Physical activity is associated with improved psychological well-being, reduced body mass, and lower levels of cardiometabolic risk factors among youth (20). The majority of young people do not meet current physical activity guidelines (18,21) with physical activity levels declining during childhood (13). The move from primary to secondary school is a key period when youth physical activity patterns change (13,14). Quantifying this change and the factors that predict change is important for developing effective strategies to maintain and increase youth physical activity.
Current interventions that have focused on increasing youth physical activity or obesity prevention with a strong focus on physical activity have yielded no, weak, or unsustained effects on body mass index (BMI) (19,23). The majority of these interventions have been delivered at school (8,23) and have focused on changes to physical education programs or school break periods. School-based interventions largely ignore the extent to which a child’s friends may affect the child’s physical activity both inside and outside of school. In light of these gaps in current knowledge and the lack of success of current school-based interventions (8), there is a need to examine the possibility of developing new peer-based physical activity interventions.
To modify a behavior such as physical activity, we need to understand the factors that influence that behavior and then alter those factors (1). We have recently shown that having “best friends” who are active and engaging in activity with best friends outside of school were associated with higher levels of activity among 10- to 11-yr-old children (10). It is not clear, however, whether general friend support, friend support for physical activity, or the overall number of friends is associated with children’s physical activity. Moreover, it is also not clear whether these friendship factors may be associated with change in physical activity after the transition to secondary school, a critical period when children’s physical activity levels decline (13). In this article, we address these questions among a UK cohort who were studied at both primary (elementary) and secondary school.
The analyses reported here represent a secondary analysis of the Personal and Environmental Associations with Children’s Health (PEACH) Project. Details of the study design have been reported elsewhere (16). Briefly, year-6 children (age 10–11 yr) were recruited from 23 state primary schools in Bristol (United Kingdom). The primary schools were selected because these had the highest transition rates (>40%) to eight local state-funded secondary schools. These eight urban secondary schools were selected on the basis of neighborhood deprivation and geographic location to represent the city. Pupils were then reassessed a year later at secondary school. This study was approved by a University of Bristol ethics committee, and written informed parental consent was obtained for all participants.
All measures were assessed in the final year of primary school (time 1) and the first year of secondary school (time 2). Physical activity was assessed using an ActiGraph accelerometer (model GT1M; ActiGraph LLC, Pensacola, FL) that was worn for 7 d and set to record data every 10 s. ActiGraph accelerometers have been shown to provide estimates of energy expenditure that are closely associated to laboratory-derived energy expenditure (24). Because children’s physical activity varies by the hours of daylight available (16), the mean hours of daylight from 3 p.m. to sunset on the first day of data collection were obtained. Height (m) was assessed using a Seca stadiometer (Hamburg, Germany), and weight (kg) was measured using a Seca digital scale. BMI (kg·m−2) was calculated and converted to an age- and gender-specific SD score (BMI SDS) (3). Pubertal status was self-reported using an established scale (17). The index of multiple deprivation (IMD) was also obtained for each participant’s home address on the basis of the postcode of the primary residence (15). The IMD is a composite area-based measure that assesses seven deprivation domains: income, employment, health, education, housing, crime, and the living environment (15).
All friend measures were assessed on a computer-based questionnaire. Friend social support measures were assessed using two subscales from an established scale (5). Friend support for physical activity was based on responses to the following four questions that asked “how often do your good friends 1) encourage you to exercise or play sports, 2) exercise or play sports with you, 3) tell you that you are doing well in exercise or sports, and 4) watch you take part in exercise or sports?” General friend support was assessed using three items: 1) “I have at least one friend who really encourages me,” 2) “I have at least one friend who makes me feel better when I am upset,” and 3) “I have at least one friend who understands my problems and worries.” In addition, participants answered three questions that asked the extent to which the participants’ friends preferred to 1) watch rather than take part in sports, 2) watch or play computer games rather than play outside, and 3) stay inside rather than play outside. These three questions were interpreted as friend sedentary preferences. Participants also reported the number of friends that they had. All items were assessed on four-point scales (strongly agree to strongly disagree). Subscale items were included in the analysis only if participants answered all items for the subscale with the number of complete responses differing for each subscale. Responses for the items included in each subscale were summed and divided by the number of items to produce a subscale score. With the exception of friend sedentary preferences (0.69), the α values for each scale were all above 0.7 indicating good internal consistency. Moreover, in a small reliability study with 46 children, the intraclass correlations were all >0.6 indicating good test–retest reliability.
Accelerometer data reduction
Periods of ≥60 min of zero values were defined as accelerometer “nonwear” time and discarded. Participants were included in the weekday analysis if they provided three or more weekdays of data with at least 480 min of data per day. Participants were included in the weekend analyses if they provided 480 min of data for at least one weekend day. Mean minutes of moderate- to vigorous-intensity physical activity (MVPA) after school on a weekday and overall weekend MVPA were obtained. Our previous global positioning system analysis indicated that very few children in this study engaged in outdoor physical activity beyond 8:30 p.m. Because we hypothesized that friends are most likely to influence outdoor physical activity, the after-school period was limited to between 3:30 and 8:30 p.m. on a weekday (4). Because a recent article by Trost et al. (22) has indicated that the accelerometer MVPA cut point of ≥2296 counts per minute of Evenson et al. (6) is the most accurate means of classifying MVPA in children, this threshold was adopted for all analyses.
Change (time 2 − time 1) was calculated for all of the accelerometer and friend variables. The change analyses were performed for each individual to provide a time 1, time 2, and change variable on all variables for all participants. To maximize power, the sample size for each analysis varied according to available data with the sample for each analysis shown in each table. Descriptive statistics were calculated, and a two-sample t-test was used to determine whether there was a difference in the baseline MVPA of participants who had valid accelerometer data at both time points when compared with participants who had valid data at baseline only. Three sets of analyses were then conducted. Cross-sectional analyses of factors associated with MVPA at time 1 and time 2 were conducted. A third longitudinal analysis examined whether change in the exposure variables between time 1 and time 2 corresponded to a change in MVPA between these two time points. For each analysis, there were two MVPA outcome variables: after-school MVPA and weekend MVPA. All analyses were stratified by sex resulting in a total of 12 separate analyses. Friend variables were the exposures with all models adjusted for BMI SDS, IMD, pubertal status, and hours of daylight. Robust SE values were used to take account of the clustering of participants in schools when calculating P values and 95% confidence intervals. Robust SE values apply a sandwich estimate of the variance structure of the data and make no assumptions about the variance structure of the data (11). Post hoc paired t-tests were used after each regression model to determine whether there were any differences in the relevant outcome variables with respect to the participants included and excluded from the model. All analyses were performed using Stata version 10.1 (College Station, TX), and α was set at P < 0.05.
Boys engaged in an average of 26.6 min of MVPA per day after school at time 1 with 45.6 min·d−1 of weekend MVPA. Girls engaged in 21.2 min of MVPA after school at time 1 and 36.3 min of weekend MVPA. Boys engaged in 4.6 fewer minutes of MVPA after school but 4.6 more minutes of weekend MVPA at time 2 than at time 1 with similar patterns for girls. A two-sample t-test indicated that there was no evidence of a difference in the baseline after-school MVPA of those with and without valid MVPA data at both time points (t = 0.052, P = 0.958). Two-sample t-tests indicated that the boys with valid data at both time points engaged in more weekend MVPA at baseline (51.70 vs 41.81 min·d−1, t = −3.04, P = 0.024) than those without valid data at both points with a similar pattern for girls (40.81 vs 32.3 min·d−1, t = −4.21, P < 0.001) (Table 1).
Time 1 regression models are presented in Table 2. Friend support for physical activity (t = 3.68, P = 0.001) was positively associated with boys’ weekday MVPA after school in a model that accounted for 13.6% of the overall variance. General friend support (t = 2.24, P = 0.037) and the number of friends (t = 2.17, P = 0.043) were positively associated with girls’ weekday MVPA after school in a model that accounted for 8.0% of the variance. The number of friends was positively associated with girls’ weekend MVPA (t = 2.26, P = 0.036) in a model that accounted for 5.5% of the variance. None of the friend variables were associated with boys’ weekend MVPA in a model that only accounted for 2.4% of the variance. Post hoc t-tests indicated that boys included in the analysis obtained more minutes of MVPA after school than those who were excluded because of missing variables (27.7 vs 22.8 min·d−1, t = −3.26, P = 0.001). There were no statistically significant (P < 0.05) differences between the girls who were included and excluded in the after-school MVPA at time 1 and no differences in the weekend MVPA or the boys and girls included and excluded in the weekend analysis at time 1 (P > 0.05).
Time 2 regression models are presented in Table 3. Friend support for physical activity was associated with boys’ after-school MVPA (t = 2.07, P = 0.049), girls’ after-school MVPA (t = 2.26, P = 0.032), and boys’ weekend MVPA (t = 2.11, P = 0.045) in models that accounted for between 5% and 6% of the overall variance. There were no statistically significant differences between the included and excluded samples for after-school or weekend MVPA at time 2.
The models that examined change in after-school and weekend MVPA are presented in Table 4. An increase in friend support for physical activity (t = 2.54, P = 0.02) and in the number of friends (t = 3.10, P = 0.006) was associated with an increase in girls’ MVPA after school during the move from primary to secondary school in a model that accounted for 4.4% of the variance. Similarly, both change in friend support for physical activity (t = 2.37, P = 0.029) and number of friends (t = 3.06, P = 0.006) were associated with an increase in girls’ weekend MVPA in a model that accounted for 9.7% of the variance. Post hoc tests showed that there was no statistically significant difference in the time 1 after-school MVPA of either the boys or the girls included and excluded from the regression model. Post hoc analysis indicated that boys included in the change in weekend MVPA regression obtained more minutes of MVPA at time 1 (53.8 vs 42.08 min·d−1, t = −3.41, P = 0.007) than those excluded with a similar pattern for girls (41.6 vs 33.3 min·d−1, t = −3.98, P < 0.001).
In this study, we found that boys engaged in approximately 26 min of MVPA after school while at primary school with girls engaging in 21 min. It is recommended that all children engage in 60 min of MVPA per day (2), and thus, in this sample, the after-school period only accounted for around a third of the recommended amount of MVPA. Weekends are also an important time for physical activity for primary school–aged boys and girls. In this study, boys and girls engaged in around 45 and 36 min of MVPA on weekends, respectively. The data presented here therefore indicate that the after-school period and weekend are key periods when youth could be more active and that strategies that focus on increasing MVPA during these times may be particularly beneficial.
There was a 16% decline in boys’ MVPA after school in their first year of secondary school compared with their primary school values, whereas girls’ after-school MVPA declined by around 12%. This finding is consistent with previous Scottish data that reported a 7% decline in self-reported physical activity after the transition from primary school to secondary school (14). It is, however, noticeable that both boys’ and girls’ mean levels of weekend MVPA increased after the move from primary to secondary school. The reasons for the differences in change in after-school and weekend MVPA are not immediately evident. However, it is possible to speculate that the increased weekend MVPA may reflect an increased license to engage in physical activity without adult supervision, which often occurs as children mature. The possible reasons for the decline in after-school MVPA are less clear and warrant further examination.Nevertheless, because the mean after-school and weekend MVPA were well below government-recommended levels at the end of primary school, a further decline in after-school MVPA after the move to secondary school is particularly concerning and suggests that the transition from primary to secondary school is a critical period for the maintenance of physical activity.
Understanding the factors that influence change in youth physical activity is an essential first phase of developing interventions to halt the decline during the transition to secondary school (1). This study was the first to investigate whether friend factors were associated with MVPA during this key period. The number of friends that girls had was positively associated with girls’ after-school and weekend MVPA at time 1 with change in the number of friends also associated with change in girls’ after-school and weekend MVPA. Inspection of the coefficients suggests that each additional friend that girls reported between primary and secondary school was associated with 3.7 more minutes of MVPA after school and 9.8 min of MVPA during the weekend. Because the average amount of after-school and weekend MVPA obtained by girls at time 1 was 21.2 and 36.3 min, respectively, the addition of a friend yielded increases of roughly 17% in after-school MVPA and 27% in weekend MVPA. These increases represent large proportional increases in the MVPA of children who are in general not meeting physical activity recommendations. Findings are consistent with our previous research in which we have shown that taking part in physical activity with friends, particularly outside of school, is associated with higher levels of physical activity among 11-yr-old girls (10). The absence of an association for the number of friends for boys may suggest that friendship effects are stronger for girls than boys. Collectively, the data reported here and elsewhere suggest that developing strategies for girls to engage with new friendship groups when they move to secondary school could have important effects on overall levels of physical activity.
Friend support for physical activity was cross-sectionally associated with boys’ after-school MVPA at time 1, boys’ after-school MVPA at time 2, boys’ weekend MVPA at time 2, and girls’ after-school MVPA at time 2. Change in friend support for physical activity was associated with change in girls’ after-school and weekend MVPA. Inspection of the coefficients indicates that each additional unit on the four-item friend support for physical activity scale was associated with 1.8 more minutes of after-school MVPA than the mean levels recorded at time 1, which would represent roughly a 5% increase in after-school MVPA. Equally, each additional unit on the friend support for physical activity was associated with 6.0 more minutes of weekend MVPA, roughly a 16% increase in the mean levels of weekend MVPA recorded at time 1. The findings are comparable to findings from the Trial of Activity in Adolescent Girls, which reported that friend support for physical activity was a mediator of girls’ physical activity (12). Previous research has reported that friend social support for physical activity is associated with adolescent physical activity (5) but did not examine whether associations differed by gender. As such, the data reported here extend current work to suggest that although friend support may be cross-sectionally associated with MVPA among boys and girls, there is a gender effect in terms of change in MVPA. Results seem to suggest that encouraging girls to support their friends’ physical activity behaviors could be important for helping girls to maintain physical activity levels. Further examination of why effects were not evident for boys is warranted.
In general, friend factors were modestly associated with girls’ physical activity with weaker associations for boys. The overall variance explained by these models ranged from 2% to 14%. This proportion of the variance is, however, comparable to studies that have examined psychological or environmental correlates of youth physical activity. For example, in a study of Belgian adolescents, psychosocial factors accounted for 25% of the variance in self-reported physical activity (7). In a study of US adolescents in Houston, TX, it was reported that objectively assessed physical environment only accounts for 3% of the variance in accelerometer-assessed MVPA (9). Collectively, these findings seem to indicate that there are many factors that affect youth physical activity, but each of these effects is on its own small. Therefore, interventions that operate on many levels and address multiple factors are therefore likely to be needed.
Strengths and limitations
The major strength of this study is the objective assessment of physical activity and self-reported friendship factors before and after the transition to secondary school, an important transition period. However, the participants were recruited from a single UK city, which hinders our ability to extrapolate findings to other contexts. We were also reliant on participant reports of friends’ influences, and as such, we have an assessment of perceived peer influences. It may have been the case that different associations would have been observed if we had obtained similar measures from peers, but it is impossible to know if those associations may have been stronger or weaker. It is also important to acknowledge that although accelerometers provide objective assessments of physical activity and can provide information about the intensity of physical activity, these cannot provide information about the type of activity in which a participant engages.
It is important to point out that because of the availability of complete data, the final sample included in the change analysis was markedly smaller than the time 1 sample, and post hoc tests indicated that the participants included in the change in weekend MVPA were more active than those excluded. As such, it is possible that our results are biased by inclusion of more active participants and need to be interpreted with caution. Moreover, to examine if there were gender differences or differences between weekend or after-school MVPA, we performed 12 separate analyses, which increases the likelihood that some of the associations were a function of chance. However, because the number of friends and friend support for physical activity were consistent predictors of girls’ MVPA, it seems unlikely that the importance of these variables was a spurious finding in this data set. Finally, the lack of change in some of the exposure variables may have limited our ability to detect change in MVPA during the transition to secondary school.
The amount of MVPA obtained after school and during the weekend declined by 12%–16% during the transition to secondary school. Having more friends and having friends who supported physical activity were associated with higher levels of physical activity and smaller reductions in MVPA among girls. The data support a need to develop strategies to prevent youth physical activity from declining after the move from primary to secondary school and suggest that fostering activity with friends and friend support for physical activity may be helpful in this process.
This work was supported by the National Prevention Research Initiative (G0501311) and World Cancer Research Fund (United Kingdom). This report is also a research arising from a Career Development Fellowship (to Dr. Jago) supported by the National Institute for Health Research. The views expressed in this publication are those of the authors and not necessarily those of the National Health Service, the National Institute for Health Research, or the Department of Health.
The authors have no conflicts of interest to declare.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Baranowski T, Jago R. Understanding mechanisms of change in children
’s physical activity programs. Exerc Sport Sci Rev. 2005; 33 (4): 163–8.
2. Cavill N, Biddle S, Sallis JF. Health enhancing physical activity for young people: statement of the United Kingdom expert consensus conference. Pediatr Exerc Sci. 2001; 13: 12–25.
3. Cole TJ, Freeman JV, Preece MA. Body mass index reference curves for the UK, 1990. Arch Dis Child. 1995; 73 (1): 25–9.
4. Cooper AR, Page AS, Wheeler BW, Hillsdon M, Griew P, Jago R. Patterns of GPS measured time outdoors after school and objective physical activity in English children
: the PEACH project. Int J Behav Nutr Phys Act. 2010; 7: 31.
5. Duncan SC, Duncan TE, Strycker LA. Sources and types of social support in youth physical activity. Health Psychol. 2005; 24 (1): 3–10.
6. Evenson KR, Catellier DJ, Gill K, Ondrak KS, McMurray RG. Calibration of two objective measures of physical activity for children
. J Sports Sci. 2008; 26 (14): 1557–65.
7. Haerens L, Craeynest M, Deforche B, Maes L, Cardon G, De Bourdeaudhuij I. The contribution of home, neighbourhood and school environmental factors in explaining physical activity among adolescents. J Environ Public Health. 2009; 2009: 320372.
8. Jago R, Baranowski T. Non-curricular approaches for increasing physical activity in youth: a review. Prev Med. 2004; 39 (1): 157–63.
9. Jago R, Baranowski T, Zakeri I, Harris M. Observed environmental features and the physical activity of adolescent males. Am J Prev Med. 2005; 29 (2): 98–104.
10. Jago R, Macdonald-Wallis K, Thompson JL, Page AS, Brockman R, Fox KR. Better with a buddy: influence of best friends on children
’s physical activity. Med Sci Sports Exerc. 2011; 43 (2): 259–65.
11. Kirkwood BR, Sterne AC. Essential Medical Statistics. London (UK): Blackwell; 2006. p. 343–70.
12. Lytle LA, Murray DM, Evenson KR, et al.. Mediators affecting girls’ levels of physical activity outside of school: findings from the trial of activity in adolescent girls. Ann Behav Med. 2009; 38 (2): 124–36.
13. Nader PR, Bradley RH, Houts RM, McRitchie SL, O’Brien M. Moderate-to-vigorous physical activity from ages 9 to 15 years. JAMA. 2008; 300 (3): 295–305.
14. Niven AG, Fawkner SG, Knowles AM, Henretty JM. From Primary to Secondary School: Changes in Scottish Girls’ Physical Activity and the Influence of Maturation and Perceptions of Competence: Research Report no. 113. Edinburgh (UK): Sport Scotland, 2009. p. 1–39.
15. Noble M, McLennan D, Wilkinson K, Whitworth A, Barnes H, Dibben C. The English Indices of Deprivation. London (UK): Department of Communities and Local Government, 2007. p. 1–129.
16. Page AS, Cooper AR, Griew P, Davis L, Hillsdon M. Independent mobility in relation to weekday and weekend physical activity in children
aged 10–11 years: the PEACH Project. Int J Behav Nutr Phys Act. 2009; 6 (2): 1–9.
17. Petersen AC, Crockett L, Richards M, Boxer A. A self-report measure of pubertal status: reliability, validity, and initial norms. J Youth Adolesc. 1988; 17: 117–33.
18. Riddoch CJ, Mattocks C, Deere K, et al.. Objective measurement of levels and patterns of physical activity. Arch Dis Child. 2007; 92 (11): 963–9.
19. Stice E, Shaw H, Marti CN. A meta-analytic review of obesity prevention programs for children
and adolescents: the skinny on interventions that work. Psychol Bull. 2006; 132 (5): 667–91.
20. Strong WB, Malina RM, Blimkie CJ, et al.. Evidence based physical activity for school-age youth. J Pediatr. 2005; 146 (6): 732–7.
21. Troiano RP, Berrigan D, Dodd KW, Mâsse LC, Tilert T, McDowell M. Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc. 2008; 40 (1): 181–8.
22. Trost SG, Loprinzi PD, Moore R, Pfeiffer KA. Comparison of accelerometer cut points for predicting activity intensity in youth. Med Sci Sports Exerc. 2011; 43 (7): 1360–8.
23. van Sluijs EM, McMinn AM, Griffin SJ. Effectiveness of interventions to promote physical activity in children
and adolescents: systematic review of controlled trials. BMJ. 2007; 335 (7622): 703.
24. Welk G, Blair SN, Wood K, Jones S, Thompson RW. A comparative evaluation of three accelerometry-based physical activity monitors. Med Sci Sports Exerc. 2000; 33 (9): S489–97.