- Email: [email protected]
After-School Physical Activity and Eating Behaviors of Middle School Students in Relation to Adult Supervision
Research Article After-School Physical Activity and Eating Behaviors of Middle School Students in Relation to Adult Supervision Wayne C. Miller, PhD1,y; Michelle Hering, MS2; Carrie Cothran, PA, MS2; Kim Croteau, RD, MS2; Rebecca Dunlap, MS2 ABSTRACT Objective: Examine after-school activity patterns, eating behaviors, and social environment of overweight and normal weight middle school students. Design: Eating and physical activity behaviors of 141 students, ages 10–14, were monitored. Students completed a diary documenting type of activity, location, adult supervision, accompanying participants, and eating habits from 3:00 PM–12:00 AM. Setting: Three middle schools, grades 6–8. Main Outcome Measures: Body mass index, estimated energy expenditure, eating behavior, active time, sedentary time, supervised time. Analysis: t tests, ANOVA, chi-square, correlation coefficients. Significance set at P < .05. Results: Children spent 76% of time sedentary, and 85% of sedentary time was under adult supervision (r ¼ 0.76). Active time related to time with friends (r ¼ 0.64) and family (r ¼ 0.46). Children spent 40% of eating time consuming unhealthful food, and adults supervised 86% of children’s eating. Overweight and normal weight children were similarly active (335 156 vs 373 194 counts per minute). Overweight girls spent more eating time (77%) eating healthfully than overweight boys (57%). Conclusions and Implications: Children should be given access to healthful food and encouraged to eat healthfully when alone and with friends. Adults should be more physically engaged with children. Children should be encouraged to eat under adult supervision and with their families. Key Words: physical activity, accelerometer, eating behavior, childhood obesity, adult supervision (J Nutr Educ Behav. 2012;44:326-334.)
INTRODUCTION It is well recognized that the prevalence of obesity in children has risen substantially over the past 3 decades. Although this prevalence trend has flattened over the past several years, childhood obesity remains a major problem in the United States.1 The high prevalence of childhood obesity experienced today is a result of complex social, environmental, and policy changes that have influenced eating and physical activity behaviors. For example, many urban and
1
suburban communities are designed in such a way that walking, bicycling, and other physical activities are discouraged. There are fewer opportunities today than a generation ago for children to be active during school, after school, and going to and from school.2,3 The data indicate that 62% of children aged 9–13 years do not participate in any organized physical activity during their nonschool hours and that 23% do not engage in any free-time physical activity.4 Children spend less of their free time engaged in physical activity outdoors
Center for Rural and Community Health, West Virginia School of Osteopathic Medicine, Lewisburg, WV 2 Department of Exercise Science, The George Washington University Medical Center, Washington, DC † Dr. Wayne C. Miller was affiliated with the Department of Exercise Science, The George Washington University Medical Center at the time this study was completed. Address for correspondence: Wayne C. Miller, PhD, Center for Rural and Community Health, West Virginia School of Osteopathic Medicine, 400 North Lee St, Lewisburg, WV 24901; Phone: (304) 647-6380; Fax: (304) 793-6586; E-mail: [email protected] .edu Ó2012 SOCIETY FOR NUTRITION EDUCATION AND BEHAVIOR doi:10.1016/j.jneb.2011.08.002
326
and more free time being sedentary, either watching TV or playing video games, than in years past.5 Evolutions in family life have caused changes in children’s eating patterns. The family can have both negative and positive effects on children’s eating behaviors.6-12 Crosssectional studies have shown that children and adolescents who regularly eat dinner with family are less likely to be overweight and more likely to have more healthful eating habits.6-8,11 However, modern families experience pressure to minimize food costs and to shorten food acquisition and preparation time, which has led many families to regularly consume unhealthful processed food that is more conveniently acquired and prepared for consumption than home-prepared, fresh food.13,14 Studies have shown that busy households, in which families dine out for meals, subject children to food that is high in energy density, as well as high in sodium.9 Over the past 2 decades, the number of working parents has increased
Journal of Nutrition Education and Behavior Volume 44, Number 4, 2012
Journal of Nutrition Education and Behavior Volume 44, Number 4, 2012 dramatically; 75% of mothers with school-aged children are now employed.15 Both parents also work an increasing number of hours during the week,15 which means that the number of children forced to care for themselves has also increased dramatically. In 2006, it was reported that 18% of children ages 9–11 years were taking care of themselves.16 The size of the family has an impact on health behaviors as well. Children in small families have more healthful eating patterns than children in large families.8 Presumably, the larger the family, the less likely a child’s eating habits will be directly supervised by an adult.8 Other researchers found that lack of adult supervision caused unsupervised children to participate in less physical activity and consume larger amounts of unhealthful food than supervised children.15 Adult supervision, therefore, seems to be a key in promoting healthful eating as well as physical activity patterns in children. It appears that there are 2 major factors that have contributed to the increase in childhood obesity over the past 3 decades: a community environment that discourages physical activity and healthful eating, and a family environment in which adults are less able to supervise childhood behaviors. Therefore, to influence/ impact rates of childhood obesity, one must examine the children’s environment, with particular focus on how children spend their time. Better understanding of how children spend their time and under what circumstances they participate in healthful or unhealthful behaviors will help in the design of more effective childhood obesity interventions. Therefore, the purpose of this study was to examine the physical activity patterns, eating behaviors, and social environment of overweight (OW) and normal weight (NW) middle school children between the hours of 3:00 PM and 12:00 AM, when children are most likely to engage in behaviors unsupervised by adults. The working hypothesis for this study was that during these after-school hours, OW children are less active, have poorer eating behaviors, and have a weaker social environment for health behaviors than NW children.
METHODS Study Design After-school physical activity patterns, eating behaviors, and social environment were examined simultaneously in middle school students. Physical activity was measured objectively through accelerometry and recorded subjectively by each student with an activity diary. Students also recorded their eating behaviors and social environment in the diary. Weight comparisons (OW vs NW) and sex comparisons were made to identify possible determinants of obesity in middle school students.
Participants This study was approved by the Institutional Review Board for Human Research at the George Washington University Medical Center. Three middle schools (grades 6–8) were selected for this study, because it is well known that physical activity declines among youth as they grow through adolescence.4,17,18 One middle school (DC) was selected from the inner city of Washington DC, a second school (VA) was selected from a mid-sized town located in northern Virginia, and the third school (GA) was selected from a rural agricultural community in Georgia. These 3 sites were selected to represent an urban metropolis, a ‘‘blue collar’’ town, and a farming community. The per capita income for the DC site was unknown, for the VA site $23,552, and for the GA site $13,657. Although direct generalizeability to the entire American population may not be possible with this sampling technique, these 3 sample cities represent 3 distinct cross-sections of the American population. Cooperation was obtained from the school board, principal, and physical education teachers of the respective schools. Students were recruited for this study via class announcements. Researchers gave a 15-minute presentation about the nature of the study to students in their physical education classes. Interested students were then given a parent authorization form and a child assent form to be completed and signed prior to participation in the study.
Miller et al 327
Protocol The DC school had 57 out of 180 potential children participating; the VA school had 48 out of 200 potential children, and the GA school 36 of 204 students participating. All of the DC students were African American; of the VA students, 90% were Caucasian and 10% were African American; and of the GA students, 39% were African American, 50% Caucasian, and 11% Hispanic. None of the students was practicing weight control methods. Students were not given any monetary incentive to participate. Researchers visited the school on the Monday of a full week of school to gather anthropometric data, equip the children with accelerometers, and provide instruction on completion of the activity logs. Monday morning, the participants were taken into a private room away from the rest of their physical education class, where their weight and height were obtained using a scale (Taylor Precision Tech #7506; Taylor Precision Products, Oak Brook, IL, 2007) and stadiometer (Secca Portable Stadiometer #213; Secca, Hanover, MD, 2007) that were accurate to 0.25 lb (0.114 kg) and 0.25 in (0.64 cm), respectively. Children were classified as NW if their body mass index (BMI) for age and sex was $ 5% and < 85% and as OW if their BMI for age and sex was $ 85%. Demographic data (height, weight, age, and sex) were input into a laptop computer so the participant could be fitted for an accelerometer. Participants were then fitted with an Actical accelerometer (Respironics Inc, Bend, OR, 2007) attached to a waistband with a velcro strap. The accelerometer was secured around the hips with the actual monitor lying just above the right iliac crest. Participants were shown how to remove the accelerometer at night and replace it in the morning. Children were told that the monitors were waterproof and thus could be worn while showering or swimming. Children were asked to keep the accelerometer on at all times from 3:00 PM until bedtime on the days of the study. The Actical accelerometer is sensitive to movement in all directions, and the storage capacity allows for
328 Miller et al several days of continuous monitoring.19,20 The Actical accelerometer is valid for detection of low-energy movements as well as high-energy movements.19,20 Along with an accelerometer, each participant was given a 4-day activity diary to complete during the experimental week, from that Monday, beginning at 3:00 PM, until that Thursday at 12:00 AM. The activity diary was based on the Bouchard 3-day physical activity diary, which has been found to be both valid and reliable for estimating daily energy expenditure for ages 10–50 years.21,22 The diary was modified so that 20-minute time intervals were used, instead of 15-minute intervals, for the participant to record the activities in which he or she took part. Our preliminary work with children this age suggested that expanding the time frame from 15–20 minutes would help with compliance in completing the diary (WCM, unpublished data, 2010). The 20-minute increments allowed for assignment of a categorical value for the dominant activity in any given time frame. A further modification to the diary was that, in addition to recording the activity in which the child participated, the child also recorded his or her location, who was with him or her, and whether there was adult supervision. Also included on the diary was an example of how to fill it out properly, with accompanying instructions. Researchers returned to the school on Friday to retrieve both the activity diaries and accelerometers.
Compression and Coding of Data Researchers coded the responses on the returned activity diaries so that location was categorized as Indoors at Home, Indoors Out of Home, and Outdoors. The ‘‘with whom’’ response was coded as either with a family member, with a friend, or alone. Adult supervision was defined as a person 18 years old or older who was in the act of overseeing the adolescent. The supervising adult must have been within voice range of the child to be considered supervising. Distinct verbal and written instructions were given to the children on how to record where they were, whom they were with, and whether an adult was supervising. The researcher who
Journal of Nutrition Education and Behavior Volume 44, Number 4, 2012 coded the diary responses was given the same set of instructions for reporting as the children. The researcher then reviewed each returned diary for consistency between the instructions and the way the children completed the diary. If a discrepancy was found, the researcher either corrected the discrepancy (according to the written coding instructions) or eliminated that particular response from the database. In addition, the participant’s activities were coded as either active or sedentary; sedentary was defined as anything done sitting or lying down (including napping). Activities were also coded as eating (yes or no) and what type of food was being consumed (healthful food or unhealthful food). Eating fast food was classified separately because of the relationship between fast-food consumption and obesity.23 Fruit, 100% fruit juice, vegetables, grains, lean meats, fish, and low-fat dairy (1% fat or lower) were classified as healthful food. High-fat snacks (eg, regular fat cheese, peanut butter, chips), sweets, sugared drinks, diet soda, high-sodium snacks (eg, chips, crackers, pretzels), and high-calorie/ low-nutrient food (eg, pizza, macaroni and cheese) were classified as unhealthful food. All of the codes were predetermined and standardized by 1 investigator. The activity diary was completed only from 3:00 PM to 12:00 AM for 4 days, yielding a total of 27 consecutive 20-minute time intervals per day, and 108 total time intervals per participant. Since the activity diary was divided into 20-minute time increments, data from the accelerometer were also compiled into 20-minute time intervals to give the total activity counts for each corresponding activity diary time interval. If there were 3 consecutive time intervals (60 consecutive minutes) of accelerometer data for where activity counts were measured to be zero, information obtained from the activity diary was used to determine whether the participant was sedentary or just not wearing the monitor. If the activity diary reported a period of being awake when there were no accelerometer counts, it was assumed that the participant was not wearing the accelerometer, and the corresponding time
interval for accelerometer data was voided. Data collection stopped for the day when the accelerometer registered zero counts and the child recorded going to bed for the night on the activity diary. Therefore, the dataset includes only activity data for the time a child was awake between 3:00 PM and 12:00 AM. Empty time cells were not included in data analysis. A reliability coefficient of 0.80 was previously recorded for students in grades 1–6 and of 0.70 for grades 7–12 when they wore the accelerometer for 4 or 5 days.24 This study examined after-school weekday activities, so a 4-day range was deemed appropriate. The need to match children’s activity, adult supervision, and eating behaviors to accelerometer counts led to extending the Bouchard 3-day activity diary to 4 days to accommodate the accuracy and reliability of the Actical accelerometer.
Statistical Analyses All statistical procedures were performed with the Systat statistical software package (version 12, SYSTAT Software, Inc., Chicago, IL, 2007). Values are reported as mean SD. Statistical significance was set at P < .05. Sample data were tested for normal distribution using the KolmogorovSmirnov test and for equal variance using the Levene median test. The 2sample t test was used to make mean group comparisons between the NW and OW and between boys and girls. ANOVA with repeated measures was used to make mean group comparisons for accelerometer counts across time periods. ANOVA was used to make group mean comparisons among schools and among races for the demographic variables and accelerometer counts, and to make comparisons among ages for accelerometer counts. A Tukey pairwise comparison was used if an ANOVA produced a significant F value. The chi-square test was used to compare differences in the frequency counts between groups for the categorical data reported in the activity diaries (ie, time spent in each of 19 categories of activity). A test for equality of 2 proportions was used to test for differences in prevalence ratios or percentages between 2 groups. Simple correlation coefficients were used to
All indicates all participants; DC, inner city of Washington, DC; GA, rural agricultural community in Georgia; NW, normal weight; OW, overweight; VA, mid-sized town in northern Virginia. *Significantly different from NW group (P < .001); **Significantly different from other schools (P < .05); ***Significantly different from other races (P < .05). Note: ANOVA was used to test for significant group differences among schools and races. Independent t tests were used to test for significant group differences between OW and NW and between sexes.
Girls (n ¼ 102) 12.4 0.8 54.6 14.2 158.4 9.3 21.6 4.8 Boys (n ¼ 39) 12.4 0.8 54.5 17.1 159.5 10.5 21.0 4.8 OW (n ¼ 36) 12.4 0.7 72.4 15.7* 160.8 12.5 27.9 4.5* NW (n ¼ 105) 12.4 0.9 48.4 8.6 158.0 8.4 19.2 2.0 All Group (n ¼ 141) Age (y) 12.4 0.8 Weight (kg) 54.6 15.0 Height (cm) 158.7 9.6 BMI 21.5 4.8
Table 1. Anthropometric and Ethnic Data for Middle School Students, mean (SD)
DC School (n ¼ 57) 12.5 0.8 59.9 17.1** 161.0 10.7** 23.0 5.8**
VA School (n ¼ 48) 12.1 0.7** 50.1 12.0 156.0 8.8 20.4 3.5
GA School (n ¼ 36) 12.8 0.8 52.0 12.6 158.7 8.2 20.4 3.3
African American (n ¼ 76) 12.6 0.9*** 58.3 16.9 161.0 10.4*** 22.4 5.5***
Caucasian (n ¼ 61) 12.1 0.7 50.4 11.3 156.7 8.3 20.3 3.3
Hispanic (n ¼ 4) 12.3 0.5 46.4 5.5 152.4 4.6 19.9 1.6
Journal of Nutrition Education and Behavior Volume 44, Number 4, 2012
Miller et al 329 measure the strength of association between 2 quantitative variables.
RESULTS The demographic information for the participants is presented in Table 1. There were no age differences between the NW and OW groups or between the sexes. The VA children were 5–7 months younger than children at the other 2 sites, and the African-American children were 2–5 months older than the Caucasians and Hispanic children. The children at the DC school were significantly taller and heavier and had a higher BMI than the children at the other 2 sites. The total physical activity measures, expressed as average accelerometer counts per minute for the entire time frame (3:00 PM to 12:00 AM) are shown in Figure 1. There were no significant group differences in average accelerometer counts from 3:00 PM to 12:00 AM when comparing the NW to OW, boys to girls, schools, races, or age groups. Energy expenditure algorithms for children using the Actical accelerometer were derived by Freedson et al.25 When these algorithms were applied to the data, there were no significant differences in estimated energy expenditure for any of the comparable groups. In fact, the range in energy expenditure for all groups was very narrow, from 0.017 0.002 to 0.019 0.002 kcal per kg body weight per minute (group data not shown). This result amounted to an average of 530 344 kcal expended from 3:00 PM to 12:00 AM, or approximately 1.0 kcal per minute during that time frame. The correlation coefficient for the relationship between the accelerometer counts within each of the 27 time intervals across the 4 days and the frequency of positive responses (in the activity diaries) for being active over the same time was 0.82 (P < .001). The plot of accelerometer counts for all children over time is shown in Figure 2. Physical activity was highest between 3:00 PM and 6:40 PM. After 6:40 PM, physical activity began to decline to where at 7:00 PM, activity was significantly lower than at its peak time (3:00 PM to 5:00 PM). Physical activity continued to decline until 11:00 PM.
330 Miller et al
Journal of Nutrition Education and Behavior Volume 44, Number 4, 2012
Figure 1. Group comparisons for average accelerometer counts per minute from 3:00 PM to 12:00 AM in middle school students. AA indicates African American; All, all participants; Cau, Caucasian; DC, District of Columbia students; GA, Georgia students; His, Hispanic; NW, normal weight; OW, overweight; VA, Virginia students. Note: 11, 12, 13, and 14 represent the respective student ages. Values represent the mean SD for the respective groups. ANOVA was used to test for group mean differences. Group numbers are reported in Table 1. Table 2 shows the percentage distribution of time between 3:00 PM and 12:00 AM spent in different activi-
ties and under different environmental situations. Middle school children spent over three fourths of their after-
school hours being sedentary, and approximately 80% of that sedentary time was spent under adult supervision. Eighty-five percent of children’s eating time was supervised by an adult, and 19%–26% of that adult-supervised eating was spent eating unhealthful food. The OW children spent more time being active, but less time being sedentary and less time under supervision than the NW children. The OW children also spent less time eating unhealthful food than the NW children. The OW children spent less of their after-school time being supervised and less sedentary time being supervised than NW children. Overweight girls spent more time being active, but also more time eating than OW boys. Overweight girls spent more of their eating time consuming healthful food than OW boys. Simple correlations revealed that the amount of time spent being active was positively related to time spent with friends for both the NW and OW children (Table 3). Active time was also related to time spent with family, but the relationship was stronger for the NW compared to the OW children. In contrast, time spent being sedentary was positively related to time spent being supervised for both the NW and the OW children. Sedentary time was more strongly related to time spent alone for the NW children compared to the OW children. Time spent eating healthful food was positively related to time spent with family, whereas time spent eating unhealthful food was related to time spent with friends for the NW and OW children. Time spent eating unhealthful food was inversely related to time spent under adult supervision as well as inversely related to time spent alone for both the NW and OW children; however, the strength of these relationships differed between groups.
DISCUSSION After-School Physical Activity Levels
Figure 2. Time plot of accelerometer counts from 3:00 PM to 12:00 AM in middle school students. Note: Values represent the mean SD for the entire student sample. ANOVA with repeated measures was used to test for significant within-group differences across time.
The amount of physical activity achieved (expressed as either activity counts or kcal) throughout the afterschool hours was similar between the NW and OW groups. Physical activity was highest from 3:00 PM to 6:40 PM,
Journal of Nutrition Education and Behavior Volume 44, Number 4, 2012 Table 2. Percentage Use of After-School Time for Middle School Students Category (%) Time spent eating Time spent with a family member Time spent with friends Time spent alone Time spent sedentary Time spent being active Time spent supervised Time spent outside Outside time being active Eating time eating healthful food Eating time eating unhealthful food Eating time supervised Supervised eating time eating unhealthful food Eating time with a family member Family eating time eating unhealthful food Eating time alone Alone eating time eating unhealthful food Sedentary time eating Sedentary time supervised
All 9.0 36.8 18.9 44.3 76.0 24.0 83.7 15.1 60.6 59.1 40.9 85.2 21.1
NW 9.6 39.2 19.7 41.1 77.3 22.7 85.8 15.5 61.4 55.0 45.0 85.7 20.7
OW 7.3 29.2* 16.5 54.3* 70.1* 29.9* 77.5* 14.6 58.6 71.9* 28.1* 83.4 22.0
OW Boys 5.4 30.6 12.6 56.8 78.4 21.6 78.9 15.6 65.3 56.9 43.1 78.5 26.2
OW Girls 8.3** 28.5 18.5** 53.0** 68.6** 31.4** 76.8 14.1 56.2** 76.7** 23.3** 80.4 19.4**
70.2 25.6
70.6 25.7
69.1 25.1
64.6 23.8
70.4 25.6
18.9 60.8
18.7 59.9
19.5 61.2
30.8 61.0
15.8** 61.4
10.9 86.6
11.5 87.7
9.9* 83.3*
6.8 82.3
10.3** 83.8
All indicates all participants; NW, normal weight; OW, overweight. *Significantly different from NW group (P < .05); **Significantly different from OW boys (P < .05). Note: A test for quality of 2 proportions was used to test for significant differences in prevalence ratios or percentages between 2 comparable groups.
and it then declined throughout the evening and night (Figure 2). This finding is not surprising in that the common routine for middle school children in the United States is free time immediately after school, dinner time, and home and family time, followed by quiet time and bed time.
Several other studies concur that the most active time in a child’s day is immediately after school.19,26-31 However, none of these studies was as complete and detailed as the present study. None of these studies sampled children from 3 distinct demographic areas (eg, urban, rural
Table 3. Relationships between Social Interaction Time and Eating or Activity Time for Middle School Students Category Active time vs time with friends Active time vs time with family Sedentary time vs time supervised Sedentary time vs time alone Time spent eating healthfully vs time with family Time spent eating unhealthfully vs time with friends Time spent eating unhealthfully vs time supervised Time spent eating unhealthfully vs time alone
All 0.64 0.46 0.76 0.88 0.87 0.89 0.82 0.66
NW 0.65 0.49 0.76 0.90 0.88 0.90 0.87 0.68
OW 0.62 0.36* 0.76 0.80* 0.81 0.87 0.78* 0.59*
All indicates all participants; NW, normal weight; OW, overweight. *Significantly different from NW group (P < .05). Note: A test for quality of 2 proportions was used to test for significant differences in prevalence ratios or percentages between the NW and OW groups. Values are correlation coefficients (P < .01).
Miller et al 331 blue collar, rural agricultural). Each of these studies sampled children from only 1 geographical location. None of these studies performed an hourby-hour analysis of physical activity levels. The few studies that evaluated physical activity with reference to time grouped the data into blocks of time.27,29-34 The time blocks analyzed were broad, representing either total time out of school24 or the block of time immediately after school until 6:00 PM and then again from 6:00 PM until 9:00 or 10:00 PM.28,29,31 Separating time frames at 6:00 PM may not have captured the best representation of physical activity patterns, because physical activity in these children decreased very rapidly from 5:00–7:00 PM. More interesting than the time of day when activities happened might be the amount of time children devoted to different types of active or sedentary behaviors. These middle school children spent over three fourths of their after-school time being sedentary (Table 2), which translates into spending 6.8 hours between 3:00 PM and 12:00 AM either sitting or lying down. Furthermore, most of the children’s sedentary time was spent under supervision by an adult (Table 2). Other researchers have suggested that the home environment created by parents likely influences the choices made by adolescents about physical activity.32 According to behavioral economic theory, a child’s choice to be sedentary or active depends on the ease of access to sedentary activities in the home.32 For instance, having many televisions in the home, or easy access to video games and computers leads to less time in an active environment. On the other hand, having dogs, for example, promotes greater amounts of physical activity.32 Other research supports this theory in that more than two thirds of fifth- and sixthgraders surveyed said they would chose to be more physically active if given the choice to do so.35 At the same time, these children reported that spending time with family was a major reason they engaged in screen-related activities. On a positive note, when children were outside, they spent about 60% of their time being active. This finding
332 Miller et al suggests that 1 area of focus in combating childhood obesity is to have children spend more time outside. Social support from parents and other family members has also been identified as 1 of the most important determinants of participation in all forms of physical activity, including sports, structured exercise, and active leisure inside and outside the home.33 Parents are responsible for providing different kinds of support in the form of transportation, equipment, paying activity fees, encouraging their children to be active, and playing with their children, all of which promote physical activity.33
Physical Activity Patterns in NW vs OW It may be surprising that the OW children spent more time being active than the NW children. These selfreport data could infer a potential reporting bias in which OW children overestimated their active behaviors. However, the accelerometer counts for the OW and NW children were similar. It may be that when the NW children were physically active, their intensity of activity was greater than that of the OW children. Under this scenario, OW children would have needed to accumulate more active time at their lower activity intensity in order to accrue the same number of accelerometer counts as the NW. Other investigators have shown that NW children do spend more time in moderate to vigorous activities than OW children.26,34,36,37 Three of these studies were consistent in finding that NW children were more active than OW children overall, and that they accrued more minutes of moderate to vigorous activity throughout the day than OW children.26,34,36 However, Purslow et al did not find any relationship between weight status and total daily activity or moderate to vigorous activity for girls.37 Furthermore, these same investigators did not find any relationship between sedentary activity and weight status for either boys or girls. None of the previous studies simultaneously used subjective and objective measures of physical activity.26,34,36,37 The authors monitored physical activity simultaneously with objective measures (accelerometers) and subjective measures (activity diaries).
Journal of Nutrition Education and Behavior Volume 44, Number 4, 2012 The correlation coefficient between the average accelerometer counts per minute and the average frequency of positive responses for being active on the diaries during each of the 27 time frames was high (r ¼ 0.82) for the OW children. This finding suggests that the OW children were accurately recording their activity time. The OW girls reported spending more time being active than the OW boys, when their estimated energy expenditure was the same. However, it may well have been that the activities reported by the OW girls were associated with lower-intensity activities such as chores and activities of daily living, whereas those of the OW boys were associated with short bouts of vigorous activity, such as organized games or sports.38 Preliminary data support this theory, in that girls scored lower than boys on the leisure index of the Fels Physical Activity Questionnaire when their total scores for the questionnaire were equivalent (WCM et al, unpublished data, 2010).39 Overweight girls spent less of their outside time being active than OW boys (Table 2), which may have occurred because the outdoor environment is more favorable to boys’ activities than girls’ activities.32 Roemmich et al discovered that the built environment affects boys and girls differently.32 These researchers reported a positive relationship between the percentage of recreational park area in a neighborhood and physical activity for boys, but no such relationship for girls.
Dietary Patterns and Behaviors The great majority of time children spent eating was under direct adult supervision and/or with a family member (Table 2). Approximately 60% of children’s total eating time was spent eating healthful food (Table 2). When children ate alone, they spent 60% of that time eating unhealthful food. These data support the idea that parents feel it is important to feed their children well so they will grow to be big and strong.13 This may be the case with the OW children, in that they spent more of their eating time consuming healthful food than the NW children (Table 2). On the other hand, it may be that the
parents of the NW children were not taking enough caution to feed their children well, thinking that their NW child will grow with his/her body, and that overweight is not their child’s concern.13 Regardless, approximately 40% of both OW and NW children’s eating time was spent consuming unhealthful food, whereas most of this time was under direct adult supervision. The reason so much unhealthful eating was allowed under adult supervision cannot be determined at this time. Adult caregivers may need more knowledge, encouragement, skills, or assistance in food preparation and how to reduce children’s access to unhealthful food while providing children more and easier access to healthful food. This conclusion seems reasonable, because the data show that adolescents who watch television during family meals have lower-quality diets than adolescents who do not watch television during family meals.12 Moreover, watching television during family meals was associated with more healthful adolescent eating than not eating meals as a family at all. In addition, children whose parents set limits on unhealthful food consumption and screen time have better eating behaviors and more physical activity than children without parental limits.35,40 The OW children spent more time consuming healthful food and less time consuming unhealthful food than the NW children. At first, this finding may seem counterintuitive. However, there are some possible explanations for this perplexing result. The tool used to record dietary behavior was not a food diary, food frequency questionnaire, or any other type of standardized dietary analysis tool. The activity diary was a behavior diary, which recorded whether the children were eating, the time they spent eating, and what type of food they were eating. Thus, the amount of food consumed during any given time period was not recorded. The energy content of the food consumed was not known, nor was the speed of food consumption. An example of a scenario that would fit the data would be that an OW child consumed a large candy bar, whereas a NW child consumed a few hard mint candies. It would take the OW child much less time to eat the candy bar, which
Journal of Nutrition Education and Behavior Volume 44, Number 4, 2012 contained more energy, than the NW child would spend sucking on a few mints, which contained less energy. An alternative that cannot be ignored is that the OW children were biased in reporting their food intake. Unfortunately, the study design did not provide a mechanism to validate food intake recording validity beyond that of self-report. Overweight girls reported spending more of their eating time consuming healthful food than OW boys (Table 2). The key to more healthful eating for the OW girls may be that they spent less of their eating time alone when compared to OW boys. On the other hand, the OW girls spent more of their sedentary time eating than OW boys. So, although OW girls were spending more time eating than OW boys, the girls were consuming more healthful food.
Study Limitations The participants in this study were students who volunteered to participate. The database generated does not provide a way to determine whether nonvolunteer students would have been characterized differently from these volunteers. African Americans were overrepresented in the database, whereas Hispanics were underrepresented. However, these racial proportions probably did not bias the data, because there were no significant differences found for activity among the races. A power analysis to determine the number of students needed to detect a significant difference, if there was one, was not conducted. Therefore, nonsignificant results should be interpreted with caution, as there may have been too few students to see a difference. The 3 study sites, representing 3 distinct cross-sections of the population, may not be representative of the entire population. In spite of these limitations, the data provide the first detailed examination of how afterschool activities of middle school students may affect their weight.
IMPLICATIONS FOR RESEARCH AND PRACTICE The data from this research provide several suggestions for foci of interven-
tions in the fight against childhood obesity. From the standpoint of physical activity, middle school-aged children should be encouraged to spend time with their friends and family members rather than spending time alone. However, the data show that more adult supervision does not guarantee more physical activity for the child. Adult caretakers should learn how to be more physically engaged with their children, rather than just supervise sedentary behavior. Alternatively, supervising adults can encourage after-school physical activity for children, even if the adult is not able to directly participate. With regard to more healthful eating, adults should better structure the home food environment such that children have easy access to healthful food, limited access to unhealthful food, and are given direction and encouragement to eat healthful food when unsupervised. In order to implement these mediators of healthful eating and physical activity more effectively than is being done now, a concerted effort needs to be made. This effort needs to promote more social interactions among children, family members, friends, and adult caretakers, where physical activity and healthful eating are more likely to occur.
ACKNOWLEDGMENTS This study was funded by the District of Columbia Department of Transportation, Washington, DC.
REFERENCES 1. Ogden CL, Carroll MD, Curtin LR, Lamb MM, Flagel KM. Prevalence of high body mass index in US children and adolescents, 2007-2008. JAMA. 2010;303:242-249. 2. McDonald NC. Active transportation to school: trends among US school children 1969–2001. Am J Prev Med. 2007;32:509-516. 3. Rye JA, O’Hara-Tompkins N, Eck E, Neal WA. Promoting youth physical activity and healthy weight through schools. WV Med J. 2008;104: 12-15. 4. Duke J, Huhman M, Heitzler C. Physical activity levels among children aged 9-13 years - United States. Morb Mortal Wkly Rep. 2003;52:33.
Miller et al 333 5. Institute of Medicine and National Research Council of the National Academies. Local Government Actions to Prevent Childhood Obesity. Washington, DC: National Academies Press; 2009; 1–126. 6. Gillman MW, Rifas-Shiman SL, Frazier AL, et al. Family dinner and diet quality among older children and adolescents. Arch Fam Med. 2000;9: 235-240. 7. Neumark-Sztainer D, Hannan PJ, Story M, Croll J, Perry C. Family meal patterns: associations with sociodemographic characteristics and improved dietary intake among adolescents. J Am Diet Assoc. 2003;103:317-322. 8. Taveras EM, Rifas-Shiman SL, Berkey CS, et al. Family dinner and adolescent overweight. Obes Res. 2005;13:900-906. 9. Wu F-E, Shu Y, Ien-Lan W. Weight control behavior among obese children: association with family related factors. J Nursing Res. 2003;11:19-29. 10. Dammann K, Smith C. Food-related attitudes and behaviors at home, school, and restaurants: perspectives from racially diverse, urban, low-income 9to 13-year-old children in Minnesota. J Nutr Educ Behav. 2010;42:389-397. 11. Burgess-Champoux TL, Larson N, Neumark-Sztainer D, Hannan PJ, Story M. Are family meal patterns associated with overall diet quality during transition from early to middle adolescence? J Nutr Educ Behav. 2009;41:79-86. 12. Feldman S, Eisenberg ME, NeumarkSztainer D, Story M. Associations between watching TV during family meals and dietary intake among adolescents. J Nutr Educ Behav. 2007;39: 257-263. 13. O’Dougherty M, Story M, Lytle L. Food choices of young AfricanAmerican and Latino adolescents: where do parents fit in? J Am Diet Assoc. 2006;106:1846-1850. 14. French SA, Harnack L, Jeffery RW. Fast food restaurant use among women in the pound of prevention study: dietary, behavioral and demographic correlates. Int J Obes. 2000;24:1353-1359. 15. US Census Bureau. Employment Status and Work Status in 1999 of Family Members - Census 2000 Summary. http:// factfinder.census.gov/servlet/QTTable?_ bm¼y&-geo_id¼D&-qr_name¼DEC_ 2000_SF3_U_QTP26&-ds_name¼ D&-_lang¼en&-redoLog¼false. Accessed March 24, 2011. 16. US Census Bureau. Children in Selfcare, by age of child, employment status
Journal of Nutrition Education and Behavior Volume 44, Number 4, 2012
334 Miller et al
17.
18.
19.
20.
21.
22.
23.
24.
of mother, and selected characteristics for children living with mother: Summer 2006. http://www.census.gov/ hhes/childcare/data/sipp/2006/tables.html. Accessed November 1, 2011. Casperson CJ, Pereira MA, Curran KM. Changes in physical activity patterns in the United States, by sex and cross-sectional age. Med Sci Sports Exerc. 2000;32:1601-1609. Webber LS, Catellier DJ, Lytle LA, et al. Promoting physical activity in middle school girls: trail of activity for adolescent girls. Am J Prev Med. 2008; 34:173-184. Puyau MR, Adolph AL, Vohra RA, Zakeri I, Butte NF. Prediction of activity energy expenditure using accelerometers in children. Med Sci Sports Exerc. 2004;36:1625-1631. Pffeiffer KA, Mciver KL, Dowda M, Almeida MJCA, Pate RR. Validation and calibration of the Actical accelerometer in preschool children. Med Sci Sports Exerc. 2006;38:152-157. Eisenmann JC, Katzmarzyk PT, Perusse L, Bouchard C, Malina RM. Estimated daily energy expenditure and blood lipids in adolescents: the Quebec family study. J Adolesc Health. 2003;33:147-153. Bouchard C, Tremblay A, Leblanc C, Lortie G, Savard R, Theriault G. A method to assess energy expenditure in children and adults. Am J Clin Nutr. 1983;37:461-467. Jeffery RW, French SA. Epidemic obesity in the United States: are fast foods and television viewing contributing? Am J Public Health. 1998;88: 277-280. Trost S, Russell P, Freedson P, Sallis J, Taylor W. Using objective physical activity measures with youth: how many days of monitoring are
25.
26.
27.
28.
29.
30.
31.
32.
33.
needed? Med Sci Sports Exerc. 2000; 32:426-431. Freedson P, Pober D, Janz KF. Calibration of accelerometer output for children. Med Sci Sports Exerc. 2005; 37(11 suppl):S523-S530. Page A, Cooper AR, Stamatakis E, et al. Physical activity patterns in nonobese and obese children assessed using minute-by-minute accelerometry. Int J Obes. 2005;29:1070-1076. Gidlow CJ, Cochrane T, Davey R, Smith H. In-school and out-of-school physical activity in primary and secondary school children. J Sports Sci. 2008; 26:1411-1419. Guinhouya CB, Sourvrier S, Vilhelm C, et al. Physical activity and sedentary lifestyle in children as timelimited functions: usefulness of the principal component analysis method. Behav Res Method. 2007;39:682-688. Jago R, Fox K, Page AS, Brockman R, Thompson JL. Physical activity and sedentary behaviour typologies of 10-11 year olds. Int J Behav Nutr Phys Act. 2010;7:59. http://www.ncbi.nlm.nih.gov/ pmc/articles/PMC2918527/. Accessed January 5, 2012. Mota J, Santos P, Guerra S, Ribeiro J, Duarte J. Patterns of daily physical activity during school days in children and adolescents. Am J Hum Biol. 2003; 15:547-553. Sirard JR, Riner WF Jr, Mciver KL, Pate RR. Physical activity and active commuting to elementary school. Med Sci Sports Exerc. 2005;37:2062-2069. Roemmich J, Leonard E, Samina R. The neighborhood and home environments: disparate relationships with physical activity and sedentary behaviors in youth. Ann Behav Med. 2007;33:29-38. Lox C. The Psychology of Exercise: Integrating Theory and Practice. 2nd
34.
35.
36.
37.
38.
39.
40.
ed. Halcomb Scottsdale, AZ: Holcomb Hathaway Publishers; 2006; 114–115. Treuth MS, Catellier DJ, Schmitz KH, et al. Weekend and weekly patterns of physical activity in overweight and normal-weight adolescent girls. Obesity. 2007;15:1782-1788. Meizi H, Piche L, Beynon C, Harris S. Screen-related sedentary behaviors: children’s and parents’ attitudes, motivations, and practices. J Nutr Educ Behav. 2010;42:17-25. Ness AR, Leary SD, Mattocks C, et al. Objectively measured physical activity and fat mass in a large cohort of children. PLoS Med. 2007;4:e97. http:// www.plosmedicine.org/article/info% 3Adoi%2F10.1371%2Fjournal.pmed .0040097. Accessed January 5, 2012. Purslow LR, Hill C, Saxton J, Corder K, Wardel J. Differences in physical activity and sedentary time in relation to weight in 8-9 year old children. Int J Behav Nutr Phys Act. 2008;5:67. http://www.ncbi .nlm.nih.gov/pmc/articles/PMC2636837/. Accessed January 5, 2012. Stager M, Harvey R, Secic M, Camlin-Shingler K, Cromer B. Self-reported physical activity and bond mineral density in urban adolescent girls. J Pediatr Adolesc Gynecol. 2006; 19:17-22. Treuth MS, Hou N, Young DR, Maynard LM. Validity and reliability of the Fels Physical Activity Questionnaire of children. Med Sci Sports Exerc. 2005;37:488-495. Nickelson J, Roseman MG, Forthofer MS. Associations between parental limits, school vending machine purchases, and soft drink consumption among Kentucky middle school students. J Nutr Educ Behav. 2010;42: 115-122.
INTRODUCTION It is well recognized that the prevalence of obesity in children has risen substantially over the past 3 decades. Although this prevalence trend has flattened over the past several years, childhood obesity remains a major problem in the United States.1 The high prevalence of childhood obesity experienced today is a result of complex social, environmental, and policy changes that have influenced eating and physical activity behaviors. For example, many urban and
1
suburban communities are designed in such a way that walking, bicycling, and other physical activities are discouraged. There are fewer opportunities today than a generation ago for children to be active during school, after school, and going to and from school.2,3 The data indicate that 62% of children aged 9–13 years do not participate in any organized physical activity during their nonschool hours and that 23% do not engage in any free-time physical activity.4 Children spend less of their free time engaged in physical activity outdoors
Center for Rural and Community Health, West Virginia School of Osteopathic Medicine, Lewisburg, WV 2 Department of Exercise Science, The George Washington University Medical Center, Washington, DC † Dr. Wayne C. Miller was affiliated with the Department of Exercise Science, The George Washington University Medical Center at the time this study was completed. Address for correspondence: Wayne C. Miller, PhD, Center for Rural and Community Health, West Virginia School of Osteopathic Medicine, 400 North Lee St, Lewisburg, WV 24901; Phone: (304) 647-6380; Fax: (304) 793-6586; E-mail: [email protected] .edu Ó2012 SOCIETY FOR NUTRITION EDUCATION AND BEHAVIOR doi:10.1016/j.jneb.2011.08.002
326
and more free time being sedentary, either watching TV or playing video games, than in years past.5 Evolutions in family life have caused changes in children’s eating patterns. The family can have both negative and positive effects on children’s eating behaviors.6-12 Crosssectional studies have shown that children and adolescents who regularly eat dinner with family are less likely to be overweight and more likely to have more healthful eating habits.6-8,11 However, modern families experience pressure to minimize food costs and to shorten food acquisition and preparation time, which has led many families to regularly consume unhealthful processed food that is more conveniently acquired and prepared for consumption than home-prepared, fresh food.13,14 Studies have shown that busy households, in which families dine out for meals, subject children to food that is high in energy density, as well as high in sodium.9 Over the past 2 decades, the number of working parents has increased
Journal of Nutrition Education and Behavior Volume 44, Number 4, 2012
Journal of Nutrition Education and Behavior Volume 44, Number 4, 2012 dramatically; 75% of mothers with school-aged children are now employed.15 Both parents also work an increasing number of hours during the week,15 which means that the number of children forced to care for themselves has also increased dramatically. In 2006, it was reported that 18% of children ages 9–11 years were taking care of themselves.16 The size of the family has an impact on health behaviors as well. Children in small families have more healthful eating patterns than children in large families.8 Presumably, the larger the family, the less likely a child’s eating habits will be directly supervised by an adult.8 Other researchers found that lack of adult supervision caused unsupervised children to participate in less physical activity and consume larger amounts of unhealthful food than supervised children.15 Adult supervision, therefore, seems to be a key in promoting healthful eating as well as physical activity patterns in children. It appears that there are 2 major factors that have contributed to the increase in childhood obesity over the past 3 decades: a community environment that discourages physical activity and healthful eating, and a family environment in which adults are less able to supervise childhood behaviors. Therefore, to influence/ impact rates of childhood obesity, one must examine the children’s environment, with particular focus on how children spend their time. Better understanding of how children spend their time and under what circumstances they participate in healthful or unhealthful behaviors will help in the design of more effective childhood obesity interventions. Therefore, the purpose of this study was to examine the physical activity patterns, eating behaviors, and social environment of overweight (OW) and normal weight (NW) middle school children between the hours of 3:00 PM and 12:00 AM, when children are most likely to engage in behaviors unsupervised by adults. The working hypothesis for this study was that during these after-school hours, OW children are less active, have poorer eating behaviors, and have a weaker social environment for health behaviors than NW children.
METHODS Study Design After-school physical activity patterns, eating behaviors, and social environment were examined simultaneously in middle school students. Physical activity was measured objectively through accelerometry and recorded subjectively by each student with an activity diary. Students also recorded their eating behaviors and social environment in the diary. Weight comparisons (OW vs NW) and sex comparisons were made to identify possible determinants of obesity in middle school students.
Participants This study was approved by the Institutional Review Board for Human Research at the George Washington University Medical Center. Three middle schools (grades 6–8) were selected for this study, because it is well known that physical activity declines among youth as they grow through adolescence.4,17,18 One middle school (DC) was selected from the inner city of Washington DC, a second school (VA) was selected from a mid-sized town located in northern Virginia, and the third school (GA) was selected from a rural agricultural community in Georgia. These 3 sites were selected to represent an urban metropolis, a ‘‘blue collar’’ town, and a farming community. The per capita income for the DC site was unknown, for the VA site $23,552, and for the GA site $13,657. Although direct generalizeability to the entire American population may not be possible with this sampling technique, these 3 sample cities represent 3 distinct cross-sections of the American population. Cooperation was obtained from the school board, principal, and physical education teachers of the respective schools. Students were recruited for this study via class announcements. Researchers gave a 15-minute presentation about the nature of the study to students in their physical education classes. Interested students were then given a parent authorization form and a child assent form to be completed and signed prior to participation in the study.
Miller et al 327
Protocol The DC school had 57 out of 180 potential children participating; the VA school had 48 out of 200 potential children, and the GA school 36 of 204 students participating. All of the DC students were African American; of the VA students, 90% were Caucasian and 10% were African American; and of the GA students, 39% were African American, 50% Caucasian, and 11% Hispanic. None of the students was practicing weight control methods. Students were not given any monetary incentive to participate. Researchers visited the school on the Monday of a full week of school to gather anthropometric data, equip the children with accelerometers, and provide instruction on completion of the activity logs. Monday morning, the participants were taken into a private room away from the rest of their physical education class, where their weight and height were obtained using a scale (Taylor Precision Tech #7506; Taylor Precision Products, Oak Brook, IL, 2007) and stadiometer (Secca Portable Stadiometer #213; Secca, Hanover, MD, 2007) that were accurate to 0.25 lb (0.114 kg) and 0.25 in (0.64 cm), respectively. Children were classified as NW if their body mass index (BMI) for age and sex was $ 5% and < 85% and as OW if their BMI for age and sex was $ 85%. Demographic data (height, weight, age, and sex) were input into a laptop computer so the participant could be fitted for an accelerometer. Participants were then fitted with an Actical accelerometer (Respironics Inc, Bend, OR, 2007) attached to a waistband with a velcro strap. The accelerometer was secured around the hips with the actual monitor lying just above the right iliac crest. Participants were shown how to remove the accelerometer at night and replace it in the morning. Children were told that the monitors were waterproof and thus could be worn while showering or swimming. Children were asked to keep the accelerometer on at all times from 3:00 PM until bedtime on the days of the study. The Actical accelerometer is sensitive to movement in all directions, and the storage capacity allows for
328 Miller et al several days of continuous monitoring.19,20 The Actical accelerometer is valid for detection of low-energy movements as well as high-energy movements.19,20 Along with an accelerometer, each participant was given a 4-day activity diary to complete during the experimental week, from that Monday, beginning at 3:00 PM, until that Thursday at 12:00 AM. The activity diary was based on the Bouchard 3-day physical activity diary, which has been found to be both valid and reliable for estimating daily energy expenditure for ages 10–50 years.21,22 The diary was modified so that 20-minute time intervals were used, instead of 15-minute intervals, for the participant to record the activities in which he or she took part. Our preliminary work with children this age suggested that expanding the time frame from 15–20 minutes would help with compliance in completing the diary (WCM, unpublished data, 2010). The 20-minute increments allowed for assignment of a categorical value for the dominant activity in any given time frame. A further modification to the diary was that, in addition to recording the activity in which the child participated, the child also recorded his or her location, who was with him or her, and whether there was adult supervision. Also included on the diary was an example of how to fill it out properly, with accompanying instructions. Researchers returned to the school on Friday to retrieve both the activity diaries and accelerometers.
Compression and Coding of Data Researchers coded the responses on the returned activity diaries so that location was categorized as Indoors at Home, Indoors Out of Home, and Outdoors. The ‘‘with whom’’ response was coded as either with a family member, with a friend, or alone. Adult supervision was defined as a person 18 years old or older who was in the act of overseeing the adolescent. The supervising adult must have been within voice range of the child to be considered supervising. Distinct verbal and written instructions were given to the children on how to record where they were, whom they were with, and whether an adult was supervising. The researcher who
Journal of Nutrition Education and Behavior Volume 44, Number 4, 2012 coded the diary responses was given the same set of instructions for reporting as the children. The researcher then reviewed each returned diary for consistency between the instructions and the way the children completed the diary. If a discrepancy was found, the researcher either corrected the discrepancy (according to the written coding instructions) or eliminated that particular response from the database. In addition, the participant’s activities were coded as either active or sedentary; sedentary was defined as anything done sitting or lying down (including napping). Activities were also coded as eating (yes or no) and what type of food was being consumed (healthful food or unhealthful food). Eating fast food was classified separately because of the relationship between fast-food consumption and obesity.23 Fruit, 100% fruit juice, vegetables, grains, lean meats, fish, and low-fat dairy (1% fat or lower) were classified as healthful food. High-fat snacks (eg, regular fat cheese, peanut butter, chips), sweets, sugared drinks, diet soda, high-sodium snacks (eg, chips, crackers, pretzels), and high-calorie/ low-nutrient food (eg, pizza, macaroni and cheese) were classified as unhealthful food. All of the codes were predetermined and standardized by 1 investigator. The activity diary was completed only from 3:00 PM to 12:00 AM for 4 days, yielding a total of 27 consecutive 20-minute time intervals per day, and 108 total time intervals per participant. Since the activity diary was divided into 20-minute time increments, data from the accelerometer were also compiled into 20-minute time intervals to give the total activity counts for each corresponding activity diary time interval. If there were 3 consecutive time intervals (60 consecutive minutes) of accelerometer data for where activity counts were measured to be zero, information obtained from the activity diary was used to determine whether the participant was sedentary or just not wearing the monitor. If the activity diary reported a period of being awake when there were no accelerometer counts, it was assumed that the participant was not wearing the accelerometer, and the corresponding time
interval for accelerometer data was voided. Data collection stopped for the day when the accelerometer registered zero counts and the child recorded going to bed for the night on the activity diary. Therefore, the dataset includes only activity data for the time a child was awake between 3:00 PM and 12:00 AM. Empty time cells were not included in data analysis. A reliability coefficient of 0.80 was previously recorded for students in grades 1–6 and of 0.70 for grades 7–12 when they wore the accelerometer for 4 or 5 days.24 This study examined after-school weekday activities, so a 4-day range was deemed appropriate. The need to match children’s activity, adult supervision, and eating behaviors to accelerometer counts led to extending the Bouchard 3-day activity diary to 4 days to accommodate the accuracy and reliability of the Actical accelerometer.
Statistical Analyses All statistical procedures were performed with the Systat statistical software package (version 12, SYSTAT Software, Inc., Chicago, IL, 2007). Values are reported as mean SD. Statistical significance was set at P < .05. Sample data were tested for normal distribution using the KolmogorovSmirnov test and for equal variance using the Levene median test. The 2sample t test was used to make mean group comparisons between the NW and OW and between boys and girls. ANOVA with repeated measures was used to make mean group comparisons for accelerometer counts across time periods. ANOVA was used to make group mean comparisons among schools and among races for the demographic variables and accelerometer counts, and to make comparisons among ages for accelerometer counts. A Tukey pairwise comparison was used if an ANOVA produced a significant F value. The chi-square test was used to compare differences in the frequency counts between groups for the categorical data reported in the activity diaries (ie, time spent in each of 19 categories of activity). A test for equality of 2 proportions was used to test for differences in prevalence ratios or percentages between 2 groups. Simple correlation coefficients were used to
All indicates all participants; DC, inner city of Washington, DC; GA, rural agricultural community in Georgia; NW, normal weight; OW, overweight; VA, mid-sized town in northern Virginia. *Significantly different from NW group (P < .001); **Significantly different from other schools (P < .05); ***Significantly different from other races (P < .05). Note: ANOVA was used to test for significant group differences among schools and races. Independent t tests were used to test for significant group differences between OW and NW and between sexes.
Girls (n ¼ 102) 12.4 0.8 54.6 14.2 158.4 9.3 21.6 4.8 Boys (n ¼ 39) 12.4 0.8 54.5 17.1 159.5 10.5 21.0 4.8 OW (n ¼ 36) 12.4 0.7 72.4 15.7* 160.8 12.5 27.9 4.5* NW (n ¼ 105) 12.4 0.9 48.4 8.6 158.0 8.4 19.2 2.0 All Group (n ¼ 141) Age (y) 12.4 0.8 Weight (kg) 54.6 15.0 Height (cm) 158.7 9.6 BMI 21.5 4.8
Table 1. Anthropometric and Ethnic Data for Middle School Students, mean (SD)
DC School (n ¼ 57) 12.5 0.8 59.9 17.1** 161.0 10.7** 23.0 5.8**
VA School (n ¼ 48) 12.1 0.7** 50.1 12.0 156.0 8.8 20.4 3.5
GA School (n ¼ 36) 12.8 0.8 52.0 12.6 158.7 8.2 20.4 3.3
African American (n ¼ 76) 12.6 0.9*** 58.3 16.9 161.0 10.4*** 22.4 5.5***
Caucasian (n ¼ 61) 12.1 0.7 50.4 11.3 156.7 8.3 20.3 3.3
Hispanic (n ¼ 4) 12.3 0.5 46.4 5.5 152.4 4.6 19.9 1.6
Journal of Nutrition Education and Behavior Volume 44, Number 4, 2012
Miller et al 329 measure the strength of association between 2 quantitative variables.
RESULTS The demographic information for the participants is presented in Table 1. There were no age differences between the NW and OW groups or between the sexes. The VA children were 5–7 months younger than children at the other 2 sites, and the African-American children were 2–5 months older than the Caucasians and Hispanic children. The children at the DC school were significantly taller and heavier and had a higher BMI than the children at the other 2 sites. The total physical activity measures, expressed as average accelerometer counts per minute for the entire time frame (3:00 PM to 12:00 AM) are shown in Figure 1. There were no significant group differences in average accelerometer counts from 3:00 PM to 12:00 AM when comparing the NW to OW, boys to girls, schools, races, or age groups. Energy expenditure algorithms for children using the Actical accelerometer were derived by Freedson et al.25 When these algorithms were applied to the data, there were no significant differences in estimated energy expenditure for any of the comparable groups. In fact, the range in energy expenditure for all groups was very narrow, from 0.017 0.002 to 0.019 0.002 kcal per kg body weight per minute (group data not shown). This result amounted to an average of 530 344 kcal expended from 3:00 PM to 12:00 AM, or approximately 1.0 kcal per minute during that time frame. The correlation coefficient for the relationship between the accelerometer counts within each of the 27 time intervals across the 4 days and the frequency of positive responses (in the activity diaries) for being active over the same time was 0.82 (P < .001). The plot of accelerometer counts for all children over time is shown in Figure 2. Physical activity was highest between 3:00 PM and 6:40 PM. After 6:40 PM, physical activity began to decline to where at 7:00 PM, activity was significantly lower than at its peak time (3:00 PM to 5:00 PM). Physical activity continued to decline until 11:00 PM.
330 Miller et al
Journal of Nutrition Education and Behavior Volume 44, Number 4, 2012
Figure 1. Group comparisons for average accelerometer counts per minute from 3:00 PM to 12:00 AM in middle school students. AA indicates African American; All, all participants; Cau, Caucasian; DC, District of Columbia students; GA, Georgia students; His, Hispanic; NW, normal weight; OW, overweight; VA, Virginia students. Note: 11, 12, 13, and 14 represent the respective student ages. Values represent the mean SD for the respective groups. ANOVA was used to test for group mean differences. Group numbers are reported in Table 1. Table 2 shows the percentage distribution of time between 3:00 PM and 12:00 AM spent in different activi-
ties and under different environmental situations. Middle school children spent over three fourths of their after-
school hours being sedentary, and approximately 80% of that sedentary time was spent under adult supervision. Eighty-five percent of children’s eating time was supervised by an adult, and 19%–26% of that adult-supervised eating was spent eating unhealthful food. The OW children spent more time being active, but less time being sedentary and less time under supervision than the NW children. The OW children also spent less time eating unhealthful food than the NW children. The OW children spent less of their after-school time being supervised and less sedentary time being supervised than NW children. Overweight girls spent more time being active, but also more time eating than OW boys. Overweight girls spent more of their eating time consuming healthful food than OW boys. Simple correlations revealed that the amount of time spent being active was positively related to time spent with friends for both the NW and OW children (Table 3). Active time was also related to time spent with family, but the relationship was stronger for the NW compared to the OW children. In contrast, time spent being sedentary was positively related to time spent being supervised for both the NW and the OW children. Sedentary time was more strongly related to time spent alone for the NW children compared to the OW children. Time spent eating healthful food was positively related to time spent with family, whereas time spent eating unhealthful food was related to time spent with friends for the NW and OW children. Time spent eating unhealthful food was inversely related to time spent under adult supervision as well as inversely related to time spent alone for both the NW and OW children; however, the strength of these relationships differed between groups.
DISCUSSION After-School Physical Activity Levels
Figure 2. Time plot of accelerometer counts from 3:00 PM to 12:00 AM in middle school students. Note: Values represent the mean SD for the entire student sample. ANOVA with repeated measures was used to test for significant within-group differences across time.
The amount of physical activity achieved (expressed as either activity counts or kcal) throughout the afterschool hours was similar between the NW and OW groups. Physical activity was highest from 3:00 PM to 6:40 PM,
Journal of Nutrition Education and Behavior Volume 44, Number 4, 2012 Table 2. Percentage Use of After-School Time for Middle School Students Category (%) Time spent eating Time spent with a family member Time spent with friends Time spent alone Time spent sedentary Time spent being active Time spent supervised Time spent outside Outside time being active Eating time eating healthful food Eating time eating unhealthful food Eating time supervised Supervised eating time eating unhealthful food Eating time with a family member Family eating time eating unhealthful food Eating time alone Alone eating time eating unhealthful food Sedentary time eating Sedentary time supervised
All 9.0 36.8 18.9 44.3 76.0 24.0 83.7 15.1 60.6 59.1 40.9 85.2 21.1
NW 9.6 39.2 19.7 41.1 77.3 22.7 85.8 15.5 61.4 55.0 45.0 85.7 20.7
OW 7.3 29.2* 16.5 54.3* 70.1* 29.9* 77.5* 14.6 58.6 71.9* 28.1* 83.4 22.0
OW Boys 5.4 30.6 12.6 56.8 78.4 21.6 78.9 15.6 65.3 56.9 43.1 78.5 26.2
OW Girls 8.3** 28.5 18.5** 53.0** 68.6** 31.4** 76.8 14.1 56.2** 76.7** 23.3** 80.4 19.4**
70.2 25.6
70.6 25.7
69.1 25.1
64.6 23.8
70.4 25.6
18.9 60.8
18.7 59.9
19.5 61.2
30.8 61.0
15.8** 61.4
10.9 86.6
11.5 87.7
9.9* 83.3*
6.8 82.3
10.3** 83.8
All indicates all participants; NW, normal weight; OW, overweight. *Significantly different from NW group (P < .05); **Significantly different from OW boys (P < .05). Note: A test for quality of 2 proportions was used to test for significant differences in prevalence ratios or percentages between 2 comparable groups.
and it then declined throughout the evening and night (Figure 2). This finding is not surprising in that the common routine for middle school children in the United States is free time immediately after school, dinner time, and home and family time, followed by quiet time and bed time.
Several other studies concur that the most active time in a child’s day is immediately after school.19,26-31 However, none of these studies was as complete and detailed as the present study. None of these studies sampled children from 3 distinct demographic areas (eg, urban, rural
Table 3. Relationships between Social Interaction Time and Eating or Activity Time for Middle School Students Category Active time vs time with friends Active time vs time with family Sedentary time vs time supervised Sedentary time vs time alone Time spent eating healthfully vs time with family Time spent eating unhealthfully vs time with friends Time spent eating unhealthfully vs time supervised Time spent eating unhealthfully vs time alone
All 0.64 0.46 0.76 0.88 0.87 0.89 0.82 0.66
NW 0.65 0.49 0.76 0.90 0.88 0.90 0.87 0.68
OW 0.62 0.36* 0.76 0.80* 0.81 0.87 0.78* 0.59*
All indicates all participants; NW, normal weight; OW, overweight. *Significantly different from NW group (P < .05). Note: A test for quality of 2 proportions was used to test for significant differences in prevalence ratios or percentages between the NW and OW groups. Values are correlation coefficients (P < .01).
Miller et al 331 blue collar, rural agricultural). Each of these studies sampled children from only 1 geographical location. None of these studies performed an hourby-hour analysis of physical activity levels. The few studies that evaluated physical activity with reference to time grouped the data into blocks of time.27,29-34 The time blocks analyzed were broad, representing either total time out of school24 or the block of time immediately after school until 6:00 PM and then again from 6:00 PM until 9:00 or 10:00 PM.28,29,31 Separating time frames at 6:00 PM may not have captured the best representation of physical activity patterns, because physical activity in these children decreased very rapidly from 5:00–7:00 PM. More interesting than the time of day when activities happened might be the amount of time children devoted to different types of active or sedentary behaviors. These middle school children spent over three fourths of their after-school time being sedentary (Table 2), which translates into spending 6.8 hours between 3:00 PM and 12:00 AM either sitting or lying down. Furthermore, most of the children’s sedentary time was spent under supervision by an adult (Table 2). Other researchers have suggested that the home environment created by parents likely influences the choices made by adolescents about physical activity.32 According to behavioral economic theory, a child’s choice to be sedentary or active depends on the ease of access to sedentary activities in the home.32 For instance, having many televisions in the home, or easy access to video games and computers leads to less time in an active environment. On the other hand, having dogs, for example, promotes greater amounts of physical activity.32 Other research supports this theory in that more than two thirds of fifth- and sixthgraders surveyed said they would chose to be more physically active if given the choice to do so.35 At the same time, these children reported that spending time with family was a major reason they engaged in screen-related activities. On a positive note, when children were outside, they spent about 60% of their time being active. This finding
332 Miller et al suggests that 1 area of focus in combating childhood obesity is to have children spend more time outside. Social support from parents and other family members has also been identified as 1 of the most important determinants of participation in all forms of physical activity, including sports, structured exercise, and active leisure inside and outside the home.33 Parents are responsible for providing different kinds of support in the form of transportation, equipment, paying activity fees, encouraging their children to be active, and playing with their children, all of which promote physical activity.33
Physical Activity Patterns in NW vs OW It may be surprising that the OW children spent more time being active than the NW children. These selfreport data could infer a potential reporting bias in which OW children overestimated their active behaviors. However, the accelerometer counts for the OW and NW children were similar. It may be that when the NW children were physically active, their intensity of activity was greater than that of the OW children. Under this scenario, OW children would have needed to accumulate more active time at their lower activity intensity in order to accrue the same number of accelerometer counts as the NW. Other investigators have shown that NW children do spend more time in moderate to vigorous activities than OW children.26,34,36,37 Three of these studies were consistent in finding that NW children were more active than OW children overall, and that they accrued more minutes of moderate to vigorous activity throughout the day than OW children.26,34,36 However, Purslow et al did not find any relationship between weight status and total daily activity or moderate to vigorous activity for girls.37 Furthermore, these same investigators did not find any relationship between sedentary activity and weight status for either boys or girls. None of the previous studies simultaneously used subjective and objective measures of physical activity.26,34,36,37 The authors monitored physical activity simultaneously with objective measures (accelerometers) and subjective measures (activity diaries).
Journal of Nutrition Education and Behavior Volume 44, Number 4, 2012 The correlation coefficient between the average accelerometer counts per minute and the average frequency of positive responses for being active on the diaries during each of the 27 time frames was high (r ¼ 0.82) for the OW children. This finding suggests that the OW children were accurately recording their activity time. The OW girls reported spending more time being active than the OW boys, when their estimated energy expenditure was the same. However, it may well have been that the activities reported by the OW girls were associated with lower-intensity activities such as chores and activities of daily living, whereas those of the OW boys were associated with short bouts of vigorous activity, such as organized games or sports.38 Preliminary data support this theory, in that girls scored lower than boys on the leisure index of the Fels Physical Activity Questionnaire when their total scores for the questionnaire were equivalent (WCM et al, unpublished data, 2010).39 Overweight girls spent less of their outside time being active than OW boys (Table 2), which may have occurred because the outdoor environment is more favorable to boys’ activities than girls’ activities.32 Roemmich et al discovered that the built environment affects boys and girls differently.32 These researchers reported a positive relationship between the percentage of recreational park area in a neighborhood and physical activity for boys, but no such relationship for girls.
Dietary Patterns and Behaviors The great majority of time children spent eating was under direct adult supervision and/or with a family member (Table 2). Approximately 60% of children’s total eating time was spent eating healthful food (Table 2). When children ate alone, they spent 60% of that time eating unhealthful food. These data support the idea that parents feel it is important to feed their children well so they will grow to be big and strong.13 This may be the case with the OW children, in that they spent more of their eating time consuming healthful food than the NW children (Table 2). On the other hand, it may be that the
parents of the NW children were not taking enough caution to feed their children well, thinking that their NW child will grow with his/her body, and that overweight is not their child’s concern.13 Regardless, approximately 40% of both OW and NW children’s eating time was spent consuming unhealthful food, whereas most of this time was under direct adult supervision. The reason so much unhealthful eating was allowed under adult supervision cannot be determined at this time. Adult caregivers may need more knowledge, encouragement, skills, or assistance in food preparation and how to reduce children’s access to unhealthful food while providing children more and easier access to healthful food. This conclusion seems reasonable, because the data show that adolescents who watch television during family meals have lower-quality diets than adolescents who do not watch television during family meals.12 Moreover, watching television during family meals was associated with more healthful adolescent eating than not eating meals as a family at all. In addition, children whose parents set limits on unhealthful food consumption and screen time have better eating behaviors and more physical activity than children without parental limits.35,40 The OW children spent more time consuming healthful food and less time consuming unhealthful food than the NW children. At first, this finding may seem counterintuitive. However, there are some possible explanations for this perplexing result. The tool used to record dietary behavior was not a food diary, food frequency questionnaire, or any other type of standardized dietary analysis tool. The activity diary was a behavior diary, which recorded whether the children were eating, the time they spent eating, and what type of food they were eating. Thus, the amount of food consumed during any given time period was not recorded. The energy content of the food consumed was not known, nor was the speed of food consumption. An example of a scenario that would fit the data would be that an OW child consumed a large candy bar, whereas a NW child consumed a few hard mint candies. It would take the OW child much less time to eat the candy bar, which
Journal of Nutrition Education and Behavior Volume 44, Number 4, 2012 contained more energy, than the NW child would spend sucking on a few mints, which contained less energy. An alternative that cannot be ignored is that the OW children were biased in reporting their food intake. Unfortunately, the study design did not provide a mechanism to validate food intake recording validity beyond that of self-report. Overweight girls reported spending more of their eating time consuming healthful food than OW boys (Table 2). The key to more healthful eating for the OW girls may be that they spent less of their eating time alone when compared to OW boys. On the other hand, the OW girls spent more of their sedentary time eating than OW boys. So, although OW girls were spending more time eating than OW boys, the girls were consuming more healthful food.
Study Limitations The participants in this study were students who volunteered to participate. The database generated does not provide a way to determine whether nonvolunteer students would have been characterized differently from these volunteers. African Americans were overrepresented in the database, whereas Hispanics were underrepresented. However, these racial proportions probably did not bias the data, because there were no significant differences found for activity among the races. A power analysis to determine the number of students needed to detect a significant difference, if there was one, was not conducted. Therefore, nonsignificant results should be interpreted with caution, as there may have been too few students to see a difference. The 3 study sites, representing 3 distinct cross-sections of the population, may not be representative of the entire population. In spite of these limitations, the data provide the first detailed examination of how afterschool activities of middle school students may affect their weight.
IMPLICATIONS FOR RESEARCH AND PRACTICE The data from this research provide several suggestions for foci of interven-
tions in the fight against childhood obesity. From the standpoint of physical activity, middle school-aged children should be encouraged to spend time with their friends and family members rather than spending time alone. However, the data show that more adult supervision does not guarantee more physical activity for the child. Adult caretakers should learn how to be more physically engaged with their children, rather than just supervise sedentary behavior. Alternatively, supervising adults can encourage after-school physical activity for children, even if the adult is not able to directly participate. With regard to more healthful eating, adults should better structure the home food environment such that children have easy access to healthful food, limited access to unhealthful food, and are given direction and encouragement to eat healthful food when unsupervised. In order to implement these mediators of healthful eating and physical activity more effectively than is being done now, a concerted effort needs to be made. This effort needs to promote more social interactions among children, family members, friends, and adult caretakers, where physical activity and healthful eating are more likely to occur.
ACKNOWLEDGMENTS This study was funded by the District of Columbia Department of Transportation, Washington, DC.
REFERENCES 1. Ogden CL, Carroll MD, Curtin LR, Lamb MM, Flagel KM. Prevalence of high body mass index in US children and adolescents, 2007-2008. JAMA. 2010;303:242-249. 2. McDonald NC. Active transportation to school: trends among US school children 1969–2001. Am J Prev Med. 2007;32:509-516. 3. Rye JA, O’Hara-Tompkins N, Eck E, Neal WA. Promoting youth physical activity and healthy weight through schools. WV Med J. 2008;104: 12-15. 4. Duke J, Huhman M, Heitzler C. Physical activity levels among children aged 9-13 years - United States. Morb Mortal Wkly Rep. 2003;52:33.
Miller et al 333 5. Institute of Medicine and National Research Council of the National Academies. Local Government Actions to Prevent Childhood Obesity. Washington, DC: National Academies Press; 2009; 1–126. 6. Gillman MW, Rifas-Shiman SL, Frazier AL, et al. Family dinner and diet quality among older children and adolescents. Arch Fam Med. 2000;9: 235-240. 7. Neumark-Sztainer D, Hannan PJ, Story M, Croll J, Perry C. Family meal patterns: associations with sociodemographic characteristics and improved dietary intake among adolescents. J Am Diet Assoc. 2003;103:317-322. 8. Taveras EM, Rifas-Shiman SL, Berkey CS, et al. Family dinner and adolescent overweight. Obes Res. 2005;13:900-906. 9. Wu F-E, Shu Y, Ien-Lan W. Weight control behavior among obese children: association with family related factors. J Nursing Res. 2003;11:19-29. 10. Dammann K, Smith C. Food-related attitudes and behaviors at home, school, and restaurants: perspectives from racially diverse, urban, low-income 9to 13-year-old children in Minnesota. J Nutr Educ Behav. 2010;42:389-397. 11. Burgess-Champoux TL, Larson N, Neumark-Sztainer D, Hannan PJ, Story M. Are family meal patterns associated with overall diet quality during transition from early to middle adolescence? J Nutr Educ Behav. 2009;41:79-86. 12. Feldman S, Eisenberg ME, NeumarkSztainer D, Story M. Associations between watching TV during family meals and dietary intake among adolescents. J Nutr Educ Behav. 2007;39: 257-263. 13. O’Dougherty M, Story M, Lytle L. Food choices of young AfricanAmerican and Latino adolescents: where do parents fit in? J Am Diet Assoc. 2006;106:1846-1850. 14. French SA, Harnack L, Jeffery RW. Fast food restaurant use among women in the pound of prevention study: dietary, behavioral and demographic correlates. Int J Obes. 2000;24:1353-1359. 15. US Census Bureau. Employment Status and Work Status in 1999 of Family Members - Census 2000 Summary. http:// factfinder.census.gov/servlet/QTTable?_ bm¼y&-geo_id¼D&-qr_name¼DEC_ 2000_SF3_U_QTP26&-ds_name¼ D&-_lang¼en&-redoLog¼false. Accessed March 24, 2011. 16. US Census Bureau. Children in Selfcare, by age of child, employment status
Journal of Nutrition Education and Behavior Volume 44, Number 4, 2012
334 Miller et al
17.
18.
19.
20.
21.
22.
23.
24.
of mother, and selected characteristics for children living with mother: Summer 2006. http://www.census.gov/ hhes/childcare/data/sipp/2006/tables.html. Accessed November 1, 2011. Casperson CJ, Pereira MA, Curran KM. Changes in physical activity patterns in the United States, by sex and cross-sectional age. Med Sci Sports Exerc. 2000;32:1601-1609. Webber LS, Catellier DJ, Lytle LA, et al. Promoting physical activity in middle school girls: trail of activity for adolescent girls. Am J Prev Med. 2008; 34:173-184. Puyau MR, Adolph AL, Vohra RA, Zakeri I, Butte NF. Prediction of activity energy expenditure using accelerometers in children. Med Sci Sports Exerc. 2004;36:1625-1631. Pffeiffer KA, Mciver KL, Dowda M, Almeida MJCA, Pate RR. Validation and calibration of the Actical accelerometer in preschool children. Med Sci Sports Exerc. 2006;38:152-157. Eisenmann JC, Katzmarzyk PT, Perusse L, Bouchard C, Malina RM. Estimated daily energy expenditure and blood lipids in adolescents: the Quebec family study. J Adolesc Health. 2003;33:147-153. Bouchard C, Tremblay A, Leblanc C, Lortie G, Savard R, Theriault G. A method to assess energy expenditure in children and adults. Am J Clin Nutr. 1983;37:461-467. Jeffery RW, French SA. Epidemic obesity in the United States: are fast foods and television viewing contributing? Am J Public Health. 1998;88: 277-280. Trost S, Russell P, Freedson P, Sallis J, Taylor W. Using objective physical activity measures with youth: how many days of monitoring are
25.
26.
27.
28.
29.
30.
31.
32.
33.
needed? Med Sci Sports Exerc. 2000; 32:426-431. Freedson P, Pober D, Janz KF. Calibration of accelerometer output for children. Med Sci Sports Exerc. 2005; 37(11 suppl):S523-S530. Page A, Cooper AR, Stamatakis E, et al. Physical activity patterns in nonobese and obese children assessed using minute-by-minute accelerometry. Int J Obes. 2005;29:1070-1076. Gidlow CJ, Cochrane T, Davey R, Smith H. In-school and out-of-school physical activity in primary and secondary school children. J Sports Sci. 2008; 26:1411-1419. Guinhouya CB, Sourvrier S, Vilhelm C, et al. Physical activity and sedentary lifestyle in children as timelimited functions: usefulness of the principal component analysis method. Behav Res Method. 2007;39:682-688. Jago R, Fox K, Page AS, Brockman R, Thompson JL. Physical activity and sedentary behaviour typologies of 10-11 year olds. Int J Behav Nutr Phys Act. 2010;7:59. http://www.ncbi.nlm.nih.gov/ pmc/articles/PMC2918527/. Accessed January 5, 2012. Mota J, Santos P, Guerra S, Ribeiro J, Duarte J. Patterns of daily physical activity during school days in children and adolescents. Am J Hum Biol. 2003; 15:547-553. Sirard JR, Riner WF Jr, Mciver KL, Pate RR. Physical activity and active commuting to elementary school. Med Sci Sports Exerc. 2005;37:2062-2069. Roemmich J, Leonard E, Samina R. The neighborhood and home environments: disparate relationships with physical activity and sedentary behaviors in youth. Ann Behav Med. 2007;33:29-38. Lox C. The Psychology of Exercise: Integrating Theory and Practice. 2nd
34.
35.
36.
37.
38.
39.
40.
ed. Halcomb Scottsdale, AZ: Holcomb Hathaway Publishers; 2006; 114–115. Treuth MS, Catellier DJ, Schmitz KH, et al. Weekend and weekly patterns of physical activity in overweight and normal-weight adolescent girls. Obesity. 2007;15:1782-1788. Meizi H, Piche L, Beynon C, Harris S. Screen-related sedentary behaviors: children’s and parents’ attitudes, motivations, and practices. J Nutr Educ Behav. 2010;42:17-25. Ness AR, Leary SD, Mattocks C, et al. Objectively measured physical activity and fat mass in a large cohort of children. PLoS Med. 2007;4:e97. http:// www.plosmedicine.org/article/info% 3Adoi%2F10.1371%2Fjournal.pmed .0040097. Accessed January 5, 2012. Purslow LR, Hill C, Saxton J, Corder K, Wardel J. Differences in physical activity and sedentary time in relation to weight in 8-9 year old children. Int J Behav Nutr Phys Act. 2008;5:67. http://www.ncbi .nlm.nih.gov/pmc/articles/PMC2636837/. Accessed January 5, 2012. Stager M, Harvey R, Secic M, Camlin-Shingler K, Cromer B. Self-reported physical activity and bond mineral density in urban adolescent girls. J Pediatr Adolesc Gynecol. 2006; 19:17-22. Treuth MS, Hou N, Young DR, Maynard LM. Validity and reliability of the Fels Physical Activity Questionnaire of children. Med Sci Sports Exerc. 2005;37:488-495. Nickelson J, Roseman MG, Forthofer MS. Associations between parental limits, school vending machine purchases, and soft drink consumption among Kentucky middle school students. J Nutr Educ Behav. 2010;42: 115-122.