JOURNALOFAPPLIEDBEHAVIORANALYSIS 2011, 44, 635–639 NUMBER3 (FALL2011) BEHAVIORAL ASSESSMENT OF PHYSICAL ACTIVITY IN OBESE PRESCHOOL CHILDREN KRISTIN M. HUSTYI, MATTHEW P. NORMAND, AND TRACY A. LARSON UNIVERSITYOFTHEPACIFIC We measured changes in physical activity in 2 obese preschool children when a package intervention was evaluated in a reversal design. Physical activity was measured via direct observation and pedometers. Although the intervention produced only modest increases in activity,theresultsprovidepreliminaryconcurrentvalidationforthedependentmeasuresused, inthatthetwomeasurescovariedandasimilardegreeofchangewasobservedwitheachacross baseline andintervention phases. Keywords: goal setting, obesity,percentile schedules,physical activity, shaping _______________________________________________________________________________ The prevalence of overweight and obese boys relative influence of eating and activity on and girls between 2 and 5 years of age is weight loss. estimated to be around 8.5%, with less Because the validity and reliability of self- conservative estimates placing the prevalence report measures are suspect (especially with rate at 24% (Ogden, Carroll, & Flegal, 2008). children) and because physical activity is an Overweight and obesity contribute to a variety important health-related behavior, researchers of serious health problems (e.g., heart disease, have developed a number of measures to assess Type2diabetes),especiallyinadulthood.There physical activity. These include direct observa- also are significant financial consequences; tionsystemssuchastheBehaviorsofEatingand annual childhood obesity-related hospital costs Activity for Children’s Health Evaluation were estimated at $127 million during 1997, System (BEACHES; McKenzie et al., 1991) 1998, and 1999 (U.S. Centers for Disease and, more recently, the Observation System for Control and Prevention, n.d.). Recording Activity in Children (OSRAC; e.g., In general, behavioral interventions for McIver, Brown, Pfeiffer, Dowda, & Pate, overweight and obesity produce outcomes 2009). The five activity levels (ranging from superior to other approaches (Jelalian & sedentary to vigorous physical activity) pro- Saelens, 1999) and typically involve package posedbytheBEACHEShavebeenvalidatedvia interventions that target eating and physical concurrent heart-rate monitoring, but no stud- activity (Epstein, Myers, Raynor, & Saelens, ies have validated the OSRAC. 1998). However, much of the published Pedometers are mechanical devices worn at research relies on weight loss and self-reports the hip that can be used to assess physical of eating and activity as the primary dependent activity (i.e., the number of steps taken by the measures; thus, the accuracy of the data is wearer). Several studies have demonstrated the questionable, and little is revealed about the utility of pedometers as dependent measures and as tools for self-monitoring in behavioral WethankScottGreenberg,EllaSargent,andtheother interventions (Normand, 2008; VanWormer, members of the behavior analysis research group at the 2004). However, the participants in these UniversityofthePacificfortheirassistancethroughoutthe studieswereadults,andtheutilityofpedometer conductof thisstudy. Address correspondence to Matthew Normand, Uni- measures with young children has yet to be versity of the Pacific, Department of Psychology, 3601 determined. The purpose of the current study PacificAve,Stockton,California95211(e-mail:mnormand@ was to compare activity data from the OSRAC pacific.edu). doi:10.1901/jaba.2011.44-635 Preschool Version with those produced by a 635 636 KRISTIN M. HUSTYI et al. pedometer. This was done in the context of a collapsed because it was hypothesized that package intervention intended to increase relative metabolic and physiological measures physical activity in obese children to assess the did not significantly differ for some codes. utility and validity of the observation system. Activity Codes 2 (stationary with limb move- ments) and 3 (slow movements) were com- bined, as were Activity Codes 4 (moderate METHOD movements) and 5 (fast movements). The latter Participants and Setting combination was indicative of MVPA, and the Two typically developing 4-year-old chil- former combination was indicative of light- dren, Meg (female) and Chris (male), partici- intensity activity. McIver et al. (2009) com- pated. Their body mass index (BMI) placed bined categories in a similar way. Moreover, them in the 95th percentile of the U.S. Centers combining the activity codes in this way for Disease Control and Prevention BMI-for- improved interobserver agreement scores. agegrowthcharts.Allsessionstookplaceduring For 31% of all sessions, two observers a 20-min outdoor recess period on a school independently recorded both activity levels playground. Various activities were available, usingtheOSRACandthesteptotalsdisplayed including bicycles, sandbox, fixed equipment, on the pedometers for each participant. An and water tables. No structured activities were agreement was scored if both observers arranged, and participants were not prompted recorded the same activity code for a given to engage in any specific activity. interval or the same step total was recorded at the end of the session. To calculate interob- Response Measurement and Reliability server agreement, we divided number of Participants wore a New Lifestyles NL-2000 agreements by the total number of agreements pedometer, calibrated for height, weight, and plus disagreements and multiplied by 100%. age,throughouteachsession.Attheendofeach Mean agreement scores across three coding session, an observer recorded the number of categories (sedentary, light, and MVPA) were steps taken by each participant, with a step 82% for Chris (range, 74% to 90%) and 80% defined as any activity that resulted in a count for Meg (range, 58% to 95%). Agreement was being recorded on the pedometer. Using the 100% for step totals. OSRAC,observersalsorecordedactivityusinga momentary time-sampling system with 5 s to Procedure observe and 25 s to record per interval. Five An intervention package comprised of goal activity codes indicated the level of physical setting, performance feedback, and reinforce- activity (see Table 1), and the highest level of ment was evaluated using a reversal design. physicalactivityobservedduringanintervalwas Throughout the study, goals were set related to recorded.Allothercategories (i.e.,activitytype, step totals. The degree of covariation between initiator, group composition, prompts, engage- the pedometer step totals and OSRAC activity ment, and context) also were coded with codeswasassessedwithinandacrossconditions. reference to the activity level code; however, Baseline. Each participant wore a pedometer only the activity-level data are reported. butdidnothaveanyperformancegoalsanddid For the purposes of analysis, the OSRAC not receive programmed consequences for activity codes were collapsed into three catego- activity. The pedometer screens were covered, ries: sedentary (i.e., stationary), light (i.e., and observers ensured that participants did not stationary with limb movements and slow uncover the screens or tamper with the movements), and moderate-to-vigorous physi- pedometers during sessions. At the end of the cal activity (MVPA). The activity codes were session, observers collected the pedometers and ASSESSMENT OF PHYSICAL ACTIVITY 637 Table 1 Activity LevelCodes Specified byMcIveretal.(2009) Level Activity Operationaldefinition 1 Stationaryormotionless Stationaryormotionlesswithnomajorlimbmovementsormajorjointmovements(e.g., sleeping,standing,ridingpassivelyinawagon). 2 Stationarywithlimbor Stationarywitheasymovementsoflimbsortrunkwithouttranslocation(e.g.,standingup, trunkmovements holdingamoderatelyheavyobject,hangingonbars). 3 Slow,easymovements Translocationataslowandeasypace(e.g.,walkingwithtranslocationofbothfeet,slowand easycycling,swingingwithoutassistanceandwithoutlegkicks). 4 Moderatemovements Translocationatamoderatepace(e.g.,walkinguphill,tworepetitionsofskippingor jumping,climbingonmonkeybars,hangingfrombarwithlegsswinging). 5 Fastmovements Translocationatafastorveryfastpace(e.g.,running). recorded the total steps displayed for each meet the goal received no prize and were participant. encouraged to reach their goal during the Goal setting and feedback. Rather than subsequent session. arbitrarily setting goals (e.g., at 10% above the baseline average; Donaldson & Normand, RESULTS AND DISCUSSION 2009), we used the percentile schedule of reinforcement equation, k 5 (m + 1)(1 2 w), Figure 1 depicts step totals and mean to determine step-total goals for each partici- activity level per session for each participant, pant (Galbicka, 1994). In the equation, w (the with the horizontal bars representing the step- densityof reinforcement) was set to0.5 andthe total goals. The number of steps taken per m value (number of responses included in the session covaried with the mean OSRAC calculation)wassetat5tomatchthenumberof activity level scored per session; both measures sessions during baseline. During the interven- indicated a similar degree of change across tion phase, the five most recent step totals were baseline and intervention phases. For Chris, arranged in ascending order. The k value (the mean step totals during baseline, intervention, thirdvalueintheorder)wasusedasthegoalfor return to baseline, and the final intervention the next session. phase were 1,478, 2,082, 1,649, and 1,809, Before each session, the primary observer respectively, and he met his goal for 78% of explainedtotheparticipantsthatthepedometer sessions. Mean percentages of intervals engaged would measure each step they took. The in MVPA per phase were 19%, 61%, 48%, observer then stated the performance goal, and 55%. For Meg, mean step totals during wrote it on a sticker, and placed it on the each phase were 769, 965, 706,and1,125, and pedometer.When10minhadelapsed fromthe she met her goal for 67% of sessions. Mean start of the session, the observer provided percentage of intervals engaged in MVPA per feedback by opening the pedometer, reading phase was 13%, 28%, 12%, and 31%. This the step total, and describing how close the correspondence suggests that either method of participant was to meeting the goal. At the end assessment can be used to evaluate interven- of the 20-min session, the experimenter pre- tions designed to increase physical activity in sented the participants who met or exceeded children. their goal with a prize box and allowed them to However, more research is necessary to choose one item identified by the teacher as determine the conditions under which step highly preferred (e.g., stickers, stamps, pencils, totals and observed activity levels would differ pens, buttons, etc.). Participants who did not enough to warrant the use of one measure over 638 KRISTIN M. HUSTYI et al. Figure1. TotalstepsandmeanactivitylevelpersessionforChrisandMeg.Horizontallinesindicatestep-totalgoals. another. For example, because of differences in displacement and do not distinguish among stridelength,runningagivendistancecanresult activities that require more or less physical in fewer steps being taken than when walking exertion. Because pedometers are likely to the same distance. Direct observation would underestimate activity, they might be preferable then indicate a higher level of activity than to direct observation in some circumstances would a pedometer. Further, pedometers can- because they produce more conservative esti- not measure activities that do not involve hip mates of behavior and behavior change. ASSESSMENT OF PHYSICAL ACTIVITY 639 The package intervention produced modest REFERENCES increasesinactivityforChriswhencomparedto Donaldson,J.M.,&Normand,M.P.(2009).Usinggoal baseline activity levels. However, there was setting, self-monitoring, and feedback to increase substantial overlap in activity levels across calorie expenditure in obese adults. Behavioral Inter- ventions,24,73–83. baseline and intervention phases for Meg; thus, Epstein, L.H., Myers,M.D.,Raynor,H.A.,&Saelens, the degree to which activity levels changed is B.E.(1998).Treatmentofpediatricobesity.Pediatrics, questionable for this participant. The interven- 101,554–570. tion might have failed to produce substantial Galbicka,G.(1994).Shapinginthe21stcentury:Moving percentile schedules into applied settings. Journal of increases in activity because the items available Applied BehaviorAnalysis,27, 739–760. intheprizeboxwerenotreinforcers.Noformal Jelalian, E., & Saelens, B. E. (1999). Empirically preference assessment was conducted; instead, supported treatments in pediatric psychology: Pedi- all of the teacher-nominated stimuli were atricobesity.JournalofPediatricPsychology,24,223– 248. included in the prize box. Future research McIver,K.,Brown,W.H.,Pfeiffer,K.A.,Dowda,M.,& should employ systematic preference assess- Pate, R. R. (2009). Assessing children’s physical ments, conducted frequently throughout the activity in their homes: The observational system for recording physical activity in children—home. Jour- intervention, to increase the likelihood that nal ofAppliedBehavior Analysis, 42, 1–16. effective reinforcers are provided. Also, time McKenzie, T. L., Sallis, J. F., Nader, P. R., Patterson, constraintsprecludedextendedphases;lengthier T. L., Elder, J. P., Berry, C. C., et al. (1991). intervention phases might produce more robust BEACHES: An observational system for assessing children’s eating and physical activity behaviors and changes in physical activity. associatedevents.JournalofAppliedBehaviorAnalysis, The specific components of the goal-setting 24,141–151. procedures also deserve further evaluation. Step Normand, M. P. (2008). Increasing physical activity totals did not closely track step-total goals, and through self-monitoring, goal setting, and feedback. Behavioral Interventions, 23,227–236. participants rarely requested additional feedback Ogden, C. L., Carroll, M. D., & Flegal, K. M. (2008). during a session or attempted to view the HighbodymassindexforageamongUSchildrenand pedometers. These results suggest that general adolescents, 2003–2006. Journal of the American Medical Association,299, 2401–2405. promptsthatdonotrequirecalculationmightbe Tudor-Locke,C.E.,McClain,J.J.,Hart,T.L.,Sisson,S. as effective as specific (percentile) goals with B.,&Washington,T.L.(2009).Expectedvaluesfor preschoolers and that performance feedback pedometer-determined physical activity in youth. might be unnecessary or ineffective. The results ResearchQuarterlyforExerciseandSport,80,164–174. U.S. Centers for Disease Control and Prevention. (n.d.). alsosuggestthatinterventionsthatteachchildren Preventing obesity and chronic diseases through good to self-monitor using a pedometer might be nutrition and physical activity. Retrieved from http:// worthy of consideration. Finally, a number of www.cdc.gov/chronicdisease/resources/publications/ mechanical devices are available to measure fact_sheets/obesity.htm VanWormer, J. J. (2004). Pedometers and brief e- physical activity or indicators thereof, such as counseling:Increasingphysicalactivityforoverweight accelerometers and heart-rate monitors. The adults. Journal of Applied Behavior Analysis, 37, utility of such devices, which are typically more 421–425. costlythanpedometersbutproducemoredetailed Received May 3,2010 information about the intensity of physical Final acceptanceAugust 20,2010 activity, warrant consideration with preschoolers. Action Editor,Rachel Thompson