired to walk home. However, they may also represent desired features that people are willing to drive to utilize, and these features are not found in most neighborhood parks.

While swimming pools were at the top of park facilities that promoted MVPA in women and girls, they are expensive to install and maintain and are not likely to be feasible to include in all neighborhood parks. That being said, the attraction of a pool did not stand out as a facility to which park users are more likely to drive. The fact that people drive from a distance to go to a specific park may suggest that it has special facilities or programming that make it sufficiently compelling to go there; once in a car, the choice of where to go is much greater than if one had to walk. Our data also show that when people drive to a park, they may not go as frequently, but they stay for a longer period of time. The facilities and amenities associated with driving to a park included a grill or firepit, a restroom, and baseball field, all of which may be associated with a longer duration of stay. A baseball field, however, has the disadvantage of being associated with lower levels of MVPA, as people are either sedentary spectators or sedentary participants sitting on the bench waiting for their turn at bat for at least half the game.

Park facilities and supervised activities were the two most modifiable factors that supported MVPA. With more facilities, there is greater variety and potentially a wider appeal to a broad group of users. Parks could install additional facilities to attract more women, such as walking paths, as this was the facility most used by adult women.

The small number of activities in parks, an average of 6.7 observed over 4 days including 2 weekend days, reflects either a lack of resources or a lack of demand. That we saw more park users in parks with marketing materials suggests that demand could be stimulated if activities were available and marketed. Supervised activities should include programming that may be attractive to women and girls. However, these typically require staffing or contracting with personnel to lead the classes and may be a management as well as financial challenge to local parks departments. And if women are responsible for children, unless childcare services are also available, they may not be able to take advantage of programming that caters to their preferences. To solve this, more equitable distribution of childcare responsibilities is needed. Further, if other factors like concerns about safety or fear of harassment exist (Lapham S et al., 2015; Yen, Scherzer, Cubbin, Gonzalez, & Winkleby, 2007), the ability to increase women’s use of parks may be further challenged regardless of renovations or program availability.

Limitations of the study include the fact that the survey response rate was just 47%, but this is equal to the median response rate for the 2016 national BRFSS survey (Okoro, Hollis, Cyrus, & Griffin-Blake, 2018), where multiple calls are made to the selected sample. Nevertheless, the limited response rate may introduce bias in our data. We also did not collect data from park users under age 18, so survey data is only representative of adults. Respondents only represent those who are taking advantage of park facilities; we did not collect data from non-users. It is possible that people who do not use the park will have different characteristics or preferences. A previous study comparing park users to a general population showed that non-park users spent more of their leisure time engaged with electronic media (Cohen et al., 2013). The parks also only represent those between 3 and 20 acres and cannot provide insights on smaller or larger parks. However, these parks represent the size of approximately half of all the urban parks in the United States (Cohen et al., 2016).

Implications for Practice and/or Policy

Because women and girls are underrepresented among park users and parks should serve all residents equally, efforts should be undertaken to achieve equity. This may require upgrading or redesigning parks to make them as attractive to women and girls as they are to men and boys, regardless of whether parks are reached on foot or through motorized transport. Amenities, facilities, and programs—including drinking fountains, walking paths, restrooms and programmed classes that are appealing to women—should be added to neighborhood parks. Given that distance was a barrier to active commuting, increasing the number and quality of local neighborhood parks may mitigate this barrier.

Conclusion

Most neighborhood parks are currently lack both the programming and facilities that are most attractive to women and girls, including swimming pools and walking paths. Because parks are an important destination and proximity is associated with visitation, the creation of parks within walking distance from homes may boost active transport and increase MVPA among local residents. Future studies should develop and test different types of facilities and programs that will support and motivate more women to visit parks as well as to engage in park-based physical activity.

Author Biographies

Deborah Cohen, MD, MPH, RAND Corp, is a Senior Physician Policy Researcher with interests in parks and physical activity and the influence of the built environment on health and health behaviors.

Stephanie Williamson, BA, RAND Corp, is a Senior Research Programmer with interests in public health and education reform.

Bing Han, PhD, RAND Corp, is a senior statistician with interests in public health and health services research.