T H E
C I E N C E
O F
L I F E S T Y L E
C H A N G
Applied Research Brief: Health Promoting Community Design
Sex Moderates Associations Between Perceptions of the Physical and Social Environments and Physical Activity in Youth Justin B. Moore, PhD, MS; Michael W. Beets, PhD, MPH, MEd; Andrew T. Kaczynski, PhD; Gina M. Besenyi, MPH; Sara F. Morris, MAT, MPH, RD; Mary Bea Kolbe, MPH, RD
Abstract Purpose. To determine i f the sex o f the child moderates the relationships between perceptions o f the physical/social environments and moderate to vigorous physical activity (MVPA) in youth. Design. Cross-sectional. Setting. North Carolina. Subjects. A fin a l sample o f 711 children, 8 to 17 years of age, was available fo r analysis. Measures. Self-reported presence o f environmental factors previously identified to be associated with physical activity in youth was collected via survey. Daily MVPA was assessed via accelerometry fo r a m inim um o f 4 days. Analysis. Multilevel linear regression models were employed, adjusted fo r clustering at the county and individual level. MVPA was first regressed onto sex and environmental perception items while controlling fo r grade and race. The interaction term between sex and environmental perception was then added to the model. Results. A significant positive association was observed in the first models between MVPA and two items related to parent permission to (1) walk and (2) ride a bike in the neighborhood. These effects were fully moderated by sex, with males indicating “yes'" on these items exhibiting 6.87 and 5.21 more minutes o f MVPA (respectively) than males indicating “no.” Conclusion. Environm ental perceptions appear to be related to MVPA, but this relationship is present only in males. Future research should be conducted to identify modifiable social and physical characteristics that are associated with MVPA in females. (Am J Health Promot 2014;29[2]:132-135.)
Key Words: Motor Activity, Environment, Child, Accelerometry, Prevention Research. Manuscript format: applied research brief; Research purpose: descriptive; Study design: nonexperimental; Outcome measure: behavioral; Setting: state; Health focus: physical activity; Strategy: built environment; Target population age: youth; Target population circumstances: geographic location
PURPOSE Youth do not achieve sufficient amounts of physical activity (PA), with fewer girls meeting 60 m in /d recom mendations for moderate to vigorous PA (MVPA) than boys (35% vs. 49%, re spectively) .1 Researchers have identified perceived environmental correlates of PA in hopes of ascertaining targets for intervention." However, many studies rely on associations with self-reported PA, which lacks the validity of objective measures.! A recent review of studies examining perceived environmental correlates and PA in youth concluded that inconsistent results were observed across studies,4 potentially stemming from the subjective nature of PA mea surement utilized. Preliminary evidence suggests the relationship between phys ical/social environment correlates and PA in youth may differ between boys and girls. Wenthe and colleagues ’ found differences by sex in the association between perceived access to facilities and MVPA among a sample of adolescent youth, with only males displaying a
J u s tin B. Moore, PhD, M S; A n d rew T. K aczynski, PhD ; a n d G ina M . Besenyi, M P H , are w ith the D epartm ent o f H e a lth Prom otion, E du ca tio n , a n d Behavior, A rn o ld School o f P ublic H ealth; J u s tin B. Moore, PhD, M S, is w ith the Office o f P ublic H ea lth Practice, A rn o ld School o f P ublic H ealth; A ndrew T. K aczynski, PhD, is w ith the P revention Research Center, A rn o ld School o f P ublic H ealth; a n d M ichael W. Beets, PhD, M P H , M E d, is w ith the D epartm ent o f Exercise Science, U niversity o f Sou th Carolina, C olum bia, S outh Carolina. Sara F. M orris, M A T, M P H , RD , is w ith the D iv isio n o f P ublic H ealth, a n d M a ry Bea Kolbe, M P H , RD, is w ith the P hysical A ctivity a n d N u tritio n B ranch, D iv isio n o f P ublic H ealth, N orth C arolina D epartm ent o f H ea lth a n d H u m a n Services, R aleigh, N orth Carolina.
Send reprint requests to Justin B. Moore, PhD, MS, Department of Health Promotion, Education, and Behavior, and Office of Public Health Practice, Arnold School of Public Health, University of South Carolina, 800 Sumter Street, Room 216, Columbia, SC 29208; [email protected]. This m anuscript was submitted October 23, 2012; revisions were requested Ju ly 12, 2013; the m anuscript was accepted fo r publication A u g u st 6, 2013. Copyright © 2 0 1 4 by American Journal o f Health Promotion, Inc. 0 8 9 0 -1 1 7 1 /1 4 /$ 5 .0 0 + 0 DOI: 1 0 .4 2 78/ajhp. 121023-ARB-516
132
American Journal of Health Promotion
November/December 2014, Vol. 29, No. 2
Table Regression Coefficients and 95% CIs for Environmental Perception Items and Item by Sex Interaction Terms With MVPA as the Dependent Variable* Step 1 ltemt _______________________ 1
It is safe to w a lk o r jo g in m y n e ig hbo rho od
Coefficient (B) 2.34
Step 2 95% Cl
- 0 . 4 2 to 5.09
S afe to w a lk * sex 2 3 4 5 6
I often s e e o th e r girls o r b o ys pla yin g in m y ne ig h b o rh o o d S ee o th e r girls o r b o ys * s ex
1.95
T h e re are lots o f lo o s e o r s c a ry do g s in m y ne ig h b o rh o o d S ca ry do g s * sex
-1 .1 3
W h e n w a lk in g in m y ne ig hb o rh o o d th e re is a lot o f s m o k e /o th e r bad sm ells S m o ke /b a d s m e lls * s ex
0.46
T h e re are s id e w a lk s on m o s t o f th e s tre e ts in m y n e ig hbo rho od T h e re are s id e w a lk s * sex
0.42
T h e re are b icycle o r w a lk in g tra ils in m y n e ig hbo rho od B icycle o r w a lk in g trails * sex
- 0 . 4 9 to 4.38 - 4 . 0 0 to 1.74 - 2 . 9 9 to 3.92 - 2 . 6 6 to 3.50
-1 .0 9
- 3 . 4 4 to 1.25
Coefficient (S)
95% Cl
1.16
- 2 . 4 6 to 4.78
2.7 7
- 2 . 7 5 to 8.29
0.07
- 3 . 2 8 to 3.41
3.85
- 0 . 8 7 to 8.57
-1 .0 2
- 4 . 7 8 to 2.74
-0 .2 6
- 5 . 8 6 to 5.35
2.10
- 2 . 4 0 to 6.60
-3 .8 9
- 1 0 .7 5 to 2.9 7
-0 .9 2
- 4 . 8 6 to 3.03
3.04
- 2 . 5 5 to 8.64
-0 .9 7
- 4 . 1 3 to 2.1 9
-0 .2 8
- 4 . 9 7 to 4.42 -3 .1 1 to 3.09
7
M y p a ren ts (or gu ard ian s) a llo w m e to w a lk in o u r n e ig h b o rh o o d on m y ow n P are nts a llo w m e to w a lk * sex
3.08
0.75 to 5.42
- 0 .0 1
8
M y p a re n ts (or g u a rd ian s) a llo w m e to bike in o u r ne ig h b o rh o o d on m y ow n P a re n ts a llo w m e to b ike * sex
2.77
0.40 to 5.14
0.52
5.21
0.55 to 9.87
9
E n viro n m e n ta l pe rc e p tio n s sc o re (range: 1 -8 )
0.73
0.06 to 1.40
-0 .0 3
- 0 . 8 9 to 0.84
1.78
0.48 to 3.08
6.87
E n viro n m e n ta l pe rc e p tio n s sc o re * s ex
2.31 to 11.42 - 2 . 5 9 to 3.6 3
* C o e fficie n ts in bold are s ig n ific a n t a t p < .05. C l in dica tes c o n fid e n ce interval; and M V P A , m od era te to v iq o ro u s p h ysica l activity t Y e s = 1; m ale = 1. a n y y-
positive association with self-reported MVPA. This association was not present for objectively measured MVPA, possibly because of the limited sample size (n = 205) or heterogeneity of the sample. Previous research suggests that a social ecological framework should guide in quiry in this area, as this framework accounts for numerous influences at the intrapersonal, interpersonal, and com munity levels.15Therefore, the objective of the present study was to employ a social ecological approach to determine if the relationships between perceptions of individual and aggregate physical/ social environment measures and MVPA were moderated by sex in a large, diverse sample ofyouth (8-l7yearsofage) from North Carolina. METHODS Design This was a cross-sectional study to examine associations between percep tions of the physical/social environ ment and MVPA by sex in youth. Sample Youth from 18 counties in North Carolina were recruited to participate in
A m e rica n Jo u rn a l o f H ealth P rom otion
baseline data collection of a larger study examining the effects of communitybased mini-grants on PA. The partici pating counties represented a racially diverse sample of North Carolina resi dents, with county demographics rang ing from 97% to 60% white (North Carolina average 72%) according to the 2010 U.S. Census. Participants repre sented a convenience sample recruited from schools and community organiza tions by local grant coordinators from the target population of their forthcom ing intervention. Participants were re cruited via consent materials sent home to parents with the child that were distributed at the targeted location. Measures Eight questions asked about the pres ence or absence of physical/social fac tors previously identified to be associated with PA in youth (Table).7 Items were utilized separately and as a summary variable (i.e., num ber of positive re sponses; range 0-8). In a similar popu lation of girls, Evenson et al.7 found that the test-retest reliability of the items ranged from .37 to .58 assessed using weighted kappa coefficients. In the present sample, the Kuder-Richardson
coefficient for the summary variable was .53. Youth also wore accelerometers (ActiGraph, Pensacola, Florida) for 7 days, removing the monitors when bathing, swimming, and sleeping. Thirty minutes of consecutive zeros was con sidered indicative of nonwear time, and days were considered complete if at least 10 valid hours was present for analysis.8 Only data from youth who provided complete questionnaire data on the items of interest and 4 valid days of accelerometer data were included. Daily MVPA was calculated by summing min utes of MVPA from valid days utilizing previously established cut points for MVPA (>2296 counts/m in).9 Loprinzi and colleagues10 examined the influ ence of five different cut points for children on meeting PA recommenda tions and select health outcomes and concluded that these cut points are the most appropriate for youth. More details regarding accelerometer data collection can be found elsewhere.11 Analyses For each of the eight items and the summary variable, two nested, multilevel linear regression models, which mod eled both random and fixed effects, were
N o v e m b e r/D e ce m b e r 2014, Vol. 29, No. 2
133
employed. MVPA was first regressed onto the main effects of sex and environ mental perceptions items individually, while controlling for grade and race. The interaction term between sex and envi ronmental perception was then added to the second model. The models adjusted for clustering at the county and individ ual level. Analyses were conducted in STATA 12.1. RESULTS Participant Characteristics A total of 1164 children provided data, with at least 4 days of accelerom eter data and complete questionnaire data available for 711 children (61%). The mean age of participants was 11.3 years (SD = 1.4); 56% were female, 65% were white, and 20% were AfricanAmerican. Participants provided an average of 12.1 (SD = 4.5) hours of accelerometer monitoring per valid day across an average of 5.5 days (SD = 1.1). Boys and girls were evenly dis tributed across the counties from which they were recruited. Regression Models Results of the unconditional multi level models indicated the mean MVPA was 40.8 m in /d for males and 29.4 m in /d for females, with 10% to 18% of children achieving the recommended 60 minutes of MVPA on a given day. County of residence accounted for 5.7% of the variance in MVPA. A significant positive association was ob served in the first models between MVPA and two items: “My parents (or guardian) allow me to walk in our neighborhood on my own,” and “My parents (or guardian) allow me to bike in our neighborhood on my own.” These effects were fully m oderated by sex, with males who indicated “yes” on these items exhibiting 5.2 to 6.9 more minutes of MVPA than males who indicated “no” (Table). Results of the model that examined the relationship between the environmental percep tions summed score and MVPA indi cated that every “yes” response was associated with .7 additional minutes of MVPA (p < .05). The relationship between the environmental percep tions summed score and MVPA was completely moderated by sex, as each additional “yes” response was associat
134
American Journal of Health Promotion
ed with 1.8 additional minutes of MVPA for males (p < .05). DISCUSSION Summary The results of the present study are consistent with the literature that reports that boys are more active than girls.1 Similarly, certain self-reported social and physical environmental features were associated with MVPA in the initial models. However, although environ mental perceptions appear to be related to MVPA in the present sample, this relationship was only present in males. This relationship was particularly pro nounced for parental support in the current sample, with the items related to parental permission to be active in the neighborhood independently associated with MVPA in boys but not girls. Howev er, the summed environmental percep tions score displayed a similar association with MVPA that was moderated by sex. It might be argued that parents are more protective of girls1' and thus not allow ing girls to go outside to be active limits their exposure to physical environment variables, thereby rendering them a nonfactor. Future research to elucidate these findings should be conducted to identify modifiable social and physical characteristics that are associated with MVPA in females. Boys (40.8 m in/d) and girls (29.4 m in/d) in this sample displayed similar levels of MPVA to those previously observed in boys (45.3 m in/d) and girls (24.6 m in/d) from a national sample,1 with important differences. Boys in this sample were considerably less active, whereas girls were considerably more active (~5 m in/d) than those in a national sample. The discrepancy ob served for girls is understandable, be cause the average age (11 years) in the present sample is lower than the age of the comparison sample (15 years), and MVPA in girls is known to decline precipitously during this age range.13 However, this was not the case for boys. Because the lower level of MVPA in the present sample of boys cannot be readily explained, the results should be inter preted in this context and with caution. Limitations These were baseline data from a study examining interventions to pro
mote PA in low-resource communities. The fact that these communities were low resource may affect the generalizability of the results. However, the current results are consistent with national estimates,1 with im portant caveats indicated above. Research by Loprinzi et al.10 clearly indicates that PA prevalence estimates vary consider ably based upon the cut point chosen, and this variability could explain the discrepancy between our data and those from national samples. Related, only a small portion of our sample (10%—18%) m et the recom m endation of 60 minutes of MVPA per day. It is possible that the relationship between environmental correlates and MVPA differs in the subsample m eeting rec ommendations, but sample size pre cludes testing this in the current study. In addition, only 61% of the sample provided enough usable data. Al though this is consistent with previous national studies,1 it is unclear how compliance affects the results. Fur therm ore, no objective measures of the physical/social environm ent were available. It is possible that objective data regarding the presence of chil dren in the neighborhood or parental permission to walk/bike in the neigh borhood would have yielded different results. Additionally, parenting prac tices14 could have mediated the rela tionships that were elucidated, but unfortunately, they were not assessed in the current study. Finally, no questions were asked about MVPA in the home environment. Because previous re search has suggested that older chil dren are active in their hom es,1’’ this is an im portant issue that should be addressed in future research. Significance This study is significant in that it establishes the moderating effect of sex on the relationship between the physi cal/social environments and PA in youth. In light of the considerable benefits of MVPA in youth16 and the noted disparity in MVPA between girls and boys in the present and national samples, these results highlight the need for more research on potential social mediators of this relationship.12,14,1 ‘ Although these findings should be rep licated in a representative sample of youth, they provide initial support for
November/December 2014, Vol. 29, No. 2
the contention that perceived social and physical environmental supports are more relevant to MVPA in boys than in girls. These results have particular rele vance for interventions targedng girls who are at risk for insufficient MVPA.
Acknowledgments This work was supported by a grant fro m the Active L ivin g Research program o f the Robert Wood Johnson Foundation. A dditional support was provided by the North Carolina Departmen t o f H ealth a n d H u m a n Services.
R eferences
1. Troiano RP, Berrigan D, Dodd KW, et al. Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc. 2008;40:181-188. 2. de Vet E, de Ridder DTD, de Wit JBF. Environmental correlates of physical activity and dietary behaviours among young people: a systematic review of reviews. Obes Rev. 2011;12:el30-el42. 3. Craggs C, Corder K, van Sluijs EMF, Griffin SJ. Determinants of change in physical activity in children and adolescents: a systematic review. Am J Prev Med. 2011;40: 645-658. 4. Ding D, Sallis JF, KerrJ, et al. Neighborhood environment and physical activity among youth: a review. Am J Prev Med. 2011;41:442-455. 5. Wenthe PJ, Janz KF, Levy SM. Gender similarities and differences in factors associated with adolescent moderatevigorous physical activity. Pediatr Exerc Sci. 2009;21:291-304. 6. McLeroy KR, Bibeau D, Steckler A, Glanz K. An ecological perspective on health promotion programs. Health Educ Q. 1988; 15:351-377. 7. Evenson KR, Scott MM, Cohen DA, Voorhees CC. Girls’ perception of neighborhood factors on physical activity, sedentary behavior, and BM1. Obesity. 2007; 15:430-445. 8. Sirard JR, Slater ME. Compliance with wearing physical activity accelerometers in high school students. J Phys Act Health. 2009;6(suppl 1):S148-S155.
SO WHAT? Implications for Health Promotion Practitioners and Researchers What is already known on this topic?
Examinations of the relationship between the physical/social environ ment and physical activity (PA) in youth have yielded conflicting re sults. These results may be a function of the lack of the analyses to consider the moderating effect of sex on these relationships. What does this article add?
This article adds evidence that sex differences exist regarding the rela tionship between perceptions of the physical/social environments and PA. These findings suggest that the association between the perceived physical/social environments and PA may be more complex than previ ously envisioned and differ by sex of the child. What are the implications for health promotion practice or research?
It is important to consider factors that may be unique to boys or girls when designing PA promoting envi ronments or interventions.
I UNITED STATES Statem ent o f O w nership, M anagem ent, a n d C irc u la tio n POSTAL SERVICE• (A ll P e riodicals P u b lic a tio n s E xcept R equester P u b lica tio n s) I.Pubacatonttumbar
American Journal of Health Promotion
°M°I N7I*M
9 -5 -20 1 4
H.ISHMHWI
ta. laauaDataforCMulaSonOataBarow
J u ly -2 0 1 4 18.Crtaal anaNaturoof Clrautadon
AuarataNa. Cagtaa No. CapiasofSkigla EachtaauaDuring Pracadlng11Manna
&Nurrbar«(laauaaptoStw Armuaty
Bim onthly
$ 1 3 9 .0 0
6
American Journal of Health Promotion, Inc. 5S E. long Lake Rd, 4466 Troy, Ml 48085-4738
a. ToolHumbarof Copaa(N*pnunjr)
Michael O'DonnsIl TeiaghonaIModtm at)» 248-682-0707
!.S IS 3 n 8 3 ^ B 5 5 M B SSB5B 5 « r3 5 5 i5 5 5 G 5 » g C S S S 5 C ? B i5 5 } A m e ric a n Jo u rn a l o f H e a lth P rom o tion , Inc. 5 5 E . L o ng La ke R d, # 4 6 6 T ro y, M l 4 8 0 8 5 -4 7 3 8
MaladOuMdaCountyPa* SubsatpfemSMadanPSfarm3S41(IneMapaid ... dasfcuton aba* norttoalrata,adnrdaa/sproof andaarf>anga«aiaa)
wET
Cl Matad
O utlet, msMa«
OkMMfca(MaltarnaMatoMMkaSstaaTlroghOorrianandCamara. PI Pat) SbaatVandora.CastorSstos.andOtiasPaMW M > OuWtoUSP*»
^
aaPS
uMmyEjt
PaMMatrftutonbyOmarOaaaaaatMai INaugr.maIMPS (ag. Fia-.CIaaaMaH) A m e ric a n Jo u rn a l o f H e a lth P ro m o tio n , Inc. 5 5 E . L o ng La ke R d, # 4 6 6 T ro y. M l 4 8 0 8 5 -4 7 3 8
Edtor (Himaridcwipow"lU t oM/aar)
————— ——————— ——
—
e. Total Pakl OMrtwaonflumsrruru ft). m*aS 4. ParsT)
Michael O’Donnell, AmericanJournal of Health Promotion, Inc. 56 E. long lake Rd. M 68 Troy, Ml 48085-4738
Kristin Anderson, Allen Press, Inc., 810 E. 10th Street, Lawrence, KS 66044
rd
y
CD FraaarNsrttoal RataOuuidaCcuntyCcgtoaIxJudadonPSFarmMat
Distributor) ft) Fraaor Nombial RatabrCauniyCapiasM aM « P* FarmJS41
C I E N C E
O F
L I F E S T Y L E
C H A N G
Applied Research Brief: Health Promoting Community Design
Sex Moderates Associations Between Perceptions of the Physical and Social Environments and Physical Activity in Youth Justin B. Moore, PhD, MS; Michael W. Beets, PhD, MPH, MEd; Andrew T. Kaczynski, PhD; Gina M. Besenyi, MPH; Sara F. Morris, MAT, MPH, RD; Mary Bea Kolbe, MPH, RD
Abstract Purpose. To determine i f the sex o f the child moderates the relationships between perceptions o f the physical/social environments and moderate to vigorous physical activity (MVPA) in youth. Design. Cross-sectional. Setting. North Carolina. Subjects. A fin a l sample o f 711 children, 8 to 17 years of age, was available fo r analysis. Measures. Self-reported presence o f environmental factors previously identified to be associated with physical activity in youth was collected via survey. Daily MVPA was assessed via accelerometry fo r a m inim um o f 4 days. Analysis. Multilevel linear regression models were employed, adjusted fo r clustering at the county and individual level. MVPA was first regressed onto sex and environmental perception items while controlling fo r grade and race. The interaction term between sex and environmental perception was then added to the model. Results. A significant positive association was observed in the first models between MVPA and two items related to parent permission to (1) walk and (2) ride a bike in the neighborhood. These effects were fully moderated by sex, with males indicating “yes'" on these items exhibiting 6.87 and 5.21 more minutes o f MVPA (respectively) than males indicating “no.” Conclusion. Environm ental perceptions appear to be related to MVPA, but this relationship is present only in males. Future research should be conducted to identify modifiable social and physical characteristics that are associated with MVPA in females. (Am J Health Promot 2014;29[2]:132-135.)
Key Words: Motor Activity, Environment, Child, Accelerometry, Prevention Research. Manuscript format: applied research brief; Research purpose: descriptive; Study design: nonexperimental; Outcome measure: behavioral; Setting: state; Health focus: physical activity; Strategy: built environment; Target population age: youth; Target population circumstances: geographic location
PURPOSE Youth do not achieve sufficient amounts of physical activity (PA), with fewer girls meeting 60 m in /d recom mendations for moderate to vigorous PA (MVPA) than boys (35% vs. 49%, re spectively) .1 Researchers have identified perceived environmental correlates of PA in hopes of ascertaining targets for intervention." However, many studies rely on associations with self-reported PA, which lacks the validity of objective measures.! A recent review of studies examining perceived environmental correlates and PA in youth concluded that inconsistent results were observed across studies,4 potentially stemming from the subjective nature of PA mea surement utilized. Preliminary evidence suggests the relationship between phys ical/social environment correlates and PA in youth may differ between boys and girls. Wenthe and colleagues ’ found differences by sex in the association between perceived access to facilities and MVPA among a sample of adolescent youth, with only males displaying a
J u s tin B. Moore, PhD, M S; A n d rew T. K aczynski, PhD ; a n d G ina M . Besenyi, M P H , are w ith the D epartm ent o f H e a lth Prom otion, E du ca tio n , a n d Behavior, A rn o ld School o f P ublic H ealth; J u s tin B. Moore, PhD, M S, is w ith the Office o f P ublic H ea lth Practice, A rn o ld School o f P ublic H ealth; A ndrew T. K aczynski, PhD, is w ith the P revention Research Center, A rn o ld School o f P ublic H ealth; a n d M ichael W. Beets, PhD, M P H , M E d, is w ith the D epartm ent o f Exercise Science, U niversity o f Sou th Carolina, C olum bia, S outh Carolina. Sara F. M orris, M A T, M P H , RD , is w ith the D iv isio n o f P ublic H ealth, a n d M a ry Bea Kolbe, M P H , RD, is w ith the P hysical A ctivity a n d N u tritio n B ranch, D iv isio n o f P ublic H ealth, N orth C arolina D epartm ent o f H ea lth a n d H u m a n Services, R aleigh, N orth Carolina.
Send reprint requests to Justin B. Moore, PhD, MS, Department of Health Promotion, Education, and Behavior, and Office of Public Health Practice, Arnold School of Public Health, University of South Carolina, 800 Sumter Street, Room 216, Columbia, SC 29208; [email protected]. This m anuscript was submitted October 23, 2012; revisions were requested Ju ly 12, 2013; the m anuscript was accepted fo r publication A u g u st 6, 2013. Copyright © 2 0 1 4 by American Journal o f Health Promotion, Inc. 0 8 9 0 -1 1 7 1 /1 4 /$ 5 .0 0 + 0 DOI: 1 0 .4 2 78/ajhp. 121023-ARB-516
132
American Journal of Health Promotion
November/December 2014, Vol. 29, No. 2
Table Regression Coefficients and 95% CIs for Environmental Perception Items and Item by Sex Interaction Terms With MVPA as the Dependent Variable* Step 1 ltemt _______________________ 1
It is safe to w a lk o r jo g in m y n e ig hbo rho od
Coefficient (B) 2.34
Step 2 95% Cl
- 0 . 4 2 to 5.09
S afe to w a lk * sex 2 3 4 5 6
I often s e e o th e r girls o r b o ys pla yin g in m y ne ig h b o rh o o d S ee o th e r girls o r b o ys * s ex
1.95
T h e re are lots o f lo o s e o r s c a ry do g s in m y ne ig h b o rh o o d S ca ry do g s * sex
-1 .1 3
W h e n w a lk in g in m y ne ig hb o rh o o d th e re is a lot o f s m o k e /o th e r bad sm ells S m o ke /b a d s m e lls * s ex
0.46
T h e re are s id e w a lk s on m o s t o f th e s tre e ts in m y n e ig hbo rho od T h e re are s id e w a lk s * sex
0.42
T h e re are b icycle o r w a lk in g tra ils in m y n e ig hbo rho od B icycle o r w a lk in g trails * sex
- 0 . 4 9 to 4.38 - 4 . 0 0 to 1.74 - 2 . 9 9 to 3.92 - 2 . 6 6 to 3.50
-1 .0 9
- 3 . 4 4 to 1.25
Coefficient (S)
95% Cl
1.16
- 2 . 4 6 to 4.78
2.7 7
- 2 . 7 5 to 8.29
0.07
- 3 . 2 8 to 3.41
3.85
- 0 . 8 7 to 8.57
-1 .0 2
- 4 . 7 8 to 2.74
-0 .2 6
- 5 . 8 6 to 5.35
2.10
- 2 . 4 0 to 6.60
-3 .8 9
- 1 0 .7 5 to 2.9 7
-0 .9 2
- 4 . 8 6 to 3.03
3.04
- 2 . 5 5 to 8.64
-0 .9 7
- 4 . 1 3 to 2.1 9
-0 .2 8
- 4 . 9 7 to 4.42 -3 .1 1 to 3.09
7
M y p a ren ts (or gu ard ian s) a llo w m e to w a lk in o u r n e ig h b o rh o o d on m y ow n P are nts a llo w m e to w a lk * sex
3.08
0.75 to 5.42
- 0 .0 1
8
M y p a re n ts (or g u a rd ian s) a llo w m e to bike in o u r ne ig h b o rh o o d on m y ow n P a re n ts a llo w m e to b ike * sex
2.77
0.40 to 5.14
0.52
5.21
0.55 to 9.87
9
E n viro n m e n ta l pe rc e p tio n s sc o re (range: 1 -8 )
0.73
0.06 to 1.40
-0 .0 3
- 0 . 8 9 to 0.84
1.78
0.48 to 3.08
6.87
E n viro n m e n ta l pe rc e p tio n s sc o re * s ex
2.31 to 11.42 - 2 . 5 9 to 3.6 3
* C o e fficie n ts in bold are s ig n ific a n t a t p < .05. C l in dica tes c o n fid e n ce interval; and M V P A , m od era te to v iq o ro u s p h ysica l activity t Y e s = 1; m ale = 1. a n y y-
positive association with self-reported MVPA. This association was not present for objectively measured MVPA, possibly because of the limited sample size (n = 205) or heterogeneity of the sample. Previous research suggests that a social ecological framework should guide in quiry in this area, as this framework accounts for numerous influences at the intrapersonal, interpersonal, and com munity levels.15Therefore, the objective of the present study was to employ a social ecological approach to determine if the relationships between perceptions of individual and aggregate physical/ social environment measures and MVPA were moderated by sex in a large, diverse sample ofyouth (8-l7yearsofage) from North Carolina. METHODS Design This was a cross-sectional study to examine associations between percep tions of the physical/social environ ment and MVPA by sex in youth. Sample Youth from 18 counties in North Carolina were recruited to participate in
A m e rica n Jo u rn a l o f H ealth P rom otion
baseline data collection of a larger study examining the effects of communitybased mini-grants on PA. The partici pating counties represented a racially diverse sample of North Carolina resi dents, with county demographics rang ing from 97% to 60% white (North Carolina average 72%) according to the 2010 U.S. Census. Participants repre sented a convenience sample recruited from schools and community organiza tions by local grant coordinators from the target population of their forthcom ing intervention. Participants were re cruited via consent materials sent home to parents with the child that were distributed at the targeted location. Measures Eight questions asked about the pres ence or absence of physical/social fac tors previously identified to be associated with PA in youth (Table).7 Items were utilized separately and as a summary variable (i.e., num ber of positive re sponses; range 0-8). In a similar popu lation of girls, Evenson et al.7 found that the test-retest reliability of the items ranged from .37 to .58 assessed using weighted kappa coefficients. In the present sample, the Kuder-Richardson
coefficient for the summary variable was .53. Youth also wore accelerometers (ActiGraph, Pensacola, Florida) for 7 days, removing the monitors when bathing, swimming, and sleeping. Thirty minutes of consecutive zeros was con sidered indicative of nonwear time, and days were considered complete if at least 10 valid hours was present for analysis.8 Only data from youth who provided complete questionnaire data on the items of interest and 4 valid days of accelerometer data were included. Daily MVPA was calculated by summing min utes of MVPA from valid days utilizing previously established cut points for MVPA (>2296 counts/m in).9 Loprinzi and colleagues10 examined the influ ence of five different cut points for children on meeting PA recommenda tions and select health outcomes and concluded that these cut points are the most appropriate for youth. More details regarding accelerometer data collection can be found elsewhere.11 Analyses For each of the eight items and the summary variable, two nested, multilevel linear regression models, which mod eled both random and fixed effects, were
N o v e m b e r/D e ce m b e r 2014, Vol. 29, No. 2
133
employed. MVPA was first regressed onto the main effects of sex and environ mental perceptions items individually, while controlling for grade and race. The interaction term between sex and envi ronmental perception was then added to the second model. The models adjusted for clustering at the county and individ ual level. Analyses were conducted in STATA 12.1. RESULTS Participant Characteristics A total of 1164 children provided data, with at least 4 days of accelerom eter data and complete questionnaire data available for 711 children (61%). The mean age of participants was 11.3 years (SD = 1.4); 56% were female, 65% were white, and 20% were AfricanAmerican. Participants provided an average of 12.1 (SD = 4.5) hours of accelerometer monitoring per valid day across an average of 5.5 days (SD = 1.1). Boys and girls were evenly dis tributed across the counties from which they were recruited. Regression Models Results of the unconditional multi level models indicated the mean MVPA was 40.8 m in /d for males and 29.4 m in /d for females, with 10% to 18% of children achieving the recommended 60 minutes of MVPA on a given day. County of residence accounted for 5.7% of the variance in MVPA. A significant positive association was ob served in the first models between MVPA and two items: “My parents (or guardian) allow me to walk in our neighborhood on my own,” and “My parents (or guardian) allow me to bike in our neighborhood on my own.” These effects were fully m oderated by sex, with males who indicated “yes” on these items exhibiting 5.2 to 6.9 more minutes of MVPA than males who indicated “no” (Table). Results of the model that examined the relationship between the environmental percep tions summed score and MVPA indi cated that every “yes” response was associated with .7 additional minutes of MVPA (p < .05). The relationship between the environmental percep tions summed score and MVPA was completely moderated by sex, as each additional “yes” response was associat
134
American Journal of Health Promotion
ed with 1.8 additional minutes of MVPA for males (p < .05). DISCUSSION Summary The results of the present study are consistent with the literature that reports that boys are more active than girls.1 Similarly, certain self-reported social and physical environmental features were associated with MVPA in the initial models. However, although environ mental perceptions appear to be related to MVPA in the present sample, this relationship was only present in males. This relationship was particularly pro nounced for parental support in the current sample, with the items related to parental permission to be active in the neighborhood independently associated with MVPA in boys but not girls. Howev er, the summed environmental percep tions score displayed a similar association with MVPA that was moderated by sex. It might be argued that parents are more protective of girls1' and thus not allow ing girls to go outside to be active limits their exposure to physical environment variables, thereby rendering them a nonfactor. Future research to elucidate these findings should be conducted to identify modifiable social and physical characteristics that are associated with MVPA in females. Boys (40.8 m in/d) and girls (29.4 m in/d) in this sample displayed similar levels of MPVA to those previously observed in boys (45.3 m in/d) and girls (24.6 m in/d) from a national sample,1 with important differences. Boys in this sample were considerably less active, whereas girls were considerably more active (~5 m in/d) than those in a national sample. The discrepancy ob served for girls is understandable, be cause the average age (11 years) in the present sample is lower than the age of the comparison sample (15 years), and MVPA in girls is known to decline precipitously during this age range.13 However, this was not the case for boys. Because the lower level of MVPA in the present sample of boys cannot be readily explained, the results should be inter preted in this context and with caution. Limitations These were baseline data from a study examining interventions to pro
mote PA in low-resource communities. The fact that these communities were low resource may affect the generalizability of the results. However, the current results are consistent with national estimates,1 with im portant caveats indicated above. Research by Loprinzi et al.10 clearly indicates that PA prevalence estimates vary consider ably based upon the cut point chosen, and this variability could explain the discrepancy between our data and those from national samples. Related, only a small portion of our sample (10%—18%) m et the recom m endation of 60 minutes of MVPA per day. It is possible that the relationship between environmental correlates and MVPA differs in the subsample m eeting rec ommendations, but sample size pre cludes testing this in the current study. In addition, only 61% of the sample provided enough usable data. Al though this is consistent with previous national studies,1 it is unclear how compliance affects the results. Fur therm ore, no objective measures of the physical/social environm ent were available. It is possible that objective data regarding the presence of chil dren in the neighborhood or parental permission to walk/bike in the neigh borhood would have yielded different results. Additionally, parenting prac tices14 could have mediated the rela tionships that were elucidated, but unfortunately, they were not assessed in the current study. Finally, no questions were asked about MVPA in the home environment. Because previous re search has suggested that older chil dren are active in their hom es,1’’ this is an im portant issue that should be addressed in future research. Significance This study is significant in that it establishes the moderating effect of sex on the relationship between the physi cal/social environments and PA in youth. In light of the considerable benefits of MVPA in youth16 and the noted disparity in MVPA between girls and boys in the present and national samples, these results highlight the need for more research on potential social mediators of this relationship.12,14,1 ‘ Although these findings should be rep licated in a representative sample of youth, they provide initial support for
November/December 2014, Vol. 29, No. 2
the contention that perceived social and physical environmental supports are more relevant to MVPA in boys than in girls. These results have particular rele vance for interventions targedng girls who are at risk for insufficient MVPA.
Acknowledgments This work was supported by a grant fro m the Active L ivin g Research program o f the Robert Wood Johnson Foundation. A dditional support was provided by the North Carolina Departmen t o f H ealth a n d H u m a n Services.
R eferences
1. Troiano RP, Berrigan D, Dodd KW, et al. Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc. 2008;40:181-188. 2. de Vet E, de Ridder DTD, de Wit JBF. Environmental correlates of physical activity and dietary behaviours among young people: a systematic review of reviews. Obes Rev. 2011;12:el30-el42. 3. Craggs C, Corder K, van Sluijs EMF, Griffin SJ. Determinants of change in physical activity in children and adolescents: a systematic review. Am J Prev Med. 2011;40: 645-658. 4. Ding D, Sallis JF, KerrJ, et al. Neighborhood environment and physical activity among youth: a review. Am J Prev Med. 2011;41:442-455. 5. Wenthe PJ, Janz KF, Levy SM. Gender similarities and differences in factors associated with adolescent moderatevigorous physical activity. Pediatr Exerc Sci. 2009;21:291-304. 6. McLeroy KR, Bibeau D, Steckler A, Glanz K. An ecological perspective on health promotion programs. Health Educ Q. 1988; 15:351-377. 7. Evenson KR, Scott MM, Cohen DA, Voorhees CC. Girls’ perception of neighborhood factors on physical activity, sedentary behavior, and BM1. Obesity. 2007; 15:430-445. 8. Sirard JR, Slater ME. Compliance with wearing physical activity accelerometers in high school students. J Phys Act Health. 2009;6(suppl 1):S148-S155.
SO WHAT? Implications for Health Promotion Practitioners and Researchers What is already known on this topic?
Examinations of the relationship between the physical/social environ ment and physical activity (PA) in youth have yielded conflicting re sults. These results may be a function of the lack of the analyses to consider the moderating effect of sex on these relationships. What does this article add?
This article adds evidence that sex differences exist regarding the rela tionship between perceptions of the physical/social environments and PA. These findings suggest that the association between the perceived physical/social environments and PA may be more complex than previ ously envisioned and differ by sex of the child. What are the implications for health promotion practice or research?
It is important to consider factors that may be unique to boys or girls when designing PA promoting envi ronments or interventions.
I UNITED STATES Statem ent o f O w nership, M anagem ent, a n d C irc u la tio n POSTAL SERVICE• (A ll P e riodicals P u b lic a tio n s E xcept R equester P u b lica tio n s) I.Pubacatonttumbar
American Journal of Health Promotion
°M°I N7I*M
9 -5 -20 1 4
H.ISHMHWI
ta. laauaDataforCMulaSonOataBarow
J u ly -2 0 1 4 18.Crtaal anaNaturoof Clrautadon
AuarataNa. Cagtaa No. CapiasofSkigla EachtaauaDuring Pracadlng11Manna
&Nurrbar«(laauaaptoStw Armuaty
Bim onthly
$ 1 3 9 .0 0
6
American Journal of Health Promotion, Inc. 5S E. long Lake Rd, 4466 Troy, Ml 48085-4738
a. ToolHumbarof Copaa(N*pnunjr)
Michael O'DonnsIl TeiaghonaIModtm at)» 248-682-0707
!.S IS 3 n 8 3 ^ B 5 5 M B SSB5B 5 « r3 5 5 i5 5 5 G 5 » g C S S S 5 C ? B i5 5 } A m e ric a n Jo u rn a l o f H e a lth P rom o tion , Inc. 5 5 E . L o ng La ke R d, # 4 6 6 T ro y, M l 4 8 0 8 5 -4 7 3 8
MaladOuMdaCountyPa* SubsatpfemSMadanPSfarm3S41(IneMapaid ... dasfcuton aba* norttoalrata,adnrdaa/sproof andaarf>anga«aiaa)
wET
Cl Matad
O utlet, msMa«
OkMMfca(MaltarnaMatoMMkaSstaaTlroghOorrianandCamara. PI Pat) SbaatVandora.CastorSstos.andOtiasPaMW M > OuWtoUSP*»
^
aaPS
uMmyEjt
PaMMatrftutonbyOmarOaaaaaatMai INaugr.maIMPS (ag. Fia-.CIaaaMaH) A m e ric a n Jo u rn a l o f H e a lth P ro m o tio n , Inc. 5 5 E . L o ng La ke R d, # 4 6 6 T ro y. M l 4 8 0 8 5 -4 7 3 8
Edtor (Himaridcwipow"lU t oM/aar)
————— ——————— ——
—
e. Total Pakl OMrtwaonflumsrruru ft). m*aS 4. ParsT)
Michael O’Donnell, AmericanJournal of Health Promotion, Inc. 56 E. long lake Rd. M 68 Troy, Ml 48085-4738
Kristin Anderson, Allen Press, Inc., 810 E. 10th Street, Lawrence, KS 66044
rd
y
CD FraaarNsrttoal RataOuuidaCcuntyCcgtoaIxJudadonPSFarmMat
Distributor) ft) Fraaor Nombial RatabrCauniyCapiasM aM « P* FarmJS41
