Home, School, and Neighborhood Environment

Pediatric Exercise Science, 2011, 23, 487-503 © 2011 Human Kinetics, Inc.

Home, School, and Neighborhood Environment Factors and Youth Physical Activity Rachel A. Millstein SDSU/UCSD Joint Doctoral Program in Clinical Psychology

Joe Strobel San Diego State University

Jacqueline Kerr University of California, San Diego

James F. Sallis San Diego State University

Gregory J. Norman University of California, San Diego

Nefertiti Durant University of Alabama – Birmingham School of Medicine

Sion Harris Harvard Medical School/Children’s Hospital Boston

Brian E. Saelens Seattle Children’s Hospital Research Institute and University of Washington This study examined the contributions of home, school, and neighborhood factors related to youth physical activity (PA). Adolescents (ages 12–18; N = 137) and parents of younger children (ages 5–11; N = 104) from three US metropolitan areas completed surveys. Youth PA was estimated from six items assessing overall physical activity. Bivariate analyses between environment factors and PA Millstein is with the SDSU/UCSD Joint Doctoral Program in Clinical Psychology, San Diego, CA. Strobel and Sallis are with San Diego State University, San Diego, CA. Kerr and Norman are with the University of California, San Diego, San Diego, CA. Durant is with the University of Alabama – Birmingham School of Medicine, Birmingham, AL. Harris is with Harvard Medical School/ Children’s Hospital, Boston, MA. Saelens is with Seattle Children’s Hospital Research Institute and University of Washington, Seattle, WA.     487

488   Millstein et al.

determined variable selection for adjusted hierarchical regression models. There were significant correlations in each environmental setting for adolescents (r’s: 0.16–0.28), but for parents of children, only for the home and neighborhood settings (r’s: 0.14–0.39). For adolescents, pieces of equipment at home, family recreation membership, equipment at school, and neighborhood aesthetics explained 15.8% of variance in PA. For younger children (based on parent report), pieces of equipment at home, neighborhood traffic safety, walking/cycling facilities, and street connectivity explained 21.4% of the variance in PA. Modifiable factors like increasing access to equipment at home and school, and improving neighborhood aesthetics may impact youth PA. To optimize explanation of youth PA, factors from multiple environments need to be considered.

A majority of young people in the United States do not meet current physical activity recommendations (46) resulting in a public health challenge that requires understanding and intervention on multiple levels of influence (12). Ecological models are helpful in conceptualizing the factors that can affect youth physical activity (31,37). Such models include levels of influence from individual to social to neighborhood and policy factors, with opportunities to intervene at each level (31,37). For youth, physical activity appears to be influenced by multiple environments, including homes, schools, and neighborhoods (10,20). However, many of the environmental factors are less-well understood for youth than for adults (3). The home environment can shape how youth spend leisure time. For example, many studies have shown that having a TV in the home or in a child’s or adolescent’s bedroom is associated with increased TV time, reduced physical activity, and likelihood of being overweight (1,8,9,34,49), although not all find these associations (41). Home environments that can promote physical activity include the presence of sport and play equipment, recreation facilities, and yard space (40). The presence of and access to physical activity equipment in and around the home is associated with greater child activity (47). Much of child play time occurs in yards (50), and time spent outdoors is correlated with physical activity (39), so outdoor play space is important for physical activity. These studies suggest that the availability of safe and convenient home-based recreation opportunities promote youth physical activity. Within the school environment, high quality physical education is a key factor for enabling and promoting youth physical activity (11,22,26,35). Important components of the school environment also include availability of school play equipment, facilities, after-school programs, and after-school availability of fields and courts (2,35,43,51). Larger campuses and access to play areas provide more opportunities for school physical activity (6). The relationships between physical education, facility and equipment availability, and school and overall physical activity appear to be consistent, and these factors are potential points of environmental modification. Neighborhood environments can also influence youth physical activity. Within the recreation domain, access and proximity to parks, playgrounds, and recreation spaces are associated with youth physical activity (7). For children targeted for decreased sedentary time, physical activity increases were positively associated with park proximity (13). Among community design factors, the presence of sidewalks, destinations within walking distance, low intersection density, and few road traffic hazards are associated with greater physical activity among youth (7). Transportation infrastructure, sidewalks, crosswalks, and traffic control features,

Environment Factors and Youth Physical Activity    489

such as traffic lights and controlled crossings are especially important for influencing whether children walk or bike to school (4). Local conditions including crime and weather have shown inconsistent relationships with physical activity (7,18). Characteristics of the home, school, and neighborhood environments are related to youth physical activity, but most studies include measures of only one type of environmental setting. While one study did examine home, school, and neighborhood factors, it focused only on adolescents (20). The present study includes data from two age groups: adolescents (using self-report data) and children (using parent-report data), and an overall physical activity index, for a more comprehensive view of youth physical activity. The goal of the current study was to simultaneously examine selected physical characteristics of these three environments and determine the independent contribution of each to understanding the variability in youth physical activity.

Methods Participants and Recruitment Data were collected as part of a larger measurement development study assessing neighborhood environments, physical activity, sedentary behaviors, and nutrition behaviors among children and adolescents (11,16,19). Three groups of participants were recruited: adolescents, parents of adolescents, and parents of children. The data from the parents of adolescents were not included in the present analyses. The first group was 189 adolescents between ages 12 and 18; the 137 with complete data were included in this study (mean age: 14.57 ±1.69). The second group was 116 parents of children ages 5–11, and 104 with complete data were included in this study (mean age: 8.27 ± 1.89). Adolescent participants were recruited from three US metropolitan areas: San Diego, CA, Boston, MA, and Cincinnati, OH. Parents of children were recruited from San Diego, CA and Cincinnati, OH only. Recruitment was designed to draw a sample from diverse neighborhoods with respect to income, race/ethnicity, walkability, and region of the country. Recruitment strategies varied by region and included phone, mail, and in-person contact at schools, recreation centers, and community centers. Specific recruitment details have been published (16). Human subjects approval was obtained from the appropriate institutions in each city. Parent participants gave written consent, and adolescent participants gave written assent.

Procedures Participants completed surveys on two occasions in 2005–2006, separated by 2–4 weeks to evaluate test-retest reliability. Surveys were given or mailed to participants. The surveys took 30–45 min to complete, and participants received $20 after completing the second survey. The average time between completion of the two surveys was 27 days. The average response rate for the first survey across study sites was 47%, with 74% of parents and 62% of adolescents completing both surveys. Only responses from the first survey were used in the current study, with the exception of test-retest reliability analyses. Parents of children answered the survey on behalf of their children and adolescents provided self-reports. For further details on survey administration, see (16,25).

490   Millstein et al.

Item Development The survey items were developed based on previous research (21,34,36,50), existing measures (24), and through formative interviews with parents and children in Cincinnati and San Diego. Participants in the formative research were not part of this larger study (25).

Measures Demographic Variables.  The demographics assessed were site, child or adolescent gender, child or adolescent race/ethnicity, and parental education. Race/ethnicity was categorized as non-Hispanic White or not, and parental income was dichotomized as earning less than $50,000 per year, or earning $50,000 per year or more. These variables were selected based on previous research in this sample (11). Home Environment.  Table 1 describes the items and test-retest reliability for

the environmental scales. The first scale consisted of 14 items that asked about the availability of sports equipment such as bike, basketball hoop, jump rope, and sports equipment (adapted from (36)). The overall scales’ ICCs were 0.69 for adolescents and 0.80 for parents of children (33). The second scale was constructed from two yes/no items asking about family membership in recreation facilities including health clubs and swimming pools. Reliability for this scale was ICC = 0.66 or above. For both scales, items were summed; higher scores indicate more availability or membership.

School Environment.  The availability of school physical activity spaces or equipment was assessed using a scale of six items, including basketball hoops, soccer goal posts, and indoor exercise machines (Table 1). A sum of items was used to create the school physical activity equipment scale, and reliability was ICC = .69 for adolescents and parents of children (11). Higher scores indicate higher availability. Neighborhood Built Environment.  The Neighborhood Environment Walkability

Scale for Youth (NEWS-Y) was used to assess neighborhood attributes believed to be related to physical activity. The NEWS-Y is organized into nine subscales that have been shown to be reliable and have support for concurrent validity (Table 1; 32). For the pedestrian and automobile traffic safety, crime safety (reverse scored), neighborhood aesthetics, walking and cycling facilities, street connectivity, land use mix-access, and residential density subscales, items were averaged and higher scores indicate better qualities.

Physical Activity Outcome.  A physical activity index was created from six items

that assessed overall physical activity, school physical education, participation in teams and classes, and active transport to and from school. The use of the index was deemed appropriate for the purpose of assessing the total environment as it relates to total physical activity. Two of the six items were used to determine whether participants met physical activity guidelines (60 minutes of activity 5 days per week; 48). Participants were asked, “For the past seven days, how many days were you/ your child physically active for a total of at least 60 minutes per day?” and “Over a typical or usual week, on how many days are you/your child physically active for

    491

20 (1 = 1–5min, 5 = 31+ min.)

7 (1 = strongly disagree, 4 = strongly agree)

6 (1 = strongly disagree, 4 = strongly agree)

Land use mix-diversity

Pedestrian and automobile traffic safety

Crime Safety

Neighborhood Environment3

School physical activity equipment

6 (0 = not available or don’t know, 1 = available)

2 (0 = no, 1 = yes)

Family recreation membership

School Environment

14 (0 = not available, 1 = available)

# items (range)

Number of pieces of equipment available at home

Home Environment

Subscale

About how long would it take you to walk from your home to the nearest stores or places listed below: convenience store, supermarket, dry cleaners, library, etc. Destinations within a 10-min walk were summed. There is so much traffic along nearby streets that makes it difficult or unpleasant to walk in my neighborhood; The speed of traffic on most nearby streets is usually slow (30 mph or less) There is a high crime rate in my neighborhood; I am worried about being outside alone around my home because I am afraid of being taken or hurt by a stranger.

Do you have any of these at your school? Basketball hoops, soccer goal posts, running/walking track, etc.

Please tell us if you have the following items in your home, yard, or apartment complex: bike, basketball hoop, jump rope, etc. Do you have a family membership to a health club or gym? Do you have a family membership to a public, private, or community pool?

Sample items (wording from adolescent survey and scale creation)

0.73

0.67

0.87

0.692

0.66

0.691

(continued)

0.87

0.74

0.80

0.69

0.76

0.801

Test-retest reliability (ICC’s) Adolescents Parents of Children

Table 1  Descriptions and Test-Retest Intraclass Correlation Coefficients (ICCs) for the Subscales Within the Home, School, and Neighborhood Environment Measures. Reported for Adolescents and Parents of Children

492

4 (1 = there are none, 5 = all residences are this type)

14 (1–1-5min, 5 = 31+ min.)

Residential density

Recreation facilities

3All

neighborhood values from Rosenberg et al., 2009.

from Durant et al., 2009.

6 (1 = strongly disagree; 4 = strongly agree)

Land use mix-access

from Rosenberg et al., 2010.

3 (1 = strongly disagree; 4 = strongly agree)

Street connectivity

2Values

3 (1 = strongly disagree, 4 = strongly agree)

Walking/cycling facilities

1Values

# items (range) 3 (1 = strongly disagree, 4 = strongly agree)

Subscale Neighborhood aesthetics

Table 1 (continued) Sample items (wording from adolescent survey and scale creation) There are trees along the streets in my neighborhood; There are many interesting tings to look at while walking in my neighborhood. There are sidewalks on most of the streets in my neighborhood; Sidewalks are separated from the road/traffic in my neighborhood by parked cars. The streets in my neighborhood do not have many cul-de-sacs (dead end streets); There are many different routes for getting from place to place in my neighborhood. Stores are within easy walking distance from my home; From my home, it is easy to walk to a transit stop. How common are each of these housing types in your neighborhood? Separate or stand alone one family homes, connected townhouses or rows of houses, multiple family or duplex homes, and apartment or condo buildings. Weights were applied to each housing type to estimate density. About how long would it take you to walk from your home to the nearest recreation place listed below? Bike/hiking/walking trails, paths, swimming pool, large/small public parks. Responses within a 10 min walk were summed. 0.73

0.82

0.62

0.72

0.73

0.56

0.79

0.56

0.58

0.66

Test-retest reliability (ICC’s) 0.60 0.75

Environment Factors and Youth Physical Activity    493

a total of at least 60 minutes?” Organized physical activity outside of school was measured by two questions: “Not counting school PE classes, how many days per week do you/your child play or practice team sports?” and “Not counting school PE classes, how many days per week do you/your child have physical activity classes or lessons not in a team sport (like martial arts, dance, tennis)?” School physical activity was measured by the question, “How many days per week do you/your child have gym or physical education (PE) class at school?” The latter three questions used the same response scale: 0 days-5 or more days. Active transport to school was assessed using the question, “In an average week, how many days do you use the following modes of transportation to get to and from school?” Response choices ranged from 0 to 5 days. Walking and bicycling to and from school were considered as one total score. Each of these six items was converted into a z-score, and the physical activity index was created by summing the z-scores. The test-retest reliabilities of this index were: adolescents: ICC= .64 and children (based on parent report): ICC=.76.

Data Analysis Descriptive statistics were run for all variables, including means and SDs, and frequencies for categorical variables. Four categories of independent variables were analyzed: demographics, home environment, school environment, and neighborhood environment. Bivariate correlations between each independent variable and the physical activity index were used to select the home and neighborhood environment variables to enter into the full model (demographics and the school variable were forced to enter). Variables with correlations of p < .15 were included in the hierarchical linear regression models. This p-value was set to be inclusive, given the small sample sizes. A hierarchal linear regression model assessed the contribution of each level of influence on overall physical activity. The blocks were entered in the following order: demographics (block 1), home environment (block 2), school environment (block 3), and neighborhood environment (block 4). The site variable was entered as a fixed effect, with San Diego as the reference group for adolescents. Since parents of children were from only two sites, site was not dummy coded for this group. The adjusted R2 and change in adjusted R2 were reported to estimate the amount of variance in physical activity accounted for by each block. Partial correlations were reported for each variable in the regression model, to reflect that variable’s unique contribution, holding all others constant. The two-tailed alpha level for significance was set at .05 for the regression analysis. Regression assumptions for skewness and multicollinearity were checked and met. All analyses were carried out using SPSS version 17.0.

Results Descriptive characteristics for the two study groups are presented in Table 2. There were 137 adolescents and 104 parents of children in the sample.

Bivariate Associations Pearson’s correlations between the physical activity index and the home, school, and neighborhood variables are presented in Table 3. For adolescents, significant correlates were found for all three settings. These small to medium correlations

Table 2  Descriptive Characteristics of Adolescents and Parents of Children Adolescents (N = 137) N Percent

Parents of Children (N = 104) N Percent

Variable Site Boston Cincinnati San Diego

30 43 64

21.9 31.4 46.7

0 51 53

0 49 51

Race White (non-Hispanic) Non-White1

84 53

61.3 38.7

81 23

77.9 22.1

Gender (of youth) Male Female

65 72

47.4 52.6

50 54

48.1 51.9

Parent Education No college degree College degree

65 72

47.4 52.6

53 51

51 49

1Non-White

includes: Non-Hispanic Black, Hispanic, Asian/Pacific Islander, and other.

Table 3  Pearson’s Correlations Between Physical Activity Index Outcome and Home, School, and Neighborhood Variables and Scale Descriptive Statistics: Adolescents and Parents of Children

Variable Neighborhood Crime safety Residential density Street connectivity Walking/cycling facilities Neighborhood aesthetics Pedestrian and automobile safety Recreation facilities Land use mix-access Land use mix-diversity Home Number of pieces of physical activity equipment at home Family recreation facility membership School School physical activity equipment

Adolescents (N = 137) SD r Mean

Parents of Children (N = 104) Mean SD r

3.28 92.85 2.77 2.87 2.91 2.73

0.75 37.64 0.74 0.85 0.76 0.5

.08 .10 -.07 .11 .16** -.07

2.41 91.86 2.78 3.04 3.17 2.44

0.83 37.01 0.71 0.86 0.66 0.55

.08 .07 .39**** .21*** .11 .15*

5.45 3.08 7.45

3.26 0.57 5.93

.09 .03 .04

3.97 2.91 5.95

3.23 0.66 5.52

-.04 .05 .06

7.44

3.04

.28****

8.06

2.73

.14*

.24****

3.90

SD: Standard Deviation *p < 0.15. **p