Before-School Running/Walking Club and Student Physical Activity Levels:An Efficacy StudyMichalis Stylianou,1 Hans van der Mars,2 Pamela Hodges Kulinna,2 Marc A. Adams,2 Matthew Mahar,3

and Eric Amazeen2

1The University of Queensland; 2Arizona State University; 3East Carolina University

ABSTRACTPurpose: Before-school programs, one of the least studied student-related comprehensive schoolphysical activity program (CSPAP) components, may be a promising strategy to help youth meet thephysical activity (PA) guidelines. This study’s purpose was to examine: (a) how much PA childrenaccrued during a before-school running/walking club and during the school day, (b) whetherchildren compensated for the PA accumulated in the before-school program by decreasing theirschool-day PA, and (c) potential sex and body mass index (BMI) differences.Method: An alternatingtreatments design with a baseline phase was first conducted at a private school (School A) and wassubsequently replicated at a public school (School B). Participants (N ¼ 88) were 3rd- and 4th-gradechildren. The before-school program involved a running/walking club that met twice per week(School A: 20min; School B: 15min). PA was measured using the NL-1000 pedometer. Data analysisincluded multilevel modeling and visual analysis. Results: Children accumulated substantialamounts of PA in the before-school programs (School A: 1,731 steps, 10:02 moderate-to-vigorous PAminutes or 50% of program duration; School B: 1,502 steps, 8:30 moderate-to-vigorous PA minutes or57% of program duration). Additionally, children did not compensate by decreasing their school-dayPA on days they attended the before-school program. Sex differences were found in before-schoolprogram PA only for School B and in school-day PA for both schools. No BMI differences were found.Conclusions: Before-school programs, as part of CSPAPs, can help children increase their PAwithout resulting in decreased school-day PA and without taking time away from academics.

ARTICLE HISTORYReceived 5 August 2015Accepted 20 June 2016

KEYWORDSBefore-school physicalactivity programming;elementary school;pedometers; physicalactivity compensation

Despite public health concerns and the extensive healthbenefits of physical activity (PA; Janssen & LeBlanc,2010), a large proportion of American youth do not meetthe national PA guidelines (Troiano et al., 2008). Schoolshave been identified as primary PA promotion sites(Institute of Medicine [IOM], 2013) because they canreach the vast majority of school-aged youth, and theyoften have the needed facilities as well as personnelwho, with sufficient training, can define PA policies andfacilitate PA programs. However, youth spend most oftheir time in school being sedentary (i.e., sitting; Abbott,Straker, & Mathiassen, 2013). Additionally, school PAopportunities (i.e., physical education and recess) havedecreased in the last few years due to an increasedemphasis on boosting academic performance (Center onEducation Policy [CEP], 2007). Thus, it is unlikely thatyouth can meet the PA guidelines through physicaleducation (PE) and/or recess alone.

Multifaceted and coordinated school-based programs(often called comprehensive or whole-school programs)

have been identified as a promising approach forincreasing PA in youth (e.g., IOM, 2013). One suchprogram is the comprehensive school physical activityprogram (CSPAP; Centers for Disease Control andPrevention [CDC], 2013), which includes five com-ponents, one of which is PA before and after school.Relevant research, however, has largely focused on after-school programs, possibly because they are moreprevalent.

Before-school physical activity programs

Before-school PA programs refer to any school-basedprogram that targets PA promotion before the start of theschool day. In the CSPAP Policy Continuum document(National Association for Sport and Physical Education[NASPE], 2012), the optimal policy related to before-school and after-school programs requires the provisionof 30min to 60min of PA, of which 50% should be spentin moderate-to-vigorous PA (MVPA). However, based

q 2016 SHAPE America

CONTACTMichalis Stylianou m.stylianou@uq.edu.au School of Human Movement and Nutrition Sciences, The University of Queensland, St Lucia, QLD 4072,Australia.

RESEARCH QUARTERLY FOR EXERCISE AND SPORT2016, VOL. 87, NO. 4, 342–353http://dx.doi.org/10.1080/02701367.2016.1214665

on relevant study findings, organized activities rarelyoccurred before school in elementary and middle schools(McKenzie, Marshall, Sallis, & Conway, 2000; Turner,Johnson, Slater, & Chaloupka, 2014), and only a smallproportion of students visited activity areas before school(McKenzie et al., 2000). Yet, evidence from descriptivestudies supports the potential of before-school time tosignificantly contribute to children’s PA (McKenzie,Crespo, Baquero, & Elder, 2010; McKenzie et al., 2000;Tudor-Locke, Lee, Morgan, Beighle, & Pangrazi, 2006).Further, there is preliminary evidence that children canaccumulate substantial MVPA during a campus-based,before-school program using an interactive multimediaPA training system (Mahar, Vuchenich, Golden, DuBose,& Raedeke, 2011). However, the specific interventionrequired a significant financial investment, which maymake it a less feasible option for schools.

A practical and cost-effective alternative for a campus-basedbefore-school PAprogrammaybe a running/walkingclub. Running and walking are lifetime activities andgenerate important health and other benefits (Lambiase,Barry, & Roemmich, 2010; Schnohr, Marott, Lange, &Jensen, 2013). The purpose of running/walking clubs is “tohelp children improve their ability to sustain continuousrunning and walking, identify walking and running asbeneficial cardiorespiratory exercise, and participate indaily walking and running outside of regularly scheduledphysical education class” (Ratliffe & Bostick, 2001, p. 24).

There has been substantial interest in assessing theefficacy of campus-based running/walking club programs(e.g., Foshay & Patterson, 2010; Ratliffe & Bostick, 2001;Xiang, McBride, & Bruene, 2006). However, studentoutcomes in relevant studies were mostly physiologicaland motivational in nature, and the interventions weredelivered mainly during school-day PE and recess. Onerunning/walking club study focused on PA (Stylianou,Kulinna, & Kloeppel, 2014), but it did not use anobjective measure. Hence, there is a need to further studythe potential contributions of running/walking clubs tostudents’ PA.

Physical activity compensation

An issue related to the value of school-based PA programsis the potential of children’s “PA compensation”—that is,whether children engage in less school-day or after-schoolPA on days with increased school-based PA program-ming, or, on the other hand, if children engage in increasedafter-school PA on school days with low PA. The conceptof compensation is based on the biological basis ofPA (Rowland, 1998), which suggests that compensatorychanges occur to maintain a stable level of PA or energyexpenditure.

There is some evidence that youth do not compensatefor increased school-based PA opportunities by engagingin less school-day PA (Mahar et al., 2011) or after-schooland daily PA (Alderman, Benham-Deal, Beighle, Erwin,& Olson, 2012; Long et al., 2013). Similarly, there isevidence that youth do not compensate for restrictedschool-based PA opportunities by engaging in moreschool-day PA (Morgan, Beighle, & Pangrazi, 2007) orafter-school PA (Dale, Corbin, & Dale, 2000). Further,children have been found to engage in more after-schooland daily PA on days with increased school-based PA(Alderman et al., 2012; Dale et al., 2000; Long et al.,2013). Collectively, it appears that expanded school-based PA opportunities can significantly help youth meetand/or exceed PA guidelines.

Research on PA compensation during the school dayis limited. However, given that about 35% to 40% of U.S.youth’s daily PA occurs during school hours (Brusseau& Hannon, 2013) and that recess, a discretionary PAperiod, is a significant contributor to youth’s school-dayand daily PA (Erwin et al., 2012), it is important tofurther examine PA compensation during the school day.

Purpose

Given the scarcity of studies focusing on before-schoolPA programs as well as running/walking clubs, thepurpose of this efficacy study was to examine: (a) howmuch PA (i.e., steps and MVPA) children receivedduring a before-school running/walking club and theschool day, and (b) whether children compensated forthe PA they received in the before-school running/walking club by being less active during the school day(excluding the before-school program). Further, asecondary purpose of this study was to examine sexand body mass index (BMI) differences in participants’before-school program and school-day PA.

Methods

Participants and settings

This study’s participants were third- and fourth-gradestudents from two schools in the Southwestern UnitedStates. The two schools were a purposive sample ofschools that were interested in participating in the studyand represented two different settings (i.e., private vs.public). The goal was not to compare the two schools, butrather to replicate the study across the two settings.

School A

School A was a K–8 private school with a total enrollmentof 273 students (primarily Caucasian). From this school,

BEFORE-SCHOOL PHYSICAL ACTIVITY PROGRAM 343

39 students provided parental consent and assent formsto participate in the study (third grade ¼ 16, fourthgrade ¼ 23; boys ¼ 14, girls ¼ 25; 89.74% Caucasian,10.26% Other). About one quarter of the participants(23.68%) were overweight or obese (i.e.,$85th percentile,based on the CDC’S BMI-for-age growth charts for boysand girls). At this school, students received two 45-min PElessons per week, as well as daily recess during lunchtime(45min for both lunch and recess). Other PAopportunitiesat this school included various after-school sport clubs (e.g.,soccer, etc.). According to student reports, active commut-ing to/from school was nonexistent among studyparticipants at this setting.

School B

School Bwas aK–6 public school with a total enrollment of451 students (57.43% Caucasian, 33.92% Hispanic, 8.65%Other), ofwhom60%were eligible for free or reduced-pricelunch. From this school, 56 students provided parentalconsent and assent forms to participate in the study (thirdgrade ¼ 28, fourth grade ¼ 28; boys ¼ 31, girls ¼ 25;53.57% Caucasian, 35.71% Hispanic, 10.71% Other).Twenty of the participants (43.49%) were overweight orobese (i.e., $85th percentile, based on the CDC’S BMI-for-age growth charts for boys and girls). At this school,students received two 30-minPE lessons perweek and dailyrecess at lunchtime (40min for both lunch and recess).Other PA opportunities at this school included an extrarecess period per week and a Fit Kids Club once a week.About 10% of this school’s participants reported activelycommuting to/from school occasionally.

Research design

This study employed a two-phase design with an initialbaseline phase followed by an alternating treatments (AT)phase. AT designs originate from applied behavior analysisand single-case experimental designs (Barlow & Hersen,1984; Cooper, Heron, & Heward, 2007). They test therelative effectiveness of two (or more) treatments/conditions and are an alternative to the traditionalbetween-group comparison designs. In AT designs, aperson (or a group) serves as its own control (i.e., the sameparticipants receive all conditions), which helps control formost threats against internal validity (Barlow & Hersen,1984), including selection, history, and maturation. More-over, the initial baseline phase helps control for regressionto the mean as well as for testing effects, which constituteparticular threats in designs with repeated measures. Thisdesign combined with the replication of the study in twodifferent settings can provide strong evidence for bothinternal and external validity, as well as program efficacy.

Phases, conditions, and number of data points

The two phases used in the study (i.e., baseline and AT)lasted 2 and 5 weeks, respectively, although the numberof data points collected was different at each school.The conditions compared in the AT phase included anon-treatment/control condition (no before-school pro-gram as in the baseline phase) and a treatment condition.

The baseline phase included 5 data points for School A(1 week for each grade level) and 9 data points for SchoolB (2 weeks; 1 day was a holiday). During the AT phase,at School A, 1 treatment data point (i.e., Tuesday orThursday) and 1 non-treatment/control data point (i.e., aday before or after a treatment point) were collected foreach grade level each week for 5 weeks, resulting in a totalof 10 data points. At School B, data were collected dailyfor 5 weeks during the AT phase except for 1 week with aholiday, resulting in a total of 24 data points. Specifically,at School B, 2 treatment data points (i.e., Tuesdaysand Thursdays) and 3 non-treatment/control data points(i.e., the remaining days) were collected each week. Thenumber of data points for each phase was decided basedon practical considerations (e.g., predetermined pro-gram-starting day, natural school breaks).

Condition sequencing and discrete conditions

The fact that the before-school programs occurred onspecific weekdays did not allow for randomly counter-balancing the non-treatment and treatment conditions,which can control for order effects. Such limitations are notunusual when conducting research in schools. However, atSchool A, non-treatment data points were manipulated sothey occurred both on days before and after treatment datapoints (at least two times before and two times after).At School B, treatment data points occurred both beforeand after non-treatment data points every week. Also, thetwo conditionswere clearly discrete,whichhelps control forpotential order effects and minimize possible carryovereffects (Barlow & Hersen, 1984).

Intervention: Before-school physical activityprogram

The before-school program in both schools involved arunning/walking club that occurred two times a week.For this study, students were considered to haveparticipated in the before-school program if they hadaccumulated at least 5min of MVPA.

School A

At School A, the program lasted 20min each time. At thisschool, the PE teacher used a reinforcement system that

344 M. STYLIANOU ET AL.

rewarded students for participating in the program. Theteacher monitored the distance students covered in theprogram and the students received “shoe”-shaped tokensfor their shoestrings or backpacks for every 8 kilometers(5 miles) they covered.

School B

At School B, the program lasted 15min each time. ThePE teacher at this school also used a reinforcementsystem to reward students for participating in theprogram. Specifically, students received a pencil for everytwo laps completed as well as a “caught being good” ticket(part of the school accountability system) each time theyparticipated in the program.

Data collection and procedures

Institutional review board approval as well as district andprincipal approvals were obtained prior to starting thestudy. Also, student assent and parental consent formswere collected. Data were collected on: (a) before-schoolPA participation, (b) PA levels during the before-schoolprogram and the school day, and (c) anthropometricmeasures (height and weight).

Before-school PA participation

Every morning during the study, participants completed abefore-school PA participation log, where they reportedwhether they participated in any PA before school. Thisinformation was used to determine fidelity to the twoconditions (e.g., whether participants engaged in any PAbefore school on baseline or non-treatment/control days).

Physical activity levels

In both schools, steps and MVPA time were assessedduring: (a) the running/walking club and (b) the schoolday. For the running/walking club, steps and minuteswere recorded right before students started running/walking as well as at the end of the program or when theydiscontinued participation. For school-day PA, steps andminutes were recorded at the beginning of homeroomtime and the end of the school day.

PA was measured using the New Lifestyles NL-1000pedometer, which uses a piezoelectric mechanism that issimilar to accelerometers. This instrument was set torecord activity greater than 3.6 metabolic equivalents,and the sampling interval was 4 s, which is consideredsuitable for children’s sporadic PA patterns. This devicehas been shown to provide valid and reliable estimates ofPA in children (Hart, Brusseau, Kulinna, McClain, &Tudor-Locke, 2011).

Prior to data collection, students were instructed inhow to use the pedometer and had the opportunity to useit during two PE lessons to avoid reactivity effects.Graphics that were posted in the school also served tovisually remind students where to place their pedometerand how to use it. Proper placement of the instrumentwas on the right hip in line with the midline of the thigh.Elastic belts were also available for students who had adifficult time adjusting the pedometer to their attire.Shake tests were conducted prior to the start of the studyas well as every 2 weeks during the study to evaluatepedometers for calibration problems.

Anthropometric measures

Height and weight measurements were obtained withoutshoes and heavy clothing using a calibrated digital scale(Seca 882Digital Scale) and stadiometer (Seca 214 PortableStadiometer). These measurements were taken during thefirst 2 weeks of the study at each school and were used tocalculate students’ BMI (weight [kg]/height [m] squared)and BMI-for-age percentile using the CDC’s BMI tool forschools. Subsequently, BMI-for-age percentiles were usedto classify students as normal-weight (,85th percentile) oroverweight/obese ($85th percentile) based on the CDC’sBMI-for-age growth charts for boys and girls.

Data analysis

Statistical analysis

Analyses were performed for both steps and MVPA timewithin a multilevel modeling framework, with dailyobservations as the Level 1 variable (subscript i inequations) and person-level variables (e.g., sex, BMI status)as Level 2 variables (subscript j in equations). Analyseswereconducted separately for each school because the purposewas to replicate the study in two different settings. Analyseswere performed using the IBM Statistical Package for theSocial Sciences software (Version 21).

To quantify the magnitude of potential effects, thepseudo-R 2 effect size was calculated (Raudenbush & Bryk,2002; Singer & Willett, 2003), which is interpreted as theproportion reduction in variance for a parameter estimatethat results from comparing the variance component (i.e.,residual/Level 1 variance, intercept/Level 2 variance) in abaseline model to the same variance component in a fullermodel (i.e., a model with more/all predictor variables).This statistic is analogous to the R 2 statistic in multipleregression and is estimated through the formula

Pseudo2 R2 ¼ s2BASE 2 s2

FULL

� �=s2

BASE; ð1Þwhere s2

BASE and s2FULL are the dependent variable

variances in the baseline model and the full model,

BEFORE-SCHOOL PHYSICAL ACTIVITY PROGRAM 345

respectively. For interpreting pseudo-R 2 values, Cohen’scriteria for R 2 values were used (i.e., .02, .13, and .26represent small, medium, and strong effects, respectively;Cohen, 1988).

Running/walking club PA. To determine the meannumber of steps and MVPA time accumulated withinthe running/walking club, a model with no predictorswas tested,

yij ¼ B0 þ u0j þ eij: ð2ÞIn this model, B0 represents the mean values of thedependent variable, u0j reflects variation in dependentvariable means across students (between-person var-iance), and eij is leftover variability across observations(within-person variance).

School-day PA. To examine the mean number of stepsand MVPA minutes accumulated during the school day,the same analyses were conducted as described above.

Sex and BMI status differences. To examine potentialdifferences in before-school program and school-day PA,two different models were tested, one with sex and onewith BMI status. For example, the model with sex was,

yij ¼ B0 þ B1ðSexjÞ þ u0j þ eij; ð3Þwhere B0 is the expected number of steps or MVPA timefor boys (sex ¼ 0), B1 is the change in PA for girls(sex ¼ 1), and u0j reflects variation in PA means acrossstudents not captured by sex.

Compensation. To determine if students compensatedon days they attended the before-school program byengaging in reduced school-day PA, a random interceptmodel was tested,

yij ¼ B0 þ B1 ðtreatmentijÞ þ u0j þ eij; ð4Þwhich allows intercepts (i.e., mean steps or MVPA time)to vary across individual students but assumes constantslopes (i.e., same influence of treatment). In this model,B0 reflects the expected number of steps or MVPA timein the absence of treatment (treatment ¼ 0), B1 is thechange in school-day PA for a 1-unit change in treatment(i.e., participation in the program), u0j reflects variationin PA means across students, and eij is leftover variabilitynot captured by treatment.

Building toward this model, some preliminaryanalyses were conducted. These analyses included testingfor the influence of PE and extra recess (School B only)on day-to-day PA,

yij ¼ B0 þ B1 ðPEijÞ þ B2 ðextra recessijÞ u0j þ eij; ð5Þto determine if they needed to be included in subsequentmodels. Preliminary analyses also included testing forpotential effects of the variables of phase and order on

day-to-day PA over and above treatment to determine ifthey needed to be included as covariates in subsequentmodels. These variables were tested in separate modelsbecause order was only present during the AT phase. Forinstance, the equation for the model including phase was:

yij ¼ B0þB1 ðtreatmentijÞ þB2 ðphaseijÞþ u0j þ eij: ð6Þ

Visual analysis

A visual analysis was also conducted to examine the PAcompensation question. For the visual analysis, graphs ofaverage school-day PA (excluding before-school programPA) were developed by phase and condition for eachschool. The visual analysis of the graphically plotted datawas based on the following criteria: variability and trendswithin and between phases/conditions, data overlapbetween phases/conditions, immediacy of change fromone phase/condition to the next, and distance betweendata paths of the different conditions (Cooper et al., 2007).Evidence for compensatory effects would include a clearand consistent reduction in school-day PA with theintroduction of the treatment condition, a clear distancebetween the data paths of the treatment and non-treatment conditions, and a lack of overlap between thedata paths of the two conditions. A large overlap of datapaths would reflect the absence of a compensatory effect.

Results

At School A, all 39 students who agreed to participate inthe study attended the before-school program at leastonce and thus were all included in the analyses. At SchoolB, however, 7 students who initially provided consent andassent forms to participate in the study did not attend theprogram at all and were thus excluded from analyses,resulting in a final sample of 49 students from this school.There were no significant differences between thenumber of students who did and did not attend theprogram in School B in terms of sex (Fisher’s Exact test,p ¼ .22) and BMI status (Fisher’s Exact test, p ¼ .51).Also, multilevel models showed that attending theprogram during the AT phase was not a significantpredictor of, and explained small percentages of variancein, baseline school-day PA when controlling for sex andBMI status (steps, p ¼ .15, pseudo-R 2 ¼ .02; MVPA,p ¼ .22, pseudo-R 2 ¼ .02).

Before-school PA participation

According to the before-school PA logs, in about 97%of thecases, participants did not engage in any PA before schoolon days without the program, which is not surprising given

346 M. STYLIANOU ET AL.

the early school start times and that sunrise (in late fall/winter) was close to the school start times.

Running/walking club and school-day physicalactivity

Mean steps andMVPAminutes accumulated in the before-school program and during the school day at each schoolare available in Tables 1 and 2, respectively. These tablesalso present PA data by sex and BMI status. Additionally,Table 1 provides information about the percentageof program duration spent in MVPA, as well as thepercentages of daily PA guidelines (12,000 steps/day[Adams, Johnson, & Tudor-Locke, 2013]; 60min/day [U.S.Department of Health and Human Services (USDHHS),2008]) and school-day PA that the before-school programsteps andMVPA time represent. Similarly, Table 2 includesinformation about the percentages of daily PA guidelinesthat the school-day steps andMVPA time represent, as wellas the percentages of daily PA guidelines that the combinedschool-day and before-school program steps and MVPAtime represent.

Sex and BMI status differences

Tables 3 and 4 present the results of sex and BMI statusmodels, respectively. BMI status was not a significantpredictor for before-school program or school-day PA ineither school and explained less than 5% of variance in

relevant models. Sex was a significant predictor forbefore-school program PA only for School B andexplained about 23% of variance in both steps andMVPA. Sex was a significant predictor for school-day PAin both schools but explained different proportions ofvariance for each school (see Table 3). In all modelswith significant sex differences, boys were more activethan girls.

Compensatory effects

According to the results of preliminary analyses, PE (forboth schools) and extra recess (only for School B) bothsignificantly contributed to the prediction of school-dayPA when controlling for each other. Additionally, whenexamining the confounding variables of phase and order,only phase significantly contributed to the prediction ofschool-day PA when controlling for treatment, PE, andextra recess. Therefore, the variables of PE, extra recess(only for School B), and phase were included insubsequent models.

Based on the random intercept model results (seeTable 5), before-school program participation did notsignificantly contribute to, and explained very smallproportions of variance in, school-day steps or MVPAtime (excluding before-school program PA) over andabove PE and phase at School A, thus suggesting a non-compensatory effect on days students attendedthe program. At School B, before-school program

Table 1. Running/walking club physical activity levels by school, sex, and body mass index (BMI) status.

Running/walking club steps

M WPSD BPSD % of daily step guideline % of school-day stepsa

School A 1,731 567 473 14.43 30.74Boys 1,953 686 398 16.28 34.68Girls 1,620 493 479 13.50 28.76BMI , 85 PCTL 1,760 564 432 14.67 31.25BMI $85 PCTL 1,645 583 604 13.71 29.21

School B 1,502 372 341 12.52 31.02Boys 1,645 389 346 13.71 33.97Girls 1,297 343 206 10.81 26.79BMI , 85 PCTL 1,506 375 378 12.55 31.10BMI $85 PCTL 1,491 368 275 12.43 30.79

Running/walking club MVPA time (minutes)

M WPSD BPSD % of program duration % of daily MVPA rec. % of school-day MVPAa

School A 10:02 3:20 2:59 50.00 16.72 47.03Boys 11:00 3:49 2:37 55.00 18.33 51.56Girls 09:32 3:02 3:07 47.67 15.89 44.69BMI , 85 PCTL 10:09 3:17 2:46 50.75 16.92 47.58BMI $85 PCTL 09:40 3:31 3:45 48.33 16.11 45.31

School B 08:30 2:24 2:07 56.67 14.17 42.50Boys 09:23 2:36 2:10 62.56 15.64 46.92Girls 07:13 2:02 1:13 48.11 12.03 36.08BMI , 85 PCTL 08:36 2:18 2:09 57.33 14.33 43.00BMI $85 PCTL 08:26 2:27 2:08 56.22 14.06 42.17

Note. Daily step guideline ¼ 12,000 steps (Adams et al., 2013); daily moderate-to-vigorous physical activity (MVPA) recommendation ¼ 60min(USDHHS, 2008); M ¼ mean; WPSD ¼ within-person standard deviation; BPSD ¼ between-person standard deviation; PCTL ¼ percentile;Rec. ¼ recommendation.aBased on school-day PA on treatment days.

BEFORE-SCHOOL PHYSICAL ACTIVITY PROGRAM 347

participation was found to significantly contribute to theprediction of school-day steps and MVPA time whencontrolling for PE, extra recess, and phase. Specifically,students at School B accumulated a significantly highernumber of steps (B1 ¼ 328) and MVPA time(B1 ¼ 1:25min) on days they attended the before-schoolprogram, but program participation explained very small

proportions of the variance in school-day PA (pseudo-R 2 , .02 [small]).

Visual analysis

Figures 1 and 2 present graphs of average school-day stepsand MVPA across phases and conditions for each school

Table 2. School-day physical activity levels by school, sex, and body mass index (BMI) status.

School-day steps

Alternating treatments phase

Baseline phase Non-treatment Treatment % of daily stepguideline

a% of daily step

guideline w/ RWC stepsa

M WPSD BPSD M WPSD BPSD M WPSD BPSD

School A 6,612 1,129 1,647 5,868 1,287 1,332 5,632 1,241 1,159 46.93 61.36Boys 7,706 1,361 1,865 6,969 1,297 1,533 6,605 1,320 1,408 55.04 71.32Girls 6,038 1,008 1,191 5,335 1,291 834 5,103 1,251 926 42.53 56.03BMI , 85 PCTL 6,666 1,161 1,812 6,002 1,347 1,393 5,879 1,297 1,163 48.99 63.66BMI $85 PCTL 6,452 1,024 1,070 5,510 1,170 1,162 4,766 1,006 604 39.72 53.43

School B 4,388 1,317 666 4,517 1,236 713 4,842 1,043 1,372 40.35 52.87Boys 4,604 1,447 726 4,739 1,348 725 5,018 1,142 1,496 41.82 55.53Girls 4,081 1,107 423 4,193 1,054 577 4,584 856 1,162 38.20 49.01BMI , 85 PCTL 4,670 1,407 713 4,823 1,418 744 5,322 1,126 1,498 44.35 56.90BMI $85 PCTL 4,226 1,262 594 4,342 1,109 648 4,555 993 1,240 37.96 50.38

School-day MVPA time (minutes)

Alternating treatments phase

Baseline phase Non-treatment Treatment% of dailyMVPA rec.

a % of daily guidelinew/ RWC PA

a

M WPSD BPSD M WPSD BPSD M WPSD BPSD

School A 24:37 6:41 6:40 21:27 6:07 6:28 21:20 6:33 5:59 35.56 52.28Boys 30:26 8:47 4:20 27:24 6:49 8:07 28:11 7:34 5:04 46.97 65.31Girls 22:03 5:38 5:47 18:47 5:49 3:01 17:49 6:09 4:12 29.69 45.58BMI , 85 PCTL 24:54 6:50 7:15 22:11 6:04 7:02 22:40 6:56 6:29 37.78 54.69BMI $85 PCTL 23:49 6:11 4:52 19:34 6:16 4:37 16:56 4:49 3:42 28.22 44.33

School B 18:03 6:32 3:41 18:26 6:35 4:26 20:00 6:25 7:19 33.33 47.50Boys 19:35 7:17 3:38 20:04 7:18 4:39 21:36 6:06 8:20 36.00 51.64Girls 15:47 5:13 2:24 16:03 5:24 2:46 17:35 4:04 4:43 29.31 41.33BMI , 85 PCTL 19:29 7:11 3:56 20:07 8:03 5:25 22:04 6:14 8:24 36.78 51.11BMI $85 PCTL 17:13 6:07 3:21 17:29 5:29 3:30 18:45 4:54 6:19 31.25 45.31

Note. Daily step guideline ¼ 12,000 steps (Adams et al., 2013); daily moderate-to-vigorous physical activity (MVPA) recommendation ¼ 60min (USDHHS, 2008);M ¼ mean; WPSD ¼ within-person standard deviation; BPSD ¼ between-person standard deviation; RWC ¼ running/walking club; PCTL ¼ percentile;Rec. ¼ recommendation.aBased on school-day PA on treatment days.

Table 3. Sex differences in before-school program and school-day physical activity.

School A

Steps MVPA (minutes)

Estimate SE p Variance explained Estimate SE p Variance explained

Before-school RWC PAB0 (Intercept) 1,924 157 10:52 00:59B1 (Sex) 2295 195 .139 6.60% 201:18 01:14 .295 1.10%

School-day PAB0 (Intercept) 7,201 318 28:24 01:24B1 (Sex) 21,693 393 , .001 36.19% 208:51 01:42 , .001 46.54%

School B

Steps MVPA (minutes)

Estimate SE p Variance explained Estimate SE p Variance explained

Before-school RWC PAB0 (Intercept) 1,646 63 09:23 00:24B1 (Sex) 2356 100 .001 22.59% 202:12 00:38 .001 23.02%

School-day PAB0 (Intercept) 4,795 169 20:01 00:47B1 (Sex) 2583 216 .024 10.49% 203:45 01:14 .004 16.30%

Note. MVPA ¼ moderate-to-vigorous physical activity; RWC ¼ running/walking club; PA ¼ physical activity; SE ¼ standard error.

348 M. STYLIANOU ET AL.

(excluding before-school program PA). According to thegraphs, there is substantial overlap between baseline andnon-treatment data paths for both step counts andMVPAtime across both schools.Moreover, in theATphase, thereis complete overlap between the treatment and non-treatment condition data paths (including the standarddeviation bars) for both schools. Collectively, based on thegraphs, children did not compensate for the PA theyreceived in the before-school running/walking club bybeing less active during the school day.

Discussion

This study aimed to examine the efficacy of a before-schoolrunning/walking club and focused on the PA (i.e., stepsand MVPA) accumulated in the before-school program,whether children compensated for the PA they received in

the program by being less active during the school day,and potential sex and BMI status differences. The studywas first conducted at a private school (School A) and wassubsequently replicated at a public school (School B).

Physical activity in the before-school program

Children from both schools accumulated substantialamounts of PA in the before-school program (1,731steps, 10:02 MVPA minutes at School A [20-minprogram]; 1,502 steps, 8:30 MVPA minutes at School B[15-min program]). The PA levels accumulated in bothprograms met or exceeded the standard of 50% of before-school program time spent in MVPA (50% at SchoolA and 56.67% at School B), as identified in the CSPAPPolicy Continuum document (NASPE, 2012). Theseresults are comparable to the results of Mahar et al.

Table 4. BMI status differences in before-school program and school-day physical activity.

School A

Steps MVPA (minutes)

Estimate SE p Variance explained Estimate SE p Variance explained

Before-school RWC PAB0 (Intercept) 1,754 110 10:07 00:41B1 (BMI Status) 298 226 .666 3.04% 200:25 01:23 .765 3.28%

School-day PAB0 (Intercept) 6,270 263 23:22 01:13B1 (BMI Status) 2695 517 .187 1.93% 203:17 02:22 .176 2.21%

School B

Steps MVPA (minutes)

Estimate SE p Variance explained Estimate SE p Variance explained

Before-school RWC PAB0 (Intercept) 1,508 69 08:26 00:26B1 (BMI Status) 217 114 .882 2.24% 200:11 00:43 .814 2.61%

School-day PAB0 (Intercept) 4,534 180 19:15 00:48B1 (BMI Status) 2246 216 .260 4.30% 202:27 01:19 .069 5.44%

Note. BMI ¼ body mass index; MVPA ¼ moderate-to-vigorous physical activity; RWC ¼ running/walking Club; PA ¼ physical activity;SE ¼ standard error.

Table 5. School-day physical activity random intercept model results.

School A

Steps MVPA (minutes)

Estimate SE p Variance explained Estimate SE p Variance explained

B0 (Intercept) 6,181 242 22:30 01:08B1 (Treatment) 2138 155 .375 2.70% 200:06 00:48 .901 1.14%B2 (Physical education) 1,008 123 , .001 14.29% 04:58 00:38 , .001 12.89%B3 (Phase) 2594 132 , .001 6.19% 202:28 00:41 , .001 2.87%

School B

Steps MVPA (minutes)

Estimate SE p Variance explained Estimate SE p Variance explained

B0 (Intercept) 3,658 128 14:30 00:43B1 (Treatment) 328 72 , .001 1.41% 01:25 00:24 , .001 0.83%B2 (Physical education) 1,009 56 , .001 9.24% 04:55 00:19 , .001 8.28%B3 (Extra recess) 1,958 80 , .001 21.44% 09:05 00:27 , .001 16.95%B4 (Phase) 209 63 .001 .70% 00:48 00:21 .024 0.29%

Note. MVPA ¼ moderate-to-vigorous physical activity; SE ¼ standard error.

BEFORE-SCHOOL PHYSICAL ACTIVITY PROGRAM 349

(2011), who found that children participating in a before-school program using an interactive multimedia PAtraining system spent an average of 46.4% (9.3min) oftheir time in the program in MVPA.

The PA accumulated through the two programsrepresents substantial portions of the daily step guidelineand MVPA recommendation (14.43% and 12.52% ofthe daily step guideline for Schools A and B, respectively;16.72% and 14.17% of the daily MVPA recommendationfor Schools A and B, respectively). The importanceof these numbers lies in the short duration of theseprograms (20min and 15min, respectively) and is furtherhighlighted when considered relative to the school-dayPA levels of children at the two schools (for School A,30.74% and 47.03% of school-day steps and MVPA time,respectively; for School B, 31.02% and 42.50% of school-day steps and MVPA time, respectively). Moreover, whencombining the before-school program and school-dayPA (see Table 2), it is clear that the running/walking clubhas the potential to help children meet and/or exceed at

least half of the daily step guideline and MVPA timerecommendation.

The contribution of the before-school program tochildren’s PA should also be considered in light of thecurrent realities of schools. Although schools areconsidered ideal PA promotion sites (IOM, 2013), schoolPA opportunities have decreased during the last few years(CEP, 2007). The two schools that participated in thisstudy provided students with daily recess but only twoperiods of PE each week. Further, active commuting to/from school was nonexistent at School A and very limitedat School B. It is not surprising, therefore, that the before-school programwas a significant source of these children’sdaily PA, which points to the importance of providingexpanded PA programming for all school-aged students.

Compensatory effects

In this study, participation in the before-school programdid not have a compensation effect on school-day PA.

Figure 1. School-day steps (excluding before-school program steps) during the baseline and alternating treatment (ATP) phases.

350 M. STYLIANOU ET AL.

In other words, children did not engage in less school-dayPA on days when they attended the before-schoolprogram. On the contrary, there was a small andstatistically significant increase in the PA children atSchool B accumulated on days they attended the program(328 steps, 01:25 MVPA minutes). However, themagnitude of this effect was small and should beinterpreted with caution. Mahar et al. (2011) also foundno significant differences in school-day PA between dayschildren did and did not attend a before-school program.

This study’s results contradict the hypothesis ofcompensatory changes to maintain a stable level of PA orenergy expenditure at the school-day level, but they mayalso be a function of other factors, including theparticipants’ age, the fact that the study focused onMVPA and not on vigorous or total PA, and the relativelyshort duration of the study. Further, children may stillcompensate after school, although the findings of relevant

studies suggest that youthdonot compensate for decreased/increased school-based PA opportunities by increasing/decreasing their after-school or daily PA (Alderman et al.,2012; Dale et al., 2000; Long et al., 2013). Further, there isevidence that children may engage in more after-schooland daily PA on days with increased school-based PA(Alderman et al., 2012; Dale et al., 2000; Long et al., 2013).

Sex and BMI status differences

In this study, sex was a significant predictor of, andexplained a substantial proportion of variance in, school-day PA. Sex differences are common in the literature (e.g.,Troiano et al., 2008) and highlight the need to particularlyfocus on girls’ PA patterns from a young age. Results aboutsex differences in before-school program PA were mixedand should be further investigated in the future. Similarly,some descriptive studies have shown no sex differences in

Figure 2. School-day MVPA (excluding before-school program MVPA) during the baseline and alternating treatment (ATP) phases.

BEFORE-SCHOOL PHYSICAL ACTIVITY PROGRAM 351

before-school PA (McKenzie et al., 2010; Tudor-Lockeet al., 2006),whereasMcKenzie et al. (2000) found that boysengaged in more MVPA than girls before school. BMIstatus was an insignificant predictor of, and explainedminimal proportions of variance in, both before-schoolprogram and school-day PA. Related studies suggest thatbody fat but not BMI may be associated with PA in youthand that body fat may correlate with vigorous but notmoderate PA (e.g., Abbott & Davies, 2004).

Strengths and limitations

This study does have limitations, including that the exactduration of participation in the before-school programwas not monitored and that data on before-school PAlevels on days without the program were not collected.However, fidelity to the two conditions was examinedthrough a daily before-school PA log. Additionally, 24-hrPA levels and potential compensation effects after schoolwere not examined, the sample size was small, and thestudy did not account for aerobic fitness and motivationto participate in the program. Given that the programonly included a running/walking club, it might not havecatered to all students’ interests.

The study’s strengths include its design, which combineda baseline phasewith anATphase, aswell as the fact that thestudy was replicated in a second setting. Combined, thesefeatures provide support both for internal and externalvalidity and the efficacy of the study. However, additionalreplicative studies are needed to support the generalizabilityof this study’s results further across different contexts.Finally, this cost-effective intervention in no way reducedacademic instruction time during the school day. Thisfinding represents an attractive dimension of the particularprogram, especially for school administrators and classroomteacherswhomay be reluctant to give up class time for PA inlight of the pressures of high-stakes testing.

Conclusion and recommendations for futureresearch

This efficacy study provides preliminary evidence that abefore-school running/walking program can significantlycontribute to children’s PA levels and can help themmeetthe daily PA guidelines. The contribution of the programto children’s PA is further highlighted by the fact thatthe participants did not compensate by decreasing theirschool-day PA on days they attended the program.Walking and running are lifetime PAs that generateimportant health benefits, and a running/walkingprogram is simple and cost-effective. Before-schoolprograms are also an attractive option for schools becausethey do not take time away from academic instruction.

Before-school programs may be the least studiedstudent-related component of CSPAPs. Thus, it isrecommended that additional studies focus on varioustypes of before-school programs (e.g., structured vs.unstructured) and examine resulting participation rates,PA levels, and sex and weight status differences. Futurestudies should also focus on other relevant outcomes,such as school attendance, classroom behavior (e.g.,Stylianou et al., 2016), and cognitive or academicperformance. Finally, there is a need for furtherinvestigation of the compensation question as a functionof individual differences and overall PA levels.

What does this article add?

This article focuses on the efficacy of a before-school PAprogram, which is perhaps the least studied student-relatedcomponent of CSPAPs, and therefore addresses a gap in theliterature examining the various CSPAP components andassociated outcomes. Specifically, this article focuses on abefore-school running/walking club, which is cost-effectiveand incorporates simple, inclusive lifetime activities. Whileseveral studies have investigated running/walking programsand associated outcomes, this study was one of the first tofocus on objectively measured PA levels accumulatedthrough a school-based running/walking program. Thisstudy provides preliminary evidence that short (15-minto 20-min) sessions of running/walking before school cansignificantly contribute to elementary-aged children’s PAlevels and can help them meet the daily PA guidelineswithout resulting in decreased school-day PA. From theperspective of policymakers, before-school PA programsmay be a more attractive option for schools than other PAprograms that occur during the school day because they donot take away fromacademic instruction time.Demonstrat-ing the positive student outcomes associated with before-school programs may encourage more schools to establishsuch programs. At a time when many elementary schoolchildren have limited PA opportunities during the schoolday, PA programming before and after school is invaluable.

References

Abbott, R. A., & Davies, P. S. W. (2004). Habitual physicalactivity and physical activity intensity: Their relation to bodycomposition in 5.0–10.5-y-old children. European Journal ofClinical Nutrition, 58, 285–291. doi:10.1038/sj.ejcn.1601780

Abbott, R. A., Straker, L. M., & Mathiassen, S. E. (2013).Patterning of children’s sedentary time at and away fromschool. Obesity, 21, E131–E133. doi:10.1002/oby.20127

Adams, M. A., Johnson, W. D., & Tudor-Locke, C. E. (2013).Steps/day translation of the moderate-to-vigorous physicalactivity guideline for children and adolescents. InternationalJournal of Behavioral Nutrition and Physical Activity, 10, 49,doi:10.1186/1479-5868-10-49

352 M. STYLIANOU ET AL.

Alderman, B. L., Benham-Deal, T., Beighle, A., Erwin, H. E., &Olson, R. L. (2012). Physical education’s contribution to dailyphysical activity among middle school youth. PediatricExercise Science, 24, 634–648.

Barlow, D. H., & Hersen, M. (1984). Alternating treatmentsdesign. In D. H. Barlow & M. Hensen (Eds.), Single caseexperimental designs: Strategies for studying behavior change(pp. 252–284). New York, NY: Pergamon.

Brusseau, T. A., & Hannon, J. C. (2013). Expected step countsof youth at school: A review. Sport Science Review, 22,329–342. doi:10.2478/ssr-2013-0016

Center on Education Policy. (2007). Choices, changes, andchallenges: Curriculum and instruction in the NCLB era.Washington, DC: Author.

Centers for Disease Control and Prevention. (2013). Comprehen-sive school physical activity programs: A guide for schools.Atlanta, GA: U.S. Department of Health and Human Services.

Cohen, J. (1988). Statistical power analysis for the behavioralsciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum.

Cooper, J. O., Heron, T. E., & Heward, W. L. (2007). Appliedbehavior analysis (2nd ed.). Upper Saddle River, NJ: Pearson.

Dale, D., Corbin, C. B., & Dale, K. S. (2000). Restrictingopportunities to be active during school time: Do childrencompensate by increasing physical activity levels after school?Research Quarterly for Exercise and Sport, 71, 240–248.doi:10.1080/02701367.2000.10608904

Erwin, H., Abel, M., Beighle, A., Noland, M. P., Worley, B., &Riggs, R. (2012). The contribution of recess to children’sschool-day physical activity. Journal of Physical Activity andHealth, 9, 442–448.

Foshay, J. D., & Patterson, M. (2010). The effects of runningclub membership on fourth graders’ achievement ofConnecticut state standard for the mile run. PhysicalEducator, 67, 197–203.

Hart, T. L., Brusseau, T., Kulinna, P. H., McClain, J. J., & Tudor-Locke, C. (2011). Evaluation of low-cost, objective instrumentsfor assessing physical activity in 10-11-year-old children.Research Quarterly for Exercise and Sport, 82, 600–609.

Institute of Medicine. (2013). Educating the student body:Taking physical activity and physical education to school.Washington, DC: National Academies Press.

Janssen, I., & LeBlanc, A.G. (2010). Systematic reviewof the healthbenefits of physical activity and fitness in school-aged childrenand youth. International Journal of Behavioral Nutrition andPhysical Activity, 7, 40. doi:10.1186/1479-5868-7-40

Lambiase, M. J., Barry, H. M., & Roemmich, J. N. (2010). Effectof a simulated active commute to school on cardiovascularstress reactivity. Medicine & Science in Sports & Exercise, 42,1609–1616. doi:10.1249/MSS.0b013e3181d0c77b

Long, M. W., Sobol, A. M., Cradock, A. L., Subramanian, S. V.,Blendon, R., & Gortmaker, S. L. (2013). School-day andoverall physical activity among youth. American Journal ofPreventive Medicine, 45, 150–157. doi:10.1016/j.amepre.2013.03.011

Mahar, M. T., Vuchenich, M. L., Golden, J., DuBose, K. D., &Raedeke, T. D. (2011). Effects of a before-school physicalactivity program on physical activity and on-task behavior[Abstract]. Medicine & Science in Sports & Exercise, 43(Suppl.1), 24.

McKenzie, T. L., Crespo, N. C., Baquero, B., & Elder, J. P.(2010). Leisure-time physical activity in elementary schools:

Analysis of contextual conditions. Journal of School Health,80, 470–477. doi:10.1111/j.1746-1561.2010.00530.x

McKenzie, T. L., Marshall, S. J., Sallis, J. F., & Conway, T. L.(2000). Leisure-time physical activity in school environ-ments: An observational study using SOPLAY. PreventiveMedicine, 30, 70–77. doi:10.1006/pmed.1999.0591

Morgan, C. F., Beighle, A., & Pangrazi, R. P. (2007). What arethe contributory and compensatory relationships betweenphysical education and physical activity in children? ResearchQuarterly for Exercise and Sport, 78, 407–412. doi:10.1080/02701367.2007.10599440

National Association for Sport and Physical Education. (2012).Comprehensive school physical activity program (CSPAP)policy continuum. Reston, VA: Author. Retrieved from http://www.shapeamerica.org/advocacy/upload/CSPAP-Policy-Continuum-2-10-12final.pdf

Ratliffe, T., & Bostick, A. (2001). Encouraging physical activitybeyond class time: Running and walking clubs. TeachingElementary Physical Education, 12, 24–26.

Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linearmodels: Applications and data analysis methods (2nd ed.).Thousand Oaks, CA: Sage.

Rowland, T. W. (1998). The biological basis of physical activity.Medicine & Science in Sports & Exercise, 30, 392–399. doi:10.1097/00005768-199803000-00009

Schnohr, P., Marott, J. L., Lange, P., & Jensen, G. B. (2013).Longevity in male and female joggers: The Copenhagen CityHeart Study.American Journal of Epidemiology, 177, 683–689.

Singer, J. D., & Willett, J. B. (2003). Applied longitudinal dataanalysis: Modeling change and event occurrence. New York,NY: Oxford University Press.

Stylianou, M., Kulinna, P. H., & Kloeppel, T. (2014). Healthyliving initiative: Running/walking club. The PhysicalEducator, 71, 157–184.

Stylianou, M., Kulinna, P. H., van der Mars, H., Mahar, M. T.,Adams, M. A., & Amazeen, E. (2016). Before-school running/walking club: Effects on student on-task behavior. PreventiveMedicine Reports, 3, 196–202. doi:10.1016/j.pmedr.2016.01.010

Troiano, R. P., Berrigan, D., Dodd, K. W., Mâsse, L. C., Tilert,T., & McDowell, M. (2008). Physical activity in the UnitedStates measured by accelerometer. Medicine & Science inSports & Exercise, 40, 181–188. doi:10.1249/mss.0b013e31815a51b3

Tudor-Locke, C., Lee, S. M., Morgan, C. F., Beighle, A., &Pangrazi, R. P. (2006). Children’s pedometer-determinedphysical activity during the segmented school day. Medicine& Science in Sports & Exercise, 38, 1732–1738. doi:10.1249/01.mss.0000230212.55119.98

Turner, L., Johnson, T. G., Slater, S. J., & Chaloupka, F. J.(2014). Physical activity practices in elementary schools andassociations with physical education staffing and training.Research Quarterly for Exercise and Sport, 85, 488–501.doi:10.1080/02701367.2014.961053

U.S. Department of Health and Human Services. (2008). 2008physical activity guidelines for Americans. Washington, DC:Author.

Xiang, P., McBride, R. E., & Bruene, A. (2006). Fourth-gradestudents’ motivational changes in an elementary physicaleducation running program. Research Quarterly for Exerciseand Sport, 77, 195–207. doi:10.1080/02701367.2006.1059935

BEFORE-SCHOOL PHYSICAL ACTIVITY PROGRAM 353