1. Introduction
Lack of physical activity among children is a serious public health issue in the United States (US). Only four in ten US children engage in the recommended amount of physical activity [
1], and 11% of children do not participate in any vigorous physical activity in a given week [
2]. When examining physical inactivity by ethnicity, Hispanic children are the most inactive; 15% of Hispanic children with US-born parents, and 17–23% of Hispanic children with immigrant parents do not engage in any vigorous physical activity (vs. 9% of non-Hispanic White children) [
2]. A low level of physical activity among Hispanic children contributes to their high prevalence of overweight and obesity (over 23%) [
3], a prevalence that has continued to increase despite recent declines among non-Hispanic White children [
4]. As Hispanic populations are the largest ethnic minority group in the US, representing 29% of children [
5], and will account for 32% of the total child population in 2060 [
6], attention has recently increased to identify factors related to physical inactivity in Hispanic children.
1.1. Neighborhood Impacts on Child Physical Activity
The neighborhood environment can affect child physical activity levels, for example, by influencing whether they can safely walk to a park or playground [
7,
8,
9,
10]. In addition, children may observe more residents engaging in physical activity when living in safe and secure neighborhoods. The active lifestyle of residents attributes to social norms, which may encourage children to engage in physical activity [
11]. In other words, living in safe and secure neighborhoods may positively affect child physical activity, mediated by social norms around engaging in activity. Empirical studies that have investigated the association between the neighborhood social environment (e.g., safety, crime, traffic) and child physical activity have reported mixed results [
8,
9,
12,
13,
14,
15,
16,
17]. The mixed findings may partly be explained by limited consideration of moderating factors such as individual-level ethnicity and cultural factors (e.g., immigrant generation, language spoken at home), in addition to variation in survey designs, sample characteristics, and measurements. Thus, the present study examines the moderating effect of individual-level cultural factors in the association between the neighborhood social environment and physical activity. In addition, given that no prior studies have investigated potential mediating mechanisms underlying an association, this study investigates the mediating role of neighborhood physical activity-related social norms in the association between the social environment and child physical activity.
1.2. Considering Cultural Factors to Understand How Neighborhood Impacts Physical Activity
Several studies have examined individual-level cultural factors, such as language spoken at home and immigrant generation, as a correlate of physical activity in Hispanic children [
2,
17,
18,
19,
20,
21,
22,
23]. Theoretically, two contrasting mechanisms have been proposed to examine the association between cultural factors and physical activity in Hispanic children. On the one hand, as physical activity is often a social behavior for children mostly involving groups of peers [
24,
25], fewer language barriers and greater exposure to US culture are likely to provide more opportunities to be physically active with peers from diverse cultural backgrounds. On the other hand, greater exposure to US culture may increase the risk of inactivity for Hispanic children given acculturation-related stress (leading to unhealthy lifestyle) and the high levels of physical inactivity in the US [
26,
27,
28]. A few empirical studies reported that being first-generation or Spanish speaking is directly associated with less physical activity among Hispanic children after adjusting for parental education and economic level [
2,
20,
21], which supports the first conceptual mechanism (see an exception in [
18]).
Cultural factors may interact with social environments to affect child physical activity although to our knowledge, no empirical research has investigated its interaction effect. Past literature showed variation across Hispanic immigrant generations in types of resources that parents use to support their children. For example, parents of first-generation Hispanic children used informal support for their children, while parents of third-generation Hispanic children used formal and structural support for their children [
29]. Living in a high-quality social environment may be an important informal support for Spanish-speaking Hispanic children and first- or second-generation Hispanic children. Safe and supportive environments may lead to more physical activity within the neighborhood thereby exposing them to visual cues of neighbors and peers being physically active in their neighborhood. Also, Spanish-speaking or first- or second-generation Hispanic families can face practical and cultural challenges to navigating and utilizing physical activity resources outside their neighborhood [
21,
30,
31,
32]. Thus, living in a health-promoting neighborhood may provide more beneficial informal support for them in comparison to non-Hispanic White children and Hispanic children from English-speaking or third-generation families who may be more likely to use formal and structural resources.
1.3. The Present Study
The present study extends past research by examining moderating and mediating mechanisms underlying the association between the social environment and physical activity in non-Hispanic White and Hispanic children. We hypothesized that (1) the protective role of the social environment in physical activity would be stronger for Spanish-speaking or first- or second-generation Hispanic children compared to non-Hispanic White children and English-speaking or third-generation Hispanic children; and (2) active and healthy social norms in the neighborhood would mediate the association between the social environment and physical activity.
3. Results
Sample characteristics are shown in
Table 1. In the analytical sample, 44% of children identified as English-speaking non-Hispanic White children, 36% identified as Hispanic children speaking English and Spanish, 14% identified as Hispanic children speaking Spanish, and 6% identified as Hispanic children speaking English at home. Regarding immigrant generation, 44% of children were second-generation Hispanic children, 7% identified as first-generation Hispanic children, and 6% identified third- or higher-generation Hispanic children. Children reported participating in any of six physical activity types 7.8 times during the past week on average.
Table 2 present the results of bivariate analyses. Bivariate statistics showed a significant difference in physical activity level between the following groups: (1) children aged 10–12 years (M = 8.2) and children aged 13–15 years (Mean = 7.1), (2) boys (Mean = 8.1) and girls (Mean = 7.4), (3) English-speaking non-Hispanic White children (Mean = 8.4) and Hispanic children speaking English and Spanish or speaking Spanish (Mean = 7.2 and 7.1, respectively), (4) US-born non-Hispanic White children (Mean = 8.4) and second generation Hispanic children (Mean = 7.2), (5) children of parents with some college (Mean = 8.0) and children of parents with less than high school (Mean = 7.0), (6) children of parents with college graduate or above (Mean = 8.3) and children of parents with less than high school (Mean = 7.0), (7) children of families with annual income
$75,001 or above (Mean = 8.4) and children of families with annual income less than
$20,000 (Mean = 7.3),
$20,000–35,000 (Mean = 7.3), and
$35,001–50,000 (Mean = 7.3), and (8) children living in high SES neighborhoods (Mean = 8.3) and children living in low or moderate SES neighborhoods (Mean = 7.6 and 7.3, respectively).
Table 3 presents associations of the social environment and language at home with physical activity. The social environment was not significantly associated with physical activity in the overall sample. Hispanic children speaking English and Spanish or speaking Spanish had less physical activity than non-Hispanic White children (b = −1.38 and −1.66, respectively,
p < 0.001) after controlling for child age and gender (Model 1). With the inclusion of parental education, household income, social environment, and neighborhood socioeconomic status in the model (Model 2), the associations remained significant (b = −0.94,
p < 0.05 and b = −1.24,
p < 0.01 respectively). There was no statistical difference in physical activity between Hispanic children speaking English and non-Hispanic White children (b = −0.62,
p > 0.05). In Model 3, there was a significant interaction effect of speaking English and Spanish at home (vs. non-Hispanic White children) in the association between the social environment and physical activity (b = 1.60,
p < 0.001). As shown in
Figure 1, the association between the social environment and physical activity was positive for Hispanic children speaking English and Spanish whereas the association was negative and flatter for non-Hispanic White children. Although it was not significant at alpha = 0.05 for Hispanic children speaking Spanish, the association between the social environment and physical activity was marginally significant and positive for Hispanic children speaking Spanish in comparison to non-Hispanic White children (b = 0.92,
p < 0.10).
Table 4 presents associations of the social environment and immigrant generation with physical activity. The social environment was not significantly associated with physical activity in the overall sample. First or second-generation Hispanics had less physical activity than US-born non-Hispanic White children after adjusting for child age and gender (b = −1.53,
p < 0.01 and b = −1.46,
p < 0.001 respectively). The associations remained significant (b = −1.12,
p < 0.05 for first-generation Hispanics and b = −1.03,
p < 0.01 for second-generation Hispanics) when parental education, household income, social environment, and neighborhood socioeconomic status were added to the model (Model 2). There was no significant difference in physical activity between third- or higher-generation Hispanics and non-Hispanic White children. In Model 3, significant interactions were found between the social environment and child immigrant generation in the association with physical activity. As shown in
Figure 2, the association between the social environment and physical activity was positive for first- and second-generation Hispanic children (b = 2.03,
p < 0.01 and b = 1.29,
p < 0.01, respectively), whereas the association was negative and flatter for non-Hispanic White children.
As a sensitivity analysis, we conducted multiple imputation for missing data and examined the interaction effect of cultural factors in the association between the social environment and physical activity in the imputed data.
Table 5 presents pooled estimates after multiple imputation. Multiple imputation changed the magnitude of coefficients but did not alter the significance.
Finally, we tested the mediation of neighborhood physical activity-related social norms in the association between the social environment and physical activity. Before the mediation analysis was performed, we stratified the data by two cultural factors and tested the association between the social environment and physical activity in the stratified models. As previously reported, the association was significant in Hispanic children speaking English and Spanish, first-generation Hispanic children, and second-generation Hispanic children, but not in non-Hispanic White children, Hispanic children speaking English or Spanish, and third- or higher-generation Hispanic children. Thus, mediation analysis was performed in Hispanic children speaking English and Spanish, first-generation Hispanic children, and second-generation Hispanic children, separately.
Table 6 and
Figure 3 present the estimates of the effects of the social environment on physical activity. As shown in
Table 6, the indirect effect estimates indicate a significant mediating role of neighborhood physical activity-related social norms for Hispanic children speaking English and Spanish (b = 0.33 [=0.30 × 1.09],
p < 0.001) and second-generation Hispanic children (b = 0.40 [=0.34 × 1.17],
p < 0.001). There was no significant mediation effect of neighborhood physical activity-related social norms for first-generation Hispanic children (b = 0.09[=0.33 × 0.29],
p > 0.05).
4. Discussion
The present study aims to extend current knowledge by investigating mediating and moderating mechanisms by which the neighborhood social environment is associated with physical activity among non-Hispanic White children and Hispanic children. In the full sample of non-Hispanic White children and Hispanic children, the social environment was not significantly associated with physical activity. However, the association in the full sample masked subgroup differences by individual-level cultural factors. We observed that Hispanic children speaking both Spanish and English, first-generation Hispanic children, and second-generation Hispanic children had greater physical activity when living in neighborhoods with high quality social environments (vs. neighborhoods with the lowest quality social environments). For example, Hispanic children speaking Spanish and English engaged in greater physical activity by 1.11 times a week (1.11 = −0.49 + 1.60) for each unit increase in the social environment. When living in neighborhoods with the highest quality social environment, Hispanic children speaking Spanish and English showed greater physical activity than non-Hispanic White children by 1.62 times a week (1.62 = −4.78 + [1.60 × 4 units]). Similar results were found in the interaction effect of immigrant generation in the association between the social environment and physical activity—first- and second-generation Hispanic children reported greater physical activity by 1.55 times a week (1.55 = −0.48 + 2.03) and 0.81 times a week (0.81 = −0.48 + 1.29) for each unit increase in the social environment, respectively. The results are aligned with prior research suggesting that informal support may be especially important for Hispanic children given that their families prefer to use informal support over formal support [
29]. Families of first- and second-generation Hispanic children and Hispanic children speaking Spanish and English at home may experience challenges with navigating and utilizing formal support outside their neighborhood [
21,
30,
31,
32]. Thus, living in health-promoting neighborhoods may provide more beneficial informal support for them in comparison to non-Hispanic White children and Hispanic children from English-speaking or third-generation families who may be more likely to use formal resources. Based on these findings, we suggest that community policies focusing on change in the quality of the social environment may have a spillover influence on increasing physical activity behaviors in first- and second-generation Hispanic children and Hispanic children speaking Spanish and English.
Our results of mediation analysis showed that neighborhood physical activity-related social norms mediate the association between the social environment and physical activity in Hispanic children speaking Spanish and English and second-generation Hispanic children (
Table 6 and
Figure 3). It is possible that visual cues of observing neighbors and peers physically active in their neighborhoods reduce fears of exercising and leads children to accept such active, healthy lifestyles as the norm. The findings suggest that health promotion policymakers need to acknowledge the protective effect of the social environment in physical activity among Hispanic children speaking Spanish and English and second-generation Hispanic children. Also, structured neighborhood physical activity programs may help to increase physical activity participation in Hispanic children speaking Spanish and English and second-generation Hispanic children.
On the other hand, we did not observe a statistically significant mediation effect for first-generation Hispanic children. The underpinnings of our observed non-significant mediation effect for the group are not clear; one possibility is due to a small sample size issue (n = 180 for first-generation Hispanic children). Also, neighborhood experiences may be different across the strength of ties to Hispanic culture. First-generation Hispanic children who still hold stronger ties to Hispanic culture may have different neighborhood experiences according to the proportion of Hispanic population in the neighborhood, which was not adequately captured in this study. Future research should examine additional features of the neighborhood environment (e.g., racial/ethnic composition, social cohesion, social support, built environment) and if specific features mediate the effect of the social environment on physical activity for first-generation Hispanic children.
Important strengths of the present study include geographical diversity in the sample. In addition, Hispanic communities and families were oversampled, which led to a large number of Hispanic children in the study (1538 out of 2749 children). The main strength of our study is the investigation of the mediating and moderating mechanisms by which the neighborhood social environment is associated with physical activity, which have not been empirically tested. The potential public health significance is great given the large ethnic disparities in child physical activity in the US and the serious potential consequences of physical inactivity on child health. Our study is also subject to notable limitations. We used a cross-sectional design which prevents us from inferring causality. Due to the cross-sectional design, there is a possibility of reverse causation in which those physically active tend to choose to live in neighborhoods with high quality social environments. Longitudinal designs would elucidate the causal relation and capture the length of exposure to the neighborhood social environment and changes in the neighborhood social environment caused by residential mobility. Also, study findings are not generalizable to all US children because the HCS is not a nationally representative study despite being national in scope. Physical activity was measured using the PABR-7 instrument. The PABR-7 instrument was designed for the HCS and has been used in associational studies [
36,
38], but the psychometric properties have not been established [
34]. Although we excluded any item explicitly asking about physical activity at school, whether physical activity occurred in a child’s residential neighborhood was not measured. In addition, our study did not include individual-level psychological factors such as self-efficacy, motivation, and social desirability [
39]. Finally, there are other factors that were not included in this analysis that may affect physical activity, such as parents’ physical activity, spending time with family, psychological stress, language barriers, acculturation, enculturation, and discriminative experiences [
40,
41]. Further research should explore these various characteristics as determinants of physical activity.