Copyright 2002 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2002
Physical activity (PA) declines during adolescence. There has been little research describing this decline or examining participation and nonparticipation in specific activities.
To describe the pattern of change in the number of physical activities, the time spent on specific activities, and the stability of participation and nonparticipation in specific activities during adolescence.
Design and Setting
A population-based 4-year longitudinal study of adolescents recruited from a single suburban school district near Pittsburgh, Pa.
A total of 782 adolescents, aged 12 to 15 years at baseline.
Main Outcome Measures
Physical activity was measured annually via questionnaire. Outcome measures include hours per week of PA, number of reported activities, and participation (yes or no) in specific activities.
Physical activity declined during the 4 years by 26%. The decline in PA was primarily due to a decrease in the number of reported activities. Adolescents who continued to report an activity during the 4 years of the study maintained or increased the time spent on that specific activity. Female adolescents were more likely to report individual activities, while male adolescents were more likely to report team activities. The probability of maintaining participation in a specific activity during the 4 years was low to moderate, 0.02 to 0.47 for female adolescents and 0.04 to 0.71 for male adolescents. The probability of not participating in a specific activity during the 4 years was extremely high and consistent for male and female adolescents, 0.70 to 1.00.
The decline in PA during adolescence is primarily due to a decrease in the number of activities in which the adolescent is participating, and there is only a moderate probability that an adolescent will continue to participate in an activity during the 4-year period from junior to senior high. Future efforts should be directed at identifying factors associated with initiating and maintaining participation in specific activities.
PHYSICAL ACTIVITY (PA) is a major determinant of morbidity and mortality,1 and has been identified as a national priority area for promoting the health of the US population.2 However, given the documented benefits of achieving and maintaining an active lifestyle, many Americans, including children and adolescents, do not meet the recommendations for PA.1,2 The adolescent years are thought to be the period during which adult health behaviors, such as dietary and PA patterns, begin to develop. In addition, the movement out of high school is characterized by many life transitions that may influence health behaviors3; thus, childhood may be a critical time for promoting PA.
Longitudinal studies have documented an age-related decline of 26% to 37% in total PA during adolescence,4- 9 with additional decreases through young adulthood.6,7 While these studies provide needed information regarding the decrease in total PA levels during adolescence, they provide little information regarding what constitutes this change. The decrease in total PA may be due to decreases in the number of activities, the time spent on each activity, or a combination of both. The questions of what constitutes the decrease in total PA and what activities adolescents are likely to continue to participate in as they move through adolescence and young adulthood need to be explored. This investigation describes the pattern of change in the number of reported activities and the time spent on specific activities, and the stability of participation and nonparticipation in specific activities during a 4-year period, in a sample of adolescents aged 12 to 15 years at baseline.
The data in this investigation were obtained as part of the Adolescent Injury Control Study.10,11 Students in the seventh, eighth, and ninth grades from a single suburban school district were recruited to participate in a 4-year longitudinal study of the contribution of sports and recreational PA to all-cause injury incidence in adolescents. A total of 1245 (89%) of the eligible students agreed to participate in the study. The study was approved by the University of Pittsburgh Institutional Review Board. Informed consent was obtained from all students and parents before enrollment into the study.
The study group was composed of similar numbers of male (n = 641) and female (n = 604) adolescents; and the racial composition was 73% white, 24% African American, and 3% Hispanic or Asian. The socioeconomic status (SES) of the student was determined by first identifying the percentage of the population below the poverty level in the township or borough where the student resided. Students were classified as having a low SES if they resided in a township or borough where more than 20% of the population was below the poverty level. Likewise, students were classified as having a middle and high SES if they resided in a township or borough where 10% to 20% and less than 10% of the population were below the poverty level, respectively. From the original cohort, 410 adolescent boys and 372 adolescent girls (63%) had complete PA data for all 4 years of the study (January 1990–May 1993) and constitute the sample for this analysis. By using baseline data, a comparison was made between those who completed all 4 years of the study and those who did not to assess possible follow-up bias. There was no difference in the age or baseline leisure time PA between those who completed and those who did not complete the study. However, those not completing the 4-year study were more likely to be minority students and to have a low SES (P = .000 for both).
The Modifiable Activity Questionnaire for Adolescents12 was used to assess past year participation in leisure time and competitive activities during each of the 4 years of the study. This questionnaire has yielded reproducible and valid estimates of past year PA in adolescents.13,14 There is potential recall bias associated with the method of assessing PA used for this study. However, there was no indication that this bias was systematically different among subgroups of the population. In a random sample, the year-to-year reproducibility of the questionnaire was 0.54 for female adolescents and 0.65 for male adolescents.14 In addition, the year-to-year correlation was consistently in the range of 0.55 to 0.67, indicating that the reproducibility did not change as the cohort aged.
The questionnaire was administered to all students during the spring of each year. Students completed the questionnaire during their regular scheduled physical education class under the supervision of trained research assistants (D.J.A. and others). The questionnaire included a menu of 26 common recreational and leisure time activities. Students were asked to indicate all activities that they participated in at least 10 times during the past year. The students were also given the opportunity to add activities that were not listed on the questionnaire. Detailed information was collected regarding the frequency and duration of participation in each activity during the past year. An estimate of the average number of hours spent on each activity was calculated, and the hours from all activities were summed to derive an overall leisure time PA estimate (hours per week) averaged over the past year.12 In addition to the stability of participation in specific activities (eg, basketball and bicycling), we were also interested in the stability of participation in different types of activities, based on the following comparisons: team sports (eg, football and soccer) vs individual sports (eg, aerobics and tennis); light- and moderate-intensity activities (eg, bowling and volleyball) vs vigorous-intensity activities (eg, running and soccer); and year-round activities (eg, aerobics and weight lifting) vs seasonal activities (eg, baseball and softball).
The distributions of the hours per week of PA (total and for specific activities) were positively skewed. Nonparametric tests for paired and multiple sample data were used to examine changes over time, and nonparametric tests for unpaired data were used to examine sex differences in the hours per week of PA.
The number of reported activities was approximated to a normal distribution. As such, a repeated-measures analysis of variance was used to examine changes in the number of reported activities over the 4 measurements, and linear regression analysis was used to examine potential predictors of the number of activities during late adolescence while controlling for sociodemographic characteristics.
Participation in specific activities (yes or no) and types of activities (yes or no) was examined for sex differences using χ2 analyses. The stability of participation and nonparticipation in specific activities and activity types was evaluated by calculating the percentage of individuals who reported an activity or activity type in year 1 and also reported the activity or activity type in year 4 (predictive value positive [PV+]) and the percentage of individuals who did not report an activity or activity type in year 1 and also did not report the activity or activity type in year 4 (predictive value negative [PV−]).
The sample consisted of 410 male and 372 female adolescents enrolled in seventh (34%), eighth (30%), and ninth (35%) grade at baseline (percentages do not total 100 because of rounding). Eighty percent of the cohort was white, and the predominant racial minority was African American. Total PA (hours per week) for the entire sample declined by 26% during the 4-year period (P = .000). Physical activity decreased by 43% in male adolescents and by 26% in female adolescents (Figure 1); however, at each year, male adolescents were significantly more active than female adolescents (P = .000).
Median hours per week of total leisure time physical activity (PA) from spring 1990 to spring 1993.
To further describe the decrease in total PA, the number of activities reported each year was examined, as were changes in the amount of time spent on specific activities. The mean number of reported activities declined by 56% from 7.05 in year 1 to 3.08 in year 4 (P = .000). A repeated-measures analysis of variance showed a significant difference in the number of reported activities during the 4 years. Significant differences were also found by sex and baseline grade (P = .000 for both), with no sex-grade interaction (Figure 2). Overall, 85% of the students reported a decrease in the number of activities and only 5% reported an increase in the number of activities during the 4-year observation period. The mean (±SD) change in the number of activities was −3.97 (±2.91). Multiple linear regression analysis indicated that the number of activities at baseline was the strongest predictor of the number of activities in year 4 for adolescent boys and girls (P = .000). In adolescent boys, fewer activities in year 4 was related to a low SES (P = .000), and in adolescent girls, being African American (P = .000) and an older baseline age (P = .03) were related to fewer activities in year 4.
Decline in mean number of reported physical activities, from spring 1990 to spring 1993, by sex and grade at baseline. A, Male adolescents. B, Female adolescents.
To examine if a portion of the decrease in total PA could also be attributed to decreased time spent on specific activities, the median hours per week for the most common activities (n = 8 for males and n = 5 for females) were compared between year 1 and year 4. Among those who reported the same activity in 1990 and 1993, the time spent on that activity did not significantly (range of P values for 12 activities, .06-.48) change during the 4 years, with the exception of a significant increase in the hours per week that male adolescents participated in basketball (P = .04). This suggests that the decrease in total PA is primarily related to a decrease in the number of activities rather than decreased time spent on specific activities.
The top 10 activities reported by male and female adolescents indicate sex and age preferences (Table 1). In 1990 and 1993, male adolescents were significantly more likely to report participating in baseball, basketball, football, street hockey, and weight lifting, while female adolescents were significantly more likely to report participating in aerobics and softball. Sex differences in 1990 for bowling, soccer, tennis, and roller-skating were not evident in 1993. The participation rates in most sports decreased during the 4 years, with the exception of increased participation in aerobics and weight lifting among female adolescents and in weight lifting, roller-skating, and softball among male adolescents. The most dramatic decrease in participation was for bicycling.
To examine the stability of participation and nonparticipation in specific activities, the PV+ and the PV− were determined for activities based on the participation (yes or no) in year 1 of the study predicting participation (yes or no) in year 4. These calculations determined the probability that a student would report either participating or not participating in a specific activity in year 4 compared with year 1. For male and female adolescents, the PV+ for specific activities was low to moderate, while the PV− was high (Table 2). The PV+ for specific activities in female adolescents ranged from a low of 0.02 (bicycling) to a high of 0.47 (aerobics), indicating that there is less than a 50% chance a female adolescent will continue to participate in a specific activity during the period spanning junior to senior high school. The PV+ for specific activities in male adolescents were similar to those of the female adolescents, with the exception of higher values for basketball and weight lifting (0.71 and 0.63, respectively). In contrast, the PV− was extremely high and consistent for male and female adolescents (0.70-1.00), indicating that if adolescents are not participating in a specific activity by the time they are in junior high school there is a high probability that they will not participate in that activity during senior high school years.
The PV+ and PV− for types of activities further illustrate sex differences in activity choices. The PV+ for male adolescents was higher for team sports compared with individual sports, while in female adolescents, it was higher for individual sports. The PV+ for male and female adolescents was higher for year-round activities compared with seasonal activities and for vigorous-intensity activities vs light- and moderate-intensity activities. For adolescent boys and girls, the ranges of PV− were fairly similar, with the highest values found for team sports and seasonal activities for adolescent boys and for seasonal activities for adolescent girls.
This report provides information on the pattern of change in PA from mid to late adolescence. Several longitudinal studies4- 9 have reported that the total amount of PA decreases during the adolescent years. The decline in PA may occur as a result of decreases in the number of activities in which the adolescent participates, decreases in the time spent on each activity, or a combination of both. Few studies4,15 have attempted to describe the changing pattern of PA as a function of time during the adolescent years, and none, to our knowledge, have detailed the stability of participation and nonparticipation in specific activities. Understanding what constitutes the decline in PA among adolescents has important public health implications for developing programs to promote PA in young people.
The pattern of PA in this adolescent cohort was characterized by changes and consistencies. The decrease in total amount of PA observed in this adolescent sample seems to be primarily a function of a decrease in the number of activities rather than a decrease in the time spent on specific activities. Dovey et al4 reported similar decreases in the number of activities reported by adolescents, with 15-year-old adolescents reporting a median of 7 activities and at the age of 18 years reporting a median of 3 activities. In addition, the present study indicates that if an adolescent continues to participate in an activity from junior high through the senior high years, the amount of time spent on that activity will remain constant or will increase. The implication of this finding is important with respect to promoting PA in adolescents and designing physical education curricula. It may be critical to have preadolescent children maximize their exposure to various activities at a young age to enhance the likelihood that they will maintain participation in some of these activities in later years. The participation in a wide variety of activities is also supported by the low to moderate PV+ and the high PV− determined for the specific activities. With less than a 50% chance that an adolescent will continue to participate in an activity during a 4-year period, but an 80% to 90% chance that an adolescent will continue to not participate in an activity during the same 4-year period, it seems prudent to have children participate in various activities during their childhood years rather than specializing in 1 or even 2 activities in which they will probably not continue to participate.
Overall, there was consistency in the most popular activities reported. However, an age preference was noted for bicycling, with the greater than 50% participation rates among adolescent boys and girls in year 1 decreasing to less than 5% by year 4. There was also a consistent sex difference in the activities reported, with some activities undertaken exclusively by male adolescents (eg, baseball) and others undertaken exclusively by female adolescents (eg, aerobics). In addition, female adolescents reported more individual activities and had a higher PV+ for individual activities compared with team activities, while male adolescents reported more team activities and had a higher PV+ for team activities compared with individual activities. These findings are in agreement with those of previous studies16,17 in which a sex preference was reported in relation to choice of activities.
In addition to continued efforts to promote PA in sedentary adolescents and adults, future research should focus on factors associated with attrition from specific activities to determine why adolescents discontinue an activity. Although only 0.2% of the adolescents in the present study reported participating in 0 leisure time activities at study enrollment, most activities had a 3-year attrition rate greater than 50%. An extensive literature review18 of the determinants of PA in children and adolescents identified several major categories of determinants, including sociodemographic, psychological, and environmental factors. However, little research has focused on factors associated with discontinuing or changing patterns of participation in specific activities.
In a qualitative analysis of PA from childhood to young adulthood,19 the primary themes influencing male activity were significant others, size and maturation, and perception of ability, while the primary themes influencing female activity were transitions, body image concerns, and significant others. According to Weiss,20 the 3 major reasons children participate in activities are perceptions of physical competence, enjoyment of the activity, and social support from parents, teachers/coaches, and peers. The opposites of these could be reasons for discontinuing participation in a particular activity, such as low skill level and lack of success, enjoyment, and support; reasons for discontinuation also include competing time factors, such as work or school. DuRant et al21 reported attrition rates of up to 26% for adolescents participating in school-sponsored sports. While injury was the leading reason for discontinuing participation (25%), 13% quit to get a job, 11% reported inconvenience associated with the game or practice schedule, 10% needed more time to study, 8% were cut because of poor grades, and 8% were not given enough playing time.
Data from the present study indicate that if adolescents continue to participate in an activity, the time spent on that activity does not decrease during a 4-year period spanning early to late adolescence. Thus, to attenuate the decline in activity, it is important to understand why adolescents stop participating in an activity and what factors are competing for their time. Elucidating this information may help to guide the development of effective interventions and strategies to prevent this attrition. Physical education programs and community recreation organizations should be encouraged to offer various activities to young children across all skill levels. The exposure to many activities may ensure that children will identify an activity that they enjoy, will have some level of success in, and can continue to participate in during their late adolescent years when competing time pressures may limit the number of activities in which they can participate. Participation in sports and recreational pursuits is voluntary. If children are not given opportunities to experience a wide range of activities and identify an activity or activities that they have access to, enjoy, and have some level of success in, they will most likely choose to be inactive. Schools should offer alternative opportunities for activity participation beyond interscholastic competition. In a recent review of tracking of PA,22 participation in competitive sports was identified as yielding the highest tracking correlations among a cohort of Finnish adolescents. However, there are limited opportunities in the United States for participating in competitive youth sports for those who are less skilled.22 There also seems to be little opportunity for children and adolescents to become exposed to different activities within the school setting, as only 49% of students in grades 9 through 12 are enrolled in physical education classes and only 27% attend physical education classes daily.1 A recent survey23 of the status of physical education in the United States reported that only Illinois and Alabama require daily physical education for all students in kindergarten through grade 8. In addition, across the United States, state-mandated requirements for physical education at the elementary school level ranged from 30 to 150 min/wk. Thus, at an organizational level, schools and communities are contributing to the decline in the number of activities participated in by adolescents and are promoting physical inactivity by (1) the systematic exclusion of adolescents who are less skilled and (2) the failure to provide sufficient opportunities for children to become exposed to various activities, in particular, individual or lifetime activities.
The results of the present study indicate that the decrease in total PA from mid to late adolescence was due to a decrease in the number of activities in which adolescents choose to participate in rather than a decrease in the time spent on each activity. Adolescents who continue to participate in a specific activity from mid to late adolescence maintain or increase their level of participation. In addition, participation in any one activity has little stability over time, while the pattern of nonparticipation seems to be highly stable. Further investigation is needed to determine the factors associated with initiating and maintaining participation in specific activities during the adolescent years.
Accepted for publication June 6, 2002.
This study was supported by grant AR39541 from the National Institute of Arthritis and Musculoskeletal and Skin Diseases, Bethesda, Md; and grant HD35607 from the National Institute of Child Health and Human Development, Bethesda.
This study was presented in part at the Annual Conference of the American College of Sports Medicine, Baltimore, Md, June 1, 2001.
Corresponding author and reprints: Deborah J. Aaron, PhD, Department of Health and Physical Education, School of Education, University of Pittsburgh, 155 Trees Hall, Pittsburgh, PA 15261 (e-mail: firstname.lastname@example.org).
Longitudinal studies have documented a decline in PA through the adolescent years. The decline in total PA may be a result of decreasing number of activities, decreasing time spent on some activities, or a combination of both. Elucidating the nature of the decline in total activity and the stability of participation in specific activities is critical for developing effective intervention programs for children and adolescents.
The results of this study indicate that the decline in total PA is a function of decreases in the number of activities the adolescents are engaging in rather than a decrease in time spent on specific activities. The probability that an adolescent continued to participate in an activity during a 4-year period was low to moderate, while the probability of continued nonparticipation was high.
Aaron DJ, Storti KL, Robertson RJ, Kriska AM, LaPorte RE. Longitudinal Study of the Number and Choice of Leisure Time Physical Activities From Mid to Late AdolescenceImplications for School Curricula and Community Recreation Programs. Arch Pediatr Adolesc Med. 2002;156(11):1075-1080. doi:10.1001/archpedi.156.11.1075