Using mixed methods to understand women’s parenting practices related to their child’s outdoor play and physical activity among families living in diverse neighborhood environments

Abstract

A convergent parallel mixed methods design was used to understand parenting practices for outdoor play, their influence on adolescent’s physical activity and outdoor play and the role of the neighborhood and child’s sex. Adolescents (n=263) and their parents completed questionnaires and wore accelerometers. Parents (n=30) participated in in-depth interviews. Parenting practices were examined by neighborhood disadvantage and child’s sex in quantitative (Chi-square and T-tests) and qualitative (comparative thematic analysis) samples. Multi-level linear mixed models examined the associations between parenting practices and two adolescent outcomes: physical activity and outdoor play. Parents in high disadvantage neighborhoods and of female adolescents imposed more restrictions on outdoor play. Restrictive parenting practices were negatively associated with outdoor play, but not physical activity. Policy and environment change that improves neighborhood conditions may be necessary to reduce parents’ fear and lessen restrictions on outdoor play.

1. Introduction

Being physically active and playing outdoors is essential for adolescent development as these activities enable cognitive, physical, emotional and motor learning, as well as elicit benefits to prevent numerous chronic diseases such as obesity.^1–3 Yet, adolescents worldwide are not meeting physical activity guidelines. This may be due in part to today’s adolescents spending less time outdoors compared to previous generations ⁴. Time spent outdoors is positively associated with greater levels of overall physical activity, yet adolescents, especially those living in disadvantaged neighborhood environments, have limited opportunities for outdoor play.^5,6 Low-socioeconomic status (SES) neighborhoods traditionally have higher crime rates, lack access to physical activity resources (e.g., parks, sidewalks) and have poor social capital ⁷. These built and social environment characteristics directly impact levels of physical activity but may also indirectly impact adolescents’ physical activity behaviors by influencing a parent’s desire or ability to let their child be outdoors.^8–12

Parents, particularly mothers, play a major role in the health-related behaviors of their children. Parental support, parenting styles and parenting behaviors have been shown to influence adolescents’ physical activity, yet parenting rules for outdoor play have rarely been studied.^13–18 Parents may support physical activity through encouragement, modeling or sharing in physical activity, or instrumental support (i.e., money for registration fees, transportation to activities, or equipment).^13,14 However, parents may also limit their child’s ability to be physically active by restricting outdoor play with rules regarding where their child can play outdoors (e.g., restriction to the backyard only), at what time of day they are able to play outside (e.g., during daylight hours only), and with whom their child can play or by whom they are supervised. Parents’ rules have been categorized as avoidance and defensive behaviors depending on whether outdoor play was avoided or modified, respectively.¹⁹ Avoidance behaviors eliminate an adolescent’s physical activity behavior entirely, such as parents driving their adolescent to school instead of allowing them to walk or forbidding unsupervised outdoor play. Defensive behaviors generate modifications like requiring parental accompaniment while walking to school or restricting outdoor play to the backyard. Rules for outdoor play may be more relevant during adolescence (10–16 years of age) when children gain the capability to play independently and/or without adult supervision than younger ages when children rely on parenting encouragement and modeling of physical activity.²⁰ Despite new found autonomy, there are significant decreases in physical activity with increasing age, especially among female adolescents.²¹ Low levels of physical activity among female adolescents may be contributed to by parents being more protective of females than males, resulting in disproportionate opportunities for independent outdoor play.^19,22 Understanding these rules, and differences by child’s sex, is important for adolescent health as the ability to play unsupervised in their neighborhood environment or actively commute to school independently results in more physical activity.^22,23

In addition to child sex, parenting behaviors may be influenced by neighborhood factors. Given the evidence linking neighborhood conditions and adolescent physical activity,^8–12 it can be hypothesized that parental rules regarding outdoor play may be an important mediator of this relationship. An Australian study found that parents were more likely to exhibit constrained behaviors for adolescents if they perceived their child was at risk for harm in their neighborhood.²⁴ However, the role of parental constrained behaviors has not been examined among adolescents in the United States living in neighborhood environments that vary in levels of disadvantage or using mixed methods. Neighborhoods with greater disadvantage, defined using an index comprised of markers of poverty (e.g., % of households below the poverty level, % receiving public assistance, % with less than a high school degree) from United States census data, traditionally present risks and lack resources that may result in parenting practices that differ from those who live in low disadvantage environments.

The examination of these relationships is supported by ecological models of behavior which posit that macro-level environmental factors (e.g., neighborhood disadvantage) may have a direct relationship with adolescents’ outdoor play and physical activity but may also indirectly influence adolescents’ behavior through interpersonal relationships (e.g., parenting practices). It is posited that ecological levels (neighborhood disadvantage may influence parenting practices) may influence individual behavior (adolescent physical activity and outdoor play).^25,26 This study fills three critical gaps in the literature using mixed methods to understand: 1) parenting practices for outdoor play 2) their influence on adolescent’s physical activity and neighborhood outdoor play; and 3) the role of the neighborhood environment and child’s sex.

2. Materials and Methods

2.1. Overview

A convergent parallel mixed methods design²⁷ was implemented to gain a comprehensive understanding of the role of parenting rules/decisions for male and female adolescents’ physical activity and neighborhood outdoor play in low and high disadvantage neighborhoods. Quantitative and qualitative data were collected during the same phase of research and aimed to collect different yet complementary data to facilitate triangulation and enrich the interpretation of results. Qualitative methods were necessary due to the limited amount of research and lack of in-depth understanding of parenting rules/decisions for adolescents’ outdoor play. All participants (n=342) were first enrolled in the Translational Investigation of Growth and Everyday Routines in Kids (TIGER Kids) Study (USDA 3092–51000-056–04A) in which they completed close-ended questionnaires, wore accelerometers to measure physical activity, and completed anthropometric measurements. All questionnaires/measures were administered by trained pediatric research staff and quantitative anthropometric and survey data were managed in REDCap.²⁸ Parents (n=30) of TIGER Kids participants were then purposively recruited by neighborhood type (low and high disadvantage) and race (African American and White) to participate in semi-structured in-depth interviews. The qualitative and quantitative data were analyzed independently and then integrated during the interpretation phase.

2.2. Participants

Adolescents aged 10 to 16 years were recruited from the greater Baton Rouge area to participate in TIGER Kids and complete quantitative assessments. Consent/assent was obtained upon each participant’s arrival at the Pennington Biomedical Research Center (PBRC) pediatric research clinic before participating in any study procedures. Children were eligible for TIGER Kids if they were between ages 10 and 16 years at their baseline visit, weighed <500 pounds, were not pregnant, were not on a restrictive diet due to illness, and did not have significant physical or mental disability.

Parents of TIGER Kids study participants (n=30) participated in in-depth interviews. The neighborhood environment of participants was characterized a-priori by geocoding each participant’s home address using geographic information systems (GIS) to ensure diverse neighborhood types were represented in the recruited sample. The census block group around each participant’s home address was characterized as low or high disadvantage using the Concentrated Disadvantage Index (CDI), which included percent of individuals below the poverty line (calculated based on family size and age of members) ²⁹, percent of households receiving public assistance, percent female-headed households, percent unemployed, percent less than age 18, and percent African-American.^30,31 A z score was calculated and the median split was used to determine low and high disadvantage for the TIGER Kids study sample. Ethical approval for this study was provided by the PBRC Institutional Review Board, and parents provided their written consent prior to being interviewed.

2.3. Quantitative Data Collection

Sociodemographics and anthropometrics.

Parents reported child’s age, sex, race, ethnicity, and annual household income. Height was measured using a stadiometer after the participant removed shoes. Weight was measured using a digital scale while the participant wore a hospital gown and undergarments. BMI percentile was calculated based on the participant’s age, height, and weight using the CDC 2000 growth charts. Weight status (normal weight, overweight, obese) was categorized using CDC definitions for BMI percentile.³²

Parenting practices for outdoor play (constrained behaviors)

were measured quantitatively using a validated questionnaire for measuring parental constraint on adolescents’ outdoor play.¹⁸ Carver and colleagues (2010) defined constrained behavior as the act of restricting adolescent’s physical activity, which is categorized as ‘avoidance,’ ‘defensive,’ or ‘neither,’ depending on whether physical activity was avoided or modified, respectively.¹⁹ Avoidance and defensive behaviors were measured by asking parents/guardians to assess their level of agreement with 14 statements (7 for avoidance and 7 for defensive). Response options are presented individually and summed to compute individual scores for avoidance and defensive behavior. Response options and subsequent coding (in parentheses) were: strongly disagree (−2); disagree (−1); neither/don’t know (0); agree (1); strongly agree (2). All responses were summed to compute overall avoidance (possible range: −14 to 14; Cronbach’s alpha=0.77) and defensive behavior (possible range: −14 to 14; Cronbach’s alpha=0.73) scores, with a higher number indicating more avoidance or defensive behavior, respectively. Carver et al. (2010) found this survey to have moderate to high internal consistency for each score (Cronbach’s alphas: avoidance behavior, 0.79; defensive behavior, 0.62) and moderate test-retest reliability (intraclass correlation ≥0.44) when used in adolescents 10–17 years of age.¹⁹

Physical Activity

was measured as time spent in moderate-to-vigorous physical activity (MVPA) using accelerometry. A triaxial accelerometer was worn on an elasticized belt on the hip (Actigraph GT3X+, Actigraph, Ft. Walton Beach, FL). Participants were encouraged to wear the accelerometer 24 hours/day for at least 7 days, including 2 weekend days.³³ Data from the nocturnal sleep period were removed based on a fully automated, validated algorithm.³⁴ Data were included if there was a minimum of 4 days with at least 10 hours of awake wear time per day, including at least one weekend day.³⁵ Evenson et al. cut points were used to determine amount of MVPA (≥574 CPE).³⁶

Neighborhood Outdoor Play

was measured using a parent-reported questionnaire adapted from the Neighborhood Impact on Kids survey.³⁷ Parents were asked how often on a 6-point Likert-scale (never [0] to 4 times per week or more [5]) during the past year their child had been physically active in the following places: 1) in your yard or common areas or in your driveway; 2) at a neighbor’s house, yard or driveway; and 3) in a local street, sidewalk, or vacant lot/field. Responses were summed to create an overall score for neighborhood outdoor play (possible range: 0–15; Cronbach’s alpha: 0.74).

2.4. Qualitative Data Collection

Parents participated in semi-structured in-depth interviews in a private room in the clinic at PBRC by trained qualitative researchers. Interview questions were focused on parenting decisions/rules around outdoor play for their child who participated in the TIGER Kids study. Questions were informed by themes in Carver’s constrained behavior questionnaire.¹⁹ Parents were instructed to answer questions for the TIGER Kids participant only. The interview guide was generated and refined by the research team and by piloting the interview guide among two employees of PBRC who each had a child between the ages of 10 and 16 years. Broad, open-ended questions were posed, followed by more specific probing questions to gain detailed data about topics mentioned by the participants. The interviewers used reflexive practice journaling of thoughts, impressions, and potential biases after each interview. The two interviewers discussed their thoughts on a weekly basis in an effort to reduce potential biases and make amendments to the interview guide (e.g., change the wording of a question) that would generate a better understanding of key concepts.³⁸ Interviews lasted an average of 40 minutes, ranging from 19 to 61 minutes.

2.5. Data Analysis

Quantitative Data Analysis

Means (standard deviations) and frequencies (percentages) were computed for variables by neighborhood type for participants with complete data for all variables in the analysis (n=263). Differences between low and high disadvantage neighborhoods and females and males within each neighborhood type were examined using Chi-square and T-tests. Multi-level linear mixed models (PROC MIXED) were used to examine the associations between parent constrained outdoor play practices and adolescent’s physical activity (average minutes per day spent in MVPA) and neighborhood outdoor play. Four separate models were run to examine the relationship between each parental constrained behavior (avoidance or defensive behavior) and each outcome (MVPA or neighborhood outdoor play). Households were treated as random effects to account for siblings within the same household. Models were adjusted for variables that remained significant in the full regressions model: age (categorized as <13 or ≥13 y), child’s sex, child weight status (obese vs. not obese) and whether study assessment occurred during the school year or not, as well as any two-way interactions between adjustment variables and age-by-parental behavior where indicated (p<0.2). Two-way interactions between parental behavior and the hypothesized moderators of neighborhood type and child sex were examined and reported. All data management and statistical analyses were conducted using SAS version 9.4. Only participants with complete data for all demographics, parenting practices (constrained behaviors), and outcomes were included in the analysis. The analysis sample (n=263, 77% of full sample) did not differ from the full sample (n=342) in any of the included variables. The significance level was set at P<0.05 and there was no adjustment for multiple comparisons.

Qualitative Data Analysis

All interviews were audio-recorded and transcribed verbatim by professionals. Interviews were anonymized prior to analysis. Data analysis was performed in NVIVOv11 and was guided by principles of thematic analysis.^39,40 Thematic analysis was chosen because the goal of the study was to identify common decisions/rules parents exhibit regarding outdoor play in the neighborhood. Guidelines by Hennink and colleagues (2011) were followed when collecting and analyzing the data.³⁸ First, four diverse participant interviews (low and high disadvantage, African American and White, male and female child) were selected, annotated and discussed by two trained staff members to establish an initial codebook.⁴¹ Focused coding⁴¹ and the constant comparison method⁴² were used to generate inductive codes. In addition to inductive codes, previous research on constrained behavior and independent mobility and topics from the interview guide were used to create deductive codes. All coding discrepancies were discussed until an initial codebook with detailed code definitions was agreed upon. An additional interview was coded by the two staff members to refine the codebook and inter-rater reliability was calculated. This process was repeated until saturation was met (after 8 interviews). At this point, all remaining interviews were coded. Two raters independently coded six of these interviews to calculate inter-rater reliability. Across all codes in these six interviews (n=25 codes), raters had an agreement of 95.0%. Once all interviews were coded, codes were compared between participants grouped in low versus high disadvantage neighborhoods and child sex to determine if differences in responses were present.

Quantitative and Qualitative Data Integration

Carver’s constrained behavior questionnaire¹⁹ was administered quantitatively to all parents of adolescents participating in the TIGER Kids study (n=342). Qualitative data collection (interviews) and analysis were informed by themes in the quantitative survey. Interview guides were developed to elicit in-depth information on parent’s avoidance and defensive behaviors. Transcripts were coded deductively to identify themes from the quantitative survey and inductively to identify new themes presented by interviewees. All themes were interpreted within the context of Carver’s larger categories that classify parenting behaviors as avoidance and defensive behaviors (Table 3).

Table 3:

Qualitative results: parent quotations illustrating avoidance and defensive behaviors by neighborhood type and child gender.

Themes and Sub-themes	Low Disadvantage (n=14)		High Disadvantage (n=16)
	Female (n=9)	Male (n=5)	Female (n=8)	Male (n=8)
Avoidance Behaviors	“I don’t want her around going just by herself because I don’t want anyone—I don’t trust people too much. It’s not only our people in the neighborhood. It’s people riding around.” – African American mother (themes: supervision, collaboration) “There is a Shell station and a convenience store across (street name), but our rule is, “You’re not allowed to cross (street name) on your own.” They’ve never been able to have that freedom to go do that.”- white mother (theme: location of play)	“He’s not allowed near the road, even at 16. We have hired a lawn service to cut the ditch. I don’t even want him out there near the road with a lawnmower.”- White mother (theme: location of play)	“I have to be present or at least within distance of eyeshot where I can see them.”- African American mother (theme: supervision) “I don’t let her get outside to play. As far as—I think that’s all. They don’t have anything for the kids to play with out there. I don’t think they have a basketball goal or anything like that.”- African American mother (themes: type of play)	“The street lights is his curfew. So that means it’s dusk and that means you need to be in the home.” –African American mother (theme: time of day) “We’ve never been [to the park] though because I wouldn’t let him go but… there’s one maybe three or four blocks.”-African American mother (theme: location of play)
Defensive Behaviors	“Even though she’s 16 she may walk (dog’s name), but she has to stay on our street unless I’m walking with her, or unless [her brothers] or somebody’s with her.”- African American mother (themes: supervision, location of play, peers and siblings) “Now that my kid is older, I’m more comfortable with [her riding bikes in the neighborhood]. When she was younger, you worry a little bit about her watching out. Although, when they were younger, the rules were they needed to be on the sidewalk to ride their bikes. Now I’m okay with them in the street.” (themes: location of play, type of play)	“It wasn’t like we felt like there had to be a parent overseeing everything, unless maybe there was a pool or some sort of a obviously dangerous situation. They were allowed a pretty decent amount of freedom to roam around, to play outside. We would always just keep track. If you hadn’t seen—if I hadn’t seen (child’s name) in a couple hours, I would start calling the neighbors saying, “Is he at your house? Is he at your house?” I guess we had felt a good bit of freedom.”- white mother (themes: supervision, collaboration) “If he’s in our yard, he can go anywhere. I trust him. If he’s going to the neighbor’s house, I just ask that he let me know. If he is going to the trail, I make sure he has someone with him at all times or is meeting someone to be with him. … if he’s walking to school or going to the trail or anything like that, they have to stick together.”- white mother (themes: supervision, peers, location of play).	“If they leave on their bike, they can go around the block, come in, check in, and then go again. So it’s like that, unless I’m with them and then we would go a farther distance.”- White mother (themes: location of play, supervision) “I tell her to stay close to the house. Always bring your phone with you so you can call me if something happens. She can go to the mailbox. I don’t mind if she walks to the mailbox, but she canť go out after dark. That’s about it.”- African American mother (themes: location of play, time of play)	“[He can go] maybe two blocks. But it’s the same block … I could walk out on the sidewalk and see him because he walks [the dog] sometimes without me going.”- white mother (themes: supervision, location of play) “The people that I know and I can trust them, then he can play with. The people that I donť know and donť, you know, that I havenť built or established a relationship of trust with, then no, he canť—they canť supervise and he canť play with their kids, unless I'm there or his dad is there.”- African American mother (themes: supervision, collaboration) “I do allow him to play a little bit more with his friends on a time schedule. “You get 30 minutes and you come back in,” or maybe it’s an hour and come back in, but I do allow him more time out, yeah.”-African American mother (theme: peers)

3. Results

3.1. Quantitative Results

On average adolescents were 12.5 (± 1.9) years of age, and approximately half of participants were female (n=140, 53.). Thirty-five percent (n=92) of the adolescents were African American and 34.6% (n=91) had obesity. On average, adolescents engaged in 34.2 (± 19.6) minutes of MVPA per day.

Child Demographics by Neighborhood Type

There were significantly more African Americans (p<0.001) and participants with obesity (p=<0.001) living in high disadvantage neighborhoods than low disadvantage neighborhoods. Seventy five percent (75%) of adolescents in high disadvantage neighborhoods had obesity versus 7% in low disadvantage neighborhoods (p<0.001). Participants in high disadvantage neighborhoods had lower annual household income than participants in low disadvantage neighborhoods (p=0.01). On average, parents in high disadvantage neighborhoods reported more defensive behaviors than parents in low disadvantage neighborhoods (p=0.01). Avoidance behaviors did not differ between parents in low and high disadvantage neighborhoods (p=0.2). The amount of MVPA (p=0.20) and parent-reported neighborhood outdoor play (p=0.88) did not differ between adolescents living in low and high disadvantage neighborhoods.

Parenting Practices by Neighborhood Type and Child’s Sex

Differences in avoidance and defensive behavior scores by neighborhood type and child’s sex are presented in Table 2. There were no significant differences between parents in low (n=132) and high disadvantage (n=131) neighborhoods for any of the seven items measuring avoidance behaviors. Five of the seven survey items measuring defensive behaviors were significantly higher for parents in high versus low disadvantage neighborhoods. More parents in high disadvantage neighborhoods agreed with the following defensive behaviors: 1) my child must be supervised while playing outside (p=0.02); 2) I prevent my child from playing outdoors in our neighborhood unless supervised (p=0.02); 3) I prevent my child from walking/cycling in our neighborhood unless supervised (p=0.01); 4) my child is learning/has learnt self-defense skills (p=0.04); and 5) my child carries a whistle/alarm to ward off unwelcome strangers (p=0.02). The majority of differences by child’s sex were within low disadvantage neighborhoods. Significantly more parents of females in low disadvantage neighborhoods than parents of males reported defensive behaviors (3 of the 7 defensive behaviors significantly differed).

Table 2:

Parenting practices (avoidance and defensive behaviors) by neighborhood disadvantage and children’s sex (N=263)

	Agree N(%)1
	Low Disadvantage				High Disadvantage				High vs. Low Disadvantage
	Total (n=132)	Female (n=71)	Male (n=61)	P-value	Total (n=131)	Female (n=69)	Male (n=62)	P-value	P-value
Avoidance Behaviors
I prevent my child from playing alone outdoors in our neighborhood.	32(24.4)	19(25.4)	14(22.9)	0.75	29(22.1)	21(30.4)	8(12.9)	0.03	0.56
I prevent my child from playing with friends outdoors in our neighborhood.	7(5.3)	6(8.4)	1(1.6)	0.08	9(6.9)	8(11.5)	1(1.6)	0.08	0.31
I don’t allow my child to spend time outside after dark.	82(62.1)	49(69.0)	33(54.1)	0.08	93(71.0)	51(73.9)	42(67.7)	0.46	0.16
I don’t allow my child to walk/ride a bike on the street after dark.	102(77.2)	58(81.7)	44(72.1)	0.22	113(86.3)	61(88.4)	52(83.8)	0.49	0.16
I prevent my child from playing alone on local streets in our neighborhood.	65(49.2)	43(60.6)	22(36.1)	0.01	71(54.2)	42(60.9)	29(46.8)	0.18	0.72
I prevent my child from playing with friends on local streets in our neighborhood.	26(19.7)	18(25.4)	8(13.1)	0.18	34(26.0)	23(33.3)	11(17.7)	0.08	0.42
I prevent my child from walking/cycling with friends in our neighborhood.	13(9.8)	11(15.5)	2(3.3)	0.06	23(17.7)	14(20.3)	9(14.8)	0.58	0.18
Defensive Behaviors
My child must be supervised while playing outside.	27(20.5)	19(26.8)	8(13.1)	0.05	46(35.1)	30(43.5)	16(25.8)	0.07	0.02
I prevent my child from playing outdoors in our neighborhood unless supervised.	23(17.4)	17(53.8)	6(9.8)	0.03	41(31.3)	28(40.6)	13(21.0)	0.04	0.02
I prevent my child from playing on local streets unless supervised.	55(41.7)	35(49.3)	20(32.8)	0.10	72(55.0)	41(59.4)	31(50.0)	0.35	0.10
I prevent my child from walking/cycling in our neighborhood unless supervised.	34(25.7)	25(35.2)	9(14.8)	0.01	55(42.0)	33(47.8)	22(35.5)	0.14	0.01
My child is learning/has learnt self-defense skills.	47(35.6)	25(35.2)	22(36.1)	0.92	62(47.3)	34(49.3)	28(45.2)	0.31	0.04
My child carries a whistle/alarm to ward off unwelcome strangers.	5(3.8)	5(7.0)	0(0.0)	0.07	10(7.6)	4(5.8)	6(9.7)	0.69	0.02
I’m always reminding my child about road safety.	114(86.4)	65(91.5)	49(80.3)	0.06	112(85.5)	61(88.4)	51(62.0)	0.42	0.60

¹Percentages are reported as the % of individuals who agreed (sum of those who agreed [1] or strongly agreed [2]) within each category (females and males within low and high disadvantage neighborhoods). For example, within low disadvantage neighborhoods, 18 parents of females reported that they agreed to the statement, ‘I prevent my child from playing alone outdoors in our neighborhood.’ The percentage (25.4) is reported as 18 divided by 71(the number of female children living in low disadvantage environments).

Associations between Parenting Practices and Physical Activity (MVPA)

Parents’ avoidance (β= 0.001, p=0.99) and defensive behaviors (β= −0.22, p=0.30) were not associated with minutes of MVPA. Associations between parents’ avoidance behaviors and MVPA did not differ by children’s sex (p=0.95) or neighborhood type (p=0.10). Similarly, associations between parents’ defensive behaviors and MVPA did not differ by children’s sex (p=0.25) or neighborhood type (p=0.09). Relationships did not differ for overall minutes of light, moderate and vigorous physical activity.

Associations between Parenting Practices and Neighborhood Outdoor Play

Parents’ avoidance (β= −0.09, p=0.02) and defensive behaviors (β= −0.10, p=0.03) were significantly associated with neighborhood outdoor play (Figure 1). Associations between parents’ avoidance behaviors and neighborhood outdoor play did not differ by neighborhood type (p=0.21) or children’s sex (p=0.28). Similarly, associations between parents’ defensive behaviors and neighborhood outdoor play did not differ by neighborhood type (p=0.40) or children’s sex (p=0.86). Overall and by neighborhood type and sex associations are presented in Figure 1.

Figure 1. Associations between Parenting Practices and Neighborhood Outdoor Play.

3.2. Qualitative results

Almost all participants (n=29, 96.7%) were mothers (average age of 45 years, range= 36–71), with one participant being a grandmother who lived with the child full-time. Half of the mothers (n=15, 50.0%) were African American, and sixteen (53.3%) lived in high disadvantage neighborhoods. Mothers were on average 45 years of age (range=36–71), and their adolescents were on average 13 years old (Table 1). Approximately half of the adolescents were female (n=17, 56.7%) and African American (n=15, 50.0%; Table 1). The qualitative sub-sample did not significantly differ from the full quantitative sample in any of the demographics, parenting practices (defensive and avoidance behaviors) or physical activity outcomes (Table 1).

Table 1:

Children’s demographics by sample and neighborhood disadvantage [N(%) or mean(std.)]

	Quantitative Sample				Qualitative Sub-sample				Sample to sub-sample comparison1
	Total (N=263)	Low Disadvantage (n=132)	High Disadvantage (n=131)	P-value	Total (n=30)	Low Disadvantage (n=14)	High Disadvantage (n=16)	P-value	P-value
Age, years	12.5(1.9)	12.6(1.9)	12.3(1.9)	0.16	12.7(2.2)	12.6(2.4)	12.8(2.0)	0.89	0.72
Female	140(53.2)	71(53.8)	69(49.8)	0.86	17(56.7)	9(64.3)	8(50.0)	0.43	0.93
African American	92(35.0)	28(21.2)	64(48.9)	<0.001	15(50.0)	4(28.6)	11(68.8)	0.03	0.21
Obese	91(34.6)	31(23.4)	60(45.8)	<0.001	13(43.3)	1(7.1)	12(75.0)	<0.001	0.54
Annual Household Income				0.01				0.02	0.74
<$40,000	62(23.6)	22(16.7)	40(30.5)	--	10(33.3)	2(14.3)	8(50.0)	--	--
$40,000- $89,999	56(21.3)	25(18.9)	31(23.6)	--	8(26.7)	3(21.4)	5(31.3)	--	--
>$90,000	128(48.7)	78(59.1)	50(38.2)	--	11(36.7)	9(64.3)	2(12.5)	--	--
Not Reported	17(6.5)	7(5.3)	10(7.6)	--	1(3.3)	0(0.0)	1(6.3)	--	--
Overall MVPA2	34.2(19.6)	32.7(17.5)	35.7(21.5)	0.20	32.7(12.7)	32.8(19.2)	32.6(23.5)	0.98	0.94
Neighborhood Outdoor Play3	7.3(4.3)	7.3(4.2)	7.25(4.4)	0.88	6.5(4.7)	7.4(3.9)	5.6(5.3)	0.31	0.86
Parent Avoidance Behavior4	−2.3(6.4)	−2.8(6.3)	−1.8(6.4)	0.22	−1.1(6.9)	−4.4(5.5) 5	2.8(6.4) 5	0.01	0.27
Parent Defensive Behavior5	−2.1(6.0)	−3.0(5.3)	−1.3(6.4)	0.01	−0.5(5.6)	−3.0(4.8) 5	2.5(5.0) 5	0.02	0.22

Bold Indicates a significant p-value defined as <0.05. Percentages do not add up to 100 due to rounding.

¹The sample is the 263 included in the quantitative analysis. The subsample includes the 30 who participated in qualitative interviews.

²Overall MVPA was measured as average daily minutes engaged in MVPA using a triaxial accelerometer.

³Neighborhood outdoor play was parent reports of how often on a 6-point Likert-scale (never [0] to 4 times per week or more [5]) during the past year their child had been physically active in the following places: 1) in your yard or common areas or in your driveway; 2) at a neighbor’s house, yard or driveway; and 3) in a local street, sidewalk, or vacant lot/field. Responses were summed to create an overall score for neighborhood activity (possible range: 0–15; Cronbach’s alpha: 0.74).

^4,5All responses were summed to compute avoidance (possible range: −14 to 14; Cronbach’s alpha: 0.77) and defensive (possible range: −14 to 14; Cronbach’s alpha: 0.73) behavior scores from a 5 point Likert scale (strongly disagree [−2] to strongly agree [2]) with a higher number indicating more avoidance or defensive behavior.

⁶22 parents (12 in low disadvantage and 10 in high disadvantage neighborhoods) who participated in qualitative interviews completed the constrained behavior (avoidance and defensive behavior) questionnaire. All other demographic data are reported on all 30 parents.

Parenting Practices for Outdoor Play

Major themes for parenting practices for outdoor play were: supervision, time of play, and location of play. Additional minor themes that were discussed to a lesser extent included: 1) collaboration among parents and neighbors, 2) peers and siblings and 3) type of play (e.g., bicycling, playing basketball, walking dog). The two types of parenting practices measured quantitatively, avoidance and defensive behaviors, were present throughout each of the qualitatively identified themes. Examples of avoidance and defensive behaviors provided by parents are presented in the following text and in Table 3.

Theme 1: Supervision

Rules and behaviors regarding adult supervision for outdoor play was the most discussed topic across parents from low and high disadvantage neighborhoods. Most parents did not allow their adolescent to play outside in the neighborhood without adult supervision from the child’s mother or father. Parents of females discussed requiring adult supervision in greater depth than parents of male adolescents. Some parents of adolescents who lived in low disadvantage neighborhoods gave their child freedom to play independently throughout the entire neighborhood. Yet, even in low disadvantage neighborhoods this freedom was not common and was never discussed among parents in high disadvantage neighborhoods. For adolescents who had freedom to play independently, it was required they ask for permission before they go anywhere and keep the parent informed about where they are at all times. Parents seemed to give a child with a mobile telephone more freedom to move around the neighborhood independently, which was more common among older adolescents and in low disadvantage neighborhoods.

Most parents exhibited defensive behaviors by allowing independent play in certain locations, with peers/siblings or with intermittent supervision (minor theme: peers). Overall, playing with peers or siblings increased adolescents’ independent mobility in the neighborhood and encouraged adolescents to play outside (minor themes: peers). However, parents did not like to leave adolescents who were playing in large groups unsupervised due to the potential for trouble between the kids (e.g., violence, bullying, and pranks). Parents who had fenced in back yards or a front yard that was far from the street often felt comfortable with their child playing in the yard without supervision. If they lived on a street with traffic, parents generally limited unsupervised play to the backyard. Overall, parents reported that the requirement for supervision limited their child’s outdoor play time as parents at other tasks (e.g., prepare dinner, help siblings with homework or clean the house). A few parents in high disadvantage neighborhoods who lacked flexibility in their schedule or worked long hours were less strict on their requirements for adult supervision and allowed their child to walk home from school/the bus stop and/or play outside without supervision.

Some parents mentioned that they collaborated with parents in the neighborhood to provide supervision for adolescents playing outside (minor theme: collaboration). It was rare, especially among high disadvantage neighborhoods, that parents reported working with and trusting other parents in the neighborhood to supervise their child when playing outdoors. Most parents had one or two neighbors they trusted to supervise their child outside, which was generally only when their child was playing in the neighbor’s yard rather than common areas (e.g., parks) in the neighborhood. Although most parents did not let neighbors be the lone supervisor for their child, some felt confident that their neighbors were looking out for their child when he/she was outside and would intervene in a dangerous situation, such as a stranger approaching their child. Parents in low disadvantage neighborhoods mentioned this confidence more than those from high disadvantage neighborhoods.

Theme 2: Time of Play

There were several parents who did not allow their adolescents to play outside during the weekdays mainly due to homework or because the family did not arrive home before sunset. Eliminating outdoor play during weekdays is an example of avoidance behavior. Parents from low disadvantage neighborhoods reported more frequent outdoor play in their neighborhood during the week than parents from high disadvantage neighborhoods. In both neighborhood types, parents limited outdoor play to daylight hours, which is an example of defensive behavior. A few parents from low disadvantage neighborhoods allowed their child to play with peers or siblings in their yard after dark (minor themes: peers). However, adult supervision was required for outdoor play in other neighborhood locations after dark. A small number of parents from high disadvantage neighborhoods exhibited defensive behaviors by limiting outdoor play to times when traffic volume in the neighborhood is lower (i.e., not during rush hours). In one case, this restricted the child’s ability to actively commute from the bus stop to home as the mother was concerned about traffic during this time. This defensive behavior was not reported by parents in low disadvantage neighborhoods.

Theme 3: Location of Play

All parents reported having rules for where their child could play independently or with peers/siblings in the neighborhood. It was uncommon for parents to exhibit avoidance behavior by not allowing their child to play in certain locations under any circumstance. Rather, parents commonly displayed defensive behaviors that required certain precautions (e.g., supervision, being in a group, communication) based on where the child was playing. In low disadvantage neighborhoods, adolescents had more freedom to play in locations outside of their yard or the sidewalk in front of their home. Yet, it was still rare for parents to allow their child to travel anywhere in the neighborhood or play in common areas (e.g., public park, basketball courts, swimming pools) without adult supervision. Furthermore, it was more uncommon for parents to allow female adolescents (vs. male adolescents) to travel independently outside of their yard. A few parents in low disadvantage areas allowed their adolescents to play with siblings or peers in common areas. In some cases these community areas had lifeguards or office staff present that parents felt would help in a dangerous situation (minor theme: collaboration). Furthermore, more parents in low disadvantage areas reported having play equipment in their yard (e.g., basketball goal, trampoline, soccer net) that promoted more regular proximal outdoor play.

Overall, adolescents in high disadvantage neighborhoods were limited to playing in their own yard, the sidewalk in front of their house, or not allowed to play outside at all and lacked community resources (e.g., parks, basketball courts, pools) that were more common in low disadvantage neighborhoods. However, among parents living in high disadvantage neighborhoods there were noticeable differences in rules based on differences in parents’ perceptions and personal history living in disadvantage environments. In particular, single mothers that recently moved to the neighborhood or described their neighborhood as lower SES than where they previously lived were more restrictive on where their child could play. Some parents, especially those who had grown up in the neighborhood or had a male child, felt that their child could fend for themselves and, therefore, allowed their child to travel freely in the neighborhood.

Parents in both neighborhood types reported rules that restricted the distance adolescents could travel from their home. For example, parents told their child they could go around the block or go as far as one of the child’s friends’ houses. Some parents allowed their child to travel further distances if they were bicycling rather than walking (minor theme: type of play). Other parents required kids to make a loop while bicycling so that they checked in at their house in regular intervals. Parents consistently reported that adolescents had to let them know where they were going, when they arrived and when they were leaving so that parents would know when to expect them home. In general, parents felt more comfortable with their child traveling further from their home if they were male or if they were with other children or siblings (minor theme: peers).

4. Conclusions

This study is the first to use mixed methods to gain a comprehensive understanding of parental restriction of outdoor play and how rules differ by neighborhood disadvantage and for male versus female adolescents. Understanding the intersection among neighborhoods, interpersonal factors (parenting practices) and individual behavior (adolescent physical activity and neighborhood outdoor play) is crucial for the health of adolescents and has been supported by ecological models of behavior change. Quantitative and qualitative results converged to indicate that parents enforce rules (e.g., distance from home; outside only during daylight hours) that vary by neighborhood and child’s sex. These rules resulted in less neighborhood outdoor play but did not influence overall physical activity measured objectively using accelerometry. This may be due to children acquiring physical activity outside their neighborhood. Today children belong to the ‘backseat generation’ in which parents’ chauffer children to structured leisure-time activities.⁶

Parents reported that they were more cautious with their female children compared to males, particularly in high disadvantage neighborhoods, due to safety concerns. However, the influence parenting behaviors had on neighborhood outdoor play or physical activity did not differ among females compared to males. This is contrary to a previous study, which found that parents’ avoidance and defensive behaviors were related to female children and adolescents’ active transport and MVPA outside of school hours, but not male adolescents’.¹⁹ The literature and parent interviews in the current study support that girls have less freedom to travel independently.^22,50,51 Furthermore, independent mobility and parenting behaviors have been identified as potential mediators between parental perceived neighborhood characteristics and physical activity for girls, but not boys.⁵² Carver found that parents who perceived their child was at risk of harm in their neighborhood were more likely to exhibit avoidance and defensive behaviors that translated into lower levels of MVPA and active commuting, particularly among girls.²⁴ Greater understanding of what factors generate fear among parents in low and high disadvantage neighborhoods is necessary to reduce parental restrictions on outdoor play for female adolescents. ^{37,46,53–55} The negative influence that parental restrictions have on females’ neighborhood outdoor play is particularly important as early adolescence is a time when females transition away from organized sport and physical activity participation.^56,57 Engaging in active travel and independent mobility becomes critical for adolescents who do not achieve sufficient physical activity through organized sport and may be a means to address disproportionately low levels of physical activity among female adolescents.^58,59

While many studies have reported the influence of disadvantaged neighborhoods on adolescent physical activity,^43–45 this study fills a critical gap by exploring the role of parenting practices in this relationship. Parents in high disadvantage neighborhoods exhibited significantly more defensive behaviors (e.g., requiring supervision or limiting locations of play) than parents in low disadvantage neighborhoods. Additionally, parents in high disadvantage neighborhoods discussed rules that may be more limiting for outdoor play than parentings in low disadvantage neighborhoods. Disadvantaged neighborhoods typically have higher crime and traffic insecurity, poor social capital among residents and lack physical activity resources (e.g., parks, sidewalks, greenspace) that have been shown to alter parental perceptions of their environment and translate to adolescents attaining lower levels of physical activity.^7,46–48 This study provides some evidence that, in addition to changing parenting perceptions, neighborhood environments also affect parenting practices (i.e., avoidance and defensive behaviors). Interventions may need to incorporate policy and environment change to improve neighborhood conditions so that parents are less fearful for their child’s safety and lessen restrictions on outdoor play. Change needs to occur in both low and high disadvantage neighborhoods to mitigate parental restrictions on outdoor play, but previous qualitative findings in this sample suggest that targets for change may differ by neighborhood type. For example, parents in high disadvantage neighborhoods indicated that access to places to play (e.g., playgrounds) and walkability (e.g., presence of a sidewalk and reduced traffic) were most important for parenting practices; whereas, parents in low disadvantage neighborhoods identified the social cohesion and social control among neighbors as potential targets for intervention. ⁵⁴ Additional research is needed to expand this work to diverse populations and settings (e.g., additional cities or rural areas, diverse age groups) and further clarify intervention targets (e.g., modifiable neighborhood characteristics or parental perceptions of the neighborhood) using quantitative methods (e.g., GIS).

Our limited knowledge with respect to constraint on independent mobility and outdoor play may be a result of measurement challenges.²³ First, objective measures of physical activity (accelerometry) combined with Global Positioning Systems (GPS) devices are necessary to generate objective measures of neighborhood activity, an objective outcome that would be directly influenced by parenting rules for neighborhood outdoor play. The current study relied on parent reports of the frequency of neighborhood outdoor play over the past year that may overestimate activity.⁶⁰ Additionally, this study did not ask adolescents about their own neighborhood outdoor play and objective measures of physical activity were not limited to physical activity acquired in the neighborhood. The use of these technologies (GPS, accelerometer) will allow for clarity on the relationship between neighborhood-based and overall physical activity. This understanding may bring insight to the differences in physical activity between adolescents in low versus high disadvantage neighborhoods and between males and females as families with greater financial flexibility have more opportunities to participate in physical activity outside the neighborhood, and males more commonly participate in organized sports.⁶¹ Secondly, more rigorous measurement of parenting practices may be warranted. A major strength of this study was the use of both qualitative and quantitative methods to examine parenting practices for outdoor play. This comprehensive approach identified intricacies to parenting practices that were not evident from quantitative methods only. Based on results from qualitative interviews, future quantitative studies striving to measure parenting practices for outdoor play may need to include additional survey questions on defensive behaviors that assess rules regarding play with peers and siblings, locations of play (i.e. yard vs. common areas) and collaboration and trust with other adults in the neighborhood.

Parenting practices examined in this study may not be representative of all geographic areas (e.g., rural communities) despite the diverse sample of neighborhoods. Parents in this study were represented by women (29 of the women interviewed were mothers and one was a grandmother); therefore, the male (or fatherly) perspective on parenting decisions for outdoor play was not represented. This limitation is true of the majority of studies examining parental influences on children’s physical activity which have focused on the mother-child relationship and, in general, lack evidence for the role of fathers.⁶² Traditionally, the female guardian’s perspective was more pertinent as women were usually the parent who was home to oversee children’s play in the neighborhood. However, these roles may be shifting which further warrants understanding the fathers’ perspectives.⁶³ Lastly, the interaction between race and neighborhood disadvantage may warrant further exploration. It was the intent to recruit a diverse sample of African American and White parents across neighborhood types; however, 49% (n-64) of participants from high disadvantage neighborhoods were African American mothers whereas 22% (n=28) of mothers from low disadvantage neighborhoods were African American (p<0.001). This unequal proportion may be due to the inclusion of ‘percent of households who are African American’ as a factor included in our definition of disadvantage and evidence that African Americans are more likely to reside in significantly poorer neighborhoods than white. ⁶⁴ Further research is necessary to determine whether parents of different races and ethnicities (e.g. white vs. African American) living in the same neighborhood environment have different parenting practices that may influence the relationship to adolescent outdoor play.

The current study provided evidence that parental restrictions are impacting adolescents’ outdoor play across neighborhoods with varying levels of disadvantage. Parents have requirements for supervision and restrictions on the location and time of play that vary depending on whether children are playing alone or with siblings/peers or the type of activity (e.g., bicycling, walking, swimming). Further understanding of whether parentings practices are more important for youth living in disadvantaged neighborhoods and females is necessary to tailor intervention efforts. These efforts should include policies and environment changes, such as zoning to increase access to safe places to play, increases in policing and traffic calming measures, improvements in social capital through social media groups or neighborhood associations that may be necessary to reduce parental restrictions, and increased neighborhood outdoor play to improve the health of developing youth.