Content uploaded by Barry M Popkin
Author content
All content in this area was uploaded by Barry M Popkin on Oct 18, 2018
Content may be subject to copyright.
301
Childhood and adolescent overweight
is a major public health problem for
American youth, particularly because
of its rapid increase in prevalence,1
persistence into adulthood,2,3 associat-
ed health consequences,4and morbidi-
ty and mortality.5
Inactivity is an important and modi-
fiable determinant of overweight. Inac-
tivity, in particular TV viewing, has
been associated with obesity in cross-
sectional studies of children, adoles-
cents, and adults.6Inactivity has been
found to show better tracking into
adulthood than physical activity.7Re-
ducing physical inactivity can increase
physical activity and have a positive
impact on obesity reduction.8,9
Physical activity, an important bio-
logic determinant of overweight, has
been associated with a wide range of
beneficial health outcomes in adults,
such as those related to bone health,
cardiovascular disease, and selected
cancers.10 Physical activity during
childhood and adolescence may have a
positive impact on growth and devel-
opment11 and psychologic and emo-
tional outcomes,12 which may continue
into adulthood.13 There is evidence
that physical activity tracks signifi-
cantly from adolescence to young
adulthood.7Studies have shown that
A
Adolescent physical activity and inactivity vary by
ethnicity: The National Longitudinal Study of
Adolescent Health
Penny Gordon-Larsen, PhD, Robert G. McMurray, PhD, and Barry M. Popkin, PhD
From the Carolina Population Center, University of North Carolina, Chapel Hill; Department of Exercise Physiolo-
gy, University of North Carolina, Chapel Hill; and Carolina Population Center and the Department of Nutrition,
School of Public Health, University of North Carolina, Chapel Hill.
Supported in part by National Institute of Child Health and Human Development grant P01-
HD31921 and the Dannon Institute Postdoctoral Fellowship in Interdisciplinary Nutrition Science.
Submitted for publication Jan 12, 1999; revision received Apr 12, 1999; accepted June 9, 1999.
Reprint requests: Penny Gordon-Larsen, PhD, University of North Carolina at Chapel Hill, Car-
olina Population Center, University Square, 123 West Franklin St, Chapel Hill, NC 27516-3997.
Copyright © 1999 by Mosby, Inc.
0022-3476/99/$8.00 + 0 9/21/100854
BMI Body mass index
METs Metabolic equivalents
NHANES National Health and Nutrition
Examination Survey
Add Health National Longitudinal Study of
Adolescent Health
RR Risk ratio
Objectives: To determine the extent to which physical activity and inactivi-
ty patterns vary by ethnicity among subpopulations of US adolescents.
Study design: Nationally representative data from the 1996 National Lon-
gitudinal Study of Adolescent Health of >14,000 US adolescents (including
3135 non-Hispanic blacks, 2446 Hispanics, and 976 Asians).
Methods: Hours per week of inactivity (TV viewing, playing video or com-
puter games) and times per week of moderate to vigorous physical activity
were collected by using questionnaire data. Multinomial logistic regression
models of physical activity and inactivity were used to adjust for sociodemo-
graphic factors.
Results: Large ethnic differences are seen for inactivity, particularly for
hours of television or video viewing per week (non-Hispanic blacks, mean =
20.4; non-Hispanic whites, mean = 13.1). Physical activity (≥5 bouts of
moderate to vigorous physical activity per week, 5-8 metabolic equivalents)
is lowest for female and minority adolescents. Ethnic differences are far
greater for inactivity than for moderate to vigorous physical activity.
Conclusion: Minority adolescents, with the exception of Asian females,
have consistently higher levels of inactivity. Results vary by sex; males have
higher inactivity and physical activity, whereas lowest physical activity is
found for non-Hispanic black and Asian females, although Asian females
also have low inactivity and low levels of overweight. Overall, efforts to
reduce the problem of adolescent overweight should focus on increasing
activity levels of adolescents, particularly female, older, and major minority
subpopulations. (J Pediatr 1999;135:301-6)
GORDON-LARSEN,MCMURRAY,AND POPKIN THE JOURNAL OF PEDIATRICS
SEPTEMBER 1999
vigorous physical activity is lower for
minority adolescents,14,15 and inactivi-
ty higher, relative to their non-Hispan-
ic white counterparts.15-17
Surprisingly few details on the types
of activity and inactivity patterns for
larger ethnic/racial groups (eg, Asians)
and ethnic subgroups (eg, Chinese
Americans) are known. Furthermore,
little is known about a range of inactivi-
ties, other than TV viewing, among
these ethnic subpopulation groups. In
addition, population-based studies on
physical activity have not provided in-
formation on adolescent sociodemo-
graphic factors that would allow adjust-
ment for a number of obesity-related
correlates.
METHODS
Sample Design
The study population consists of over
20,000 adolescents enrolled in the Na-
tional Longitudinal Study of Adolescent
Health, a nationally representative,
school-based sample of adolescents in
grades 7 to 12 in the United States. The
collection of data followed informed
consent procedures established by the
institutional review board of the Univer-
sity of North Carolina at Chapel Hill.
Add Health was longitudinal; the
first wave of data (wave 1) was ob-
tained between April and December
1995, and the second wave (wave 2)
was collected after a 1-year interval.
Add Health includes a core sample and
additional subsamples of selected eth-
nic and other groupings. This analysis
focuses primarily on the wave 2 sam-
ple (14,438 eligible adolescents who
were still enrolled in high school dur-
ing 1996, including dropouts [88% fol-
low-up rate]). Pregnant females are
excluded from the analyses, as are dis-
abled adolescents who used a walking
aid device (eg, cane, crutches, or
wheelchair). American Indians are ex-
cluded from the present analysis be-
cause of small sample size (n = 173).
Our final analysis sample totals 13,157
for prevalence estimates and multino-
mial logistic regression analyses. The
survey design and sampling frame
have been described elsewhere.18
Anthropometry
Height and weight were measured
according to standardized procedures
and protocol with subjects wearing
street clothes without shoes. Body
mass index is used as an index of ado-
lescent overweight.19-21 Overweight
status was determined by comparing
the adolescents’ BMI to age- and sex-
appropriate reference data from Na-
tional Health and Nutrition Examina-
tion Survey I21,22 for BMI ≥85th and
95th percentiles.
Physical Activity
Physical activity was assessed by
using standard questionnaire method-
ology23-25 with an array of questions
similar to those validated in many other
smaller studies on physical activity to
categorize adolescents into high, medi-
um, and low activity patterns with rea-
sonable reliability and validity. During
wave 2, each adolescent was asked
about the times per week spent en-
gaged in various physical activities.
Each activity grouping was assigned a
metabolic equivalent value (1 MET =
3.5 mL O2/kg body weight/min or rest-
ing metabolic rate) based on the Com-
pendium of Physical Activity26 devel-
oped for adults to categorize activity as
low, moderate, or vigorous. It is known
that the energy cost of activities is
about 10% higher in children; however,
at present, no norms exist for children.
Higher intensity activities such as skat-
ing, cycling, dance, martial arts activi-
ties, and active sports were assigned 5
to 8 METs and are thus considered
moderate to vigorous physical activity.
Physical Inactivity
Physical inactivity was assessed by
using a standard 7-day recall question-
naire to categorize adolescents into
high, medium, and low inactivity pat-
terns. The inactivity questions follow
the same structure as those validated
for physical activity. The quantifica-
tion of inactivity has received far less
attention than physical activity,27 and
there is little published in the literature
regarding reliability and validity of
measures of inactivity. In addition to
the standard TV viewing question used
in other studies, the number of hours
and minutes of video viewing and
video or computer game use during the
previous week were assessed. The
302
Figure. Percent of adolescents with BMI ≥85th and 95th percentiles NHANES I22 by sex and eth-
nic group, weighted to be nationally representative with the error terms corrected for design effects.
THE JOURNAL OF PEDIATRICS GORDON-LARSEN,MCMURRAY,AND POPKIN
VOLUME 135, NUMBER 3
quantification of inactivity is more
straightforward than that of physical
activity, because there is no concern
with intensity and most sedentary ac-
tivities have the same relative energy
cost. In this study all 3 inactivity vari-
ables were assigned 1 MET each. A
composite inactivity score was calcu-
lated by using the number of hours and
minutes that each adolescent spent en-
gaged in TV and video viewing and
playing video or computer games.
Other Measures
Add Health collected separate self-
reported race (census categorization:
white, black, Asian/Pacific Islander,
Native American) and ethnic identifi-
cation (eg, Hispanic origin, Cuban
background) from each adolescent and
from parents. Ethnic groups include
the following: non-Hispanic whites,
non-Hispanic blacks, Hispanics
(Cuban, Puerto Rican, Central/South
American, Mexican/Chicano, and
Other Hispanic), Asians (Chinese, Fil-
ipino, and Other Asians). Age was de-
termined at time of survey.
Statistical Analysis
Statistical analyses were carried out
with STATA. 28 Post-stratification sam-
ple weights were used in all analyses to
allow these results to be comparable
with the US population. Because the
Add Health survey is based on com-
plex sample designs, design effects of
multiple stages of sampling were also
controlled in all analyses. Multinomial
logistic regression models of physical
activity and inactivity were used to in-
vestigate sex and ethnic interactions
controlling for age, urban residence,
education of mother, family income,
ethnicity, generation of residence in
the United States, in-school status, and
month of interview.
RESULTS
Overweight Prevalence
Overweight prevalence (BMI ≥85th/
95th percentiles, Figure) is highest
among non-Hispanic black females
(39.4/19.8%) and Hispanic males (30.2/
12.6%) and females (29.8/13.5%) and
lowest for Asian females (11.1/4.2%).
Rates among the ethnic subpopula-
tions were high for Cuban females
(37.8/16.2%) and Puerto Rican males
(34.1/14.7%).
Inactivity
Mean hours of TV viewing per week
were highest for males and particularly
for non-Hispanic black males and fe-
males (Table I). Values were lowest for
non-Hispanic white males and females.
Composite inactivity hours, presented
in Table II, show similar patterns to
those shown for TV viewing. Compos-
ite inactivity was higher for males, and
this sex difference was greater than
that seen for TV viewing. Ethnic differ-
ences in weekly inactivity were highest
among non-Hispanic black males and
females and lowest for non-Hispanic
white males and females. TV viewing
and composite inactivity decreased
with age for both males and females.
Physical Activity
Percentages of adolescents reporting
total moderate to vigorous physical ac-
tivity (5-8 METs) are shown in Table
III. One third (33.2%) of the adoles-
cents in Add Health reported that they
participated in 5 or more episodes of
moderate to vigorous physical activity
per week; activity was higher for males
than females. Ethnic differences were
small among males. In comparison, eth-
nic differences were greater for females,
303
No. Males No. Females No. Total
Ethnicity
White 3,540 14.4 (.40) 3,665 11.9 (.43) 7,205 13.1 (.37)
Black 1,288 20.8 (.77) 1,482 20.0 (.89) 2,770 20.4 (.73)
Hispanic 1,139 16.6 (.64) 1,121 14.6 (.82) 2,260 15.6 (.55)
Asian 489 15.0 (1.02) 433 12.8 (1.09) 922 14.0 (.87)
Total 6,456 15.7 (.40) 6,701 13.5 (.47) 13,157 14.6 (.40)
Age category
12-15 y 1,985 17.1 (0.49) 2,334 14.7 (0.63) 4,319 15.8 (0.48)
16-17 y 2,803 15.2 (0.55) 2,901 12.6 (0.57) 5,704 13.9 (0.51)
18-22 y 1,668 13.9 (0.56) 1,466 12.9 (0.74) 3,134 13.4 (0.54)
*Weighted to be nationally representative with the error terms corrected for design effects.
Table I. Mean (SEM) hours of television per week*
No. Males No. Females No. Total
Ethnicity
White 3,540 22.1 (.58) 3,665 16.5 (.52) 7,205 19.3 (.50)
Black 1,288 31.8 (1.12) 1,482 27.6 (1.06) 2,770 29.7 (.85)
Hispanic 1,139 24.4 (.99) 1,121 19.9 (1.21) 2,260 22.2 (.86)
Asian 489 23.7 (1.45) 433 18.6 (1.67) 922 21.3 (1.20)
Total 6,456 24.0 (.59) 6,701 18.7 (.59) 13,157 21.4 (.54)
Age category
12-15 y 1,985 26.03 (0.74) 2,334 20.4 (0.78) 4,319 23.1 (0.64)
16-17 y 2,803 22.99 (0.78) 2,901 17.6 (0.72) 5,704 20.3 (0.69)
18-22 y 1,668 21.87 (0.84) 1,466 17.3 (0.93) 3,134 19.8 (0.73)
*Weighted to be nationally representative with the error terms corrected for design effects.
Table II. Mean (SEM) composite inactivity hours per week based on TV and video
viewing and computer and video game playing*
GORDON-LARSEN,MCMURRAY,AND POPKIN THE JOURNAL OF PEDIATRICS
SEPTEMBER 1999
with much higher levels of moderate to
vigorous physical activity for non-His-
panic white females than minority fe-
males. Approximately half of the non-
Hispanic black females and slightly
fewer Asian and Hispanic females re-
ported that they engaged in 2 or fewer
bouts of moderate to vigorous weekly
physical activity. Participation in mod-
erate to vigorous physical activity de-
creased substantially with age.
Multinomial Logistic Regression
Models
To study the distribution of inactivity
patterns by ethnicity and gender, per-
centages of adolescents predicted to
participate in low, medium, or high ter-
tiles of weekly inactivity (TV and video
viewing and video or computer game
use) are prepared from a multinomial
logistic regression model of weekly
composite inactivity adjusting for age,
urban residence, socioeconomic status,
ethnicity, in-school status, and month
of interview (Table IV). Risk ratios
were highest for non-Hispanic blacks
(1.78; 95% CI, 1.38-2.30; P≤.0001)
and Asians (1.81; 95% CI, 1.11-2.97; P
≤.02). Ethnicity and sex interactions
were significant for non-Hispanic black
females (RR, 1.56; 95% CI, 1.19-2.05;
P≤.001) and Hispanic females (RR,
1.43; 95% CI, 1.06-1.93; P≤.02). Inac-
tivity was highest for non-Hispanic
black males (48.3%) and females
(40.3%) and Asian males (45.3%) and
lowest for non-Hispanic white (21.7%)
and Hispanic females (27.7%).
For moderate to vigorous physical
activity, again adjusting for sex, age,
urban residence, socioeconomic status,
ethnicity, in-school status, and month
of interview (Table V), there were
more similarities across ethnic groups
and within sexes than for the inactivity
data. The likelihood of engaging in
moderate to vigorous physical activity
ranged from 37.9% (Asian males) to
42.2% (non-Hispanic black males). In
contrast, non-Hispanic white females
(25.6%) were most likely to engage in
moderate to vigorous physical activity,
and Asian females (12.5%) were least
likely to have high physical activity
levels. The ethnicity and sex interac-
tions were statistically significant for
non-Hispanic black females (RR, 0.43;
95% CI, 0.31-0.61; P≤.0001), Asian
females (RR, 0.41; 95% CI, 0.22-0.77;
P≤.006), and Hispanic females (RR,
0.62; 95% CI, 0.40-0.94; P≤.027).
In addition, inactivity and moderate to
vigorous physical activity were similarly
modeled (not shown here) for sub-
groups of Hispanics (Cuban, Puerto
Rican, Central/South American, Mexi-
can-Chicano, and Other Hispanic) and
Asians (Chinese, Filipino, and Other
Asian). Non-Hispanic blacks (44.3%),
Filipinos (43.1%), Other Asians
(38.8%), and Cubans (37.5%) were
most likely to be inactive, whereas
Other Hispanics (25.1%), non-Hispanic
whites (28.0%), and Mexican-Chicanos
(29.9%) were least likely to be inactive.
Participation in inactivity was signifi-
cant for Filipinos (RR, 2.68; 95% CI,
1.74-4.14; P≤.0001), Cubans (RR, 2.47;
95% CI, 1.24-4.90; P≤.01), non-His-
panic blacks (RR, 2.27; 95% CI, 1.81-
2.84; P≤.0001), Puerto Ricans (RR,
2.04; 95% CI, 1.25-3.32; P≤.005), and
Central/South Americans (RR, 1.55;
95% CI, 1.07-2.24; P≤.02). Ethnic dif-
ferences for physical activity were small.
DISCUSSION
Add Health provides the unique op-
portunity to examine patterns of over-
weight, physical activity, and inactivity
304
0-2 3-4 ≥5
Ethnic/age group No. times/wk times/wk times/wk
Non-Hispanic white—total 7,205 31.2 (1.2) 33.6 (.80) 35.2 (1.2)
Males 3,540 25.6 (1.3) 31.8 (1.1) 42.6 (1.6)
Females 3,665 37.2 (1.6) 35.3 (1.1) 27.6 (1.4)
Non-Hispanic black—total 2,770 36.7 (1.7) 34.8 (1.1) 28.5 (1.5)
Males 1,288 24.0 (1.5) 36.0 (1.3) 39.9 (2.1)
Females 1,482 49.5 (2.4) 33.5 (1.8) 17.1 (1.8)
Hispanic—total 2,260 33.9 (2.0) 36.3 (1.9) 29.8 (1.6)
Males 1,139 26.8 (2.3) 33.7 (2.2) 39.5 (2.0)
Females 1,121 41.6 (3.1) 38.9 (2.6) 19.5 (2.3)
Asian—total 922 32.7 (3.2) 37.7 (3.0) 29.6 (2.8)
Males 489 23.1 (3.0) 34.1 (3.0) 42.8 (3.9)
Females 433 44.1 (4.8) 41.6 (4.7) 14.3 (2.4)
12-15 y—total 4,319 20.9 (1.3) 35.6 (1.1) 43.5 (1.4)
Males 1,985 16.0 (1.2) 31.4 (1.5) 52.6 (1.8)
Females 2,334 25.5 (1.6) 39.5 (1.5) 35.0 (1.7)
16-17 y—total 5,704 37.2 (1.0) 33.5 (.86) 29.3 (.85)
Males 2,803 27.3 (1.3) 33.0 (1.3) 39.6 (1.6)
Females 2,901 47.0 (1.6) 33.9 (1.2) 19.1 (1.2)
18-22 y—total 3,134 47.6 (1.4) 32.9 (1.4) 19.5 (1.1)
Males 1,668 39.0 (1.6) 35.1 (1.7) 25.9 (1.7)
Females 1,466 57.7 (2.3) 30.3 (2.0) 11.9 (1.2)
Total sample 13,157 32.6 (1.0) 34.2 (.65) 33.2 (.99)
Males 6,456 25.4 (.97) 32.8 (.84) 41.8 (1.2)
Females 6,701 39.9 (1.4) 35.6 (.89) 24.5 (1.2)
Values are expressed as percents with SEMs in parentheses. Error terms are corrected for design
effects.
Table III. Participation in sessions of moderate-vigorous physical activity per week,
weighted to be nationally representative
THE JOURNAL OF PEDIATRICS GORDON-LARSEN,MCMURRAY,AND POPKIN
VOLUME 135, NUMBER 3
among large and nationally represen-
tative samples of ethnic subpopulation
groups (eg, Chinese, Puerto Ricans,
and Cubans). Other national surveys
have collected these data for selected
ethnic groups (mainly whites, Hispan-
ics or Mexican-Americans, and non-
Hispanic blacks or blacks) but lack the
depth of coverage and the detailed so-
ciodemographic data of Add Health.
As noted in the text, overweight lev-
els are very high for adolescents, par-
ticularly for non-Hispanic black fe-
males, Cuban females, and Puerto
Rican males. These results, which are
similar to those found in the NHANES
III survey, are of considerable concern
for health care professionals treating
American adolescents.1
The data presented here show that
activity and inactivity patterns differ
by ethnicity, with minority groups en-
gaging in less physical activity and
more inactivity than their non-Hispan-
ic white counterparts. The exception
was Asian females who have low levels
of physical activity, inactivity, and
overweight. This potential contradic-
tion might be the result of the later age
of maturation among Asian females or
reduced inactivity and hence more
light-moderate activity. Alternatively,
this contradiction might result from a
larger proportion of first-generation
Asian females who follow a healthier,
more traditional diet.18
The Add Health survey, with its
much larger sample and direct inter-
view of adolescents and their parents,
provides a more detailed set of activity,
inactivity, and sociodemographic data
than previously published studies. It
includes estimates of physical activity
in between those of the NHANES III
survey15 and the Youth Risk Behavior
Survey.14 Add Health found that
74.6% of males and 60.1% of females
were active 3 or more times per week.
The current public health recommen-
dation is that adolescents and youth en-
gage in at least 3 bouts of continuous
moderate to vigorous physical activity
per week, an accumulation of at least
30 minutes of daily moderate level
physical activity.29 One third of adoles-
cents in Add Health (25.4% males and
39.9% females) failed to achieve these
recommendations. Failure to achieve
these recommendations was highest for
non-Hispanic blacks (36.7%) and low-
est for non-Hispanic whites (31.2%).
In particular, a substantial proportion
of non-Hispanic black females
(49.5%), Asian females (44.1%), and
Hispanic females (41.6%) reported
zero to two sessions of moderate to vig-
orous activity per week. Therefore it
appears that minorities, specifically
Hispanics and non-Hispanic blacks, do
not get sufficient physical activity. In
addition, they spend more time watch-
ing TV and have higher overall inactiv-
ity levels than non-Hispanic whites.
Both physical activity and inactivity
decreased with age; however, the de-
crease in inactivity was not as dramatic
as that seen for physical activity. This is
comparable to other research.14,15
In Add Health both inactivity and
physical activity were considerably
higher for males than females. The inac-
tivity patterns for Add Health adoles-
cents are similar to those found for other
studies of US adolescents in which more
limited measures of inactivity were
used.15-17 As noted, key subpopulations
were much more inactive (eg, Non-His-
panic blacks, Hispanics, and Filipinos).
Overall, the results suggest that inactivi-
ty is a potentially more important area
305
Non-Hispanic white Non-Hispanic black Hispanic Asian
Inactivity level Males Females Males Females Males Females Males Females
Low 31.4 47.6 25.3 33.0 28.0 37.2 23.1 41.2
Medium 34.1 30.6 26.4 26.7 36.6 35.1 31.6 26.6
High 34.5 21.7 48.3 40.3 35.3 27.7 45.3 32.2
*Adjusted by using multinomial logistic regressions controlling for sex, age, urban residence, socioeconomic status, ethnicity, in-school status, and
month of interview.
Table IV. The adjusted proportion of adolescents participating in given tertiles of composite inactivity*
Non-Hispanic white Non-Hispanic black Hispanic Asian
Activity level Males Females Males Females Males Females Males Females
Low 27.2 39.5 22.8 48.9 26.1 43.3 25.3 44.5
Medium 35.0 34.9 35.1 33.1 33.7 38.5 36.8 43.0
High 42.1 25.6 42.2 18.1 40.1 18.1 37.9 12.5
*Adjusted by using multinomial logistic regressions controlling for sex, age, urban residence, socioeconomic status, ethnicity, in-school status, and
month of interview.
Table V. The adjusted proportion of adolescents participating in given tertiles of moderate to vigorous physical activity*
GORDON-LARSEN,MCMURRAY,AND POPKIN THE JOURNAL OF PEDIATRICS
SEPTEMBER 1999
for public health efforts aimed at Ameri-
can adolescents than physical activity.
Dietz27 states that inactivity is an under-
studied behavior amenable to cost-effec-
tive strategies for reducing obesity and
preventing cardiovascular disease. TV
viewing, the number one leisure activity
among children, has received the most
attention and has been negatively asso-
ciated with obesity, physical activity,
and fitness patterns.30,31 Add Health
shows that other components of inactiv-
ity, such as video viewing and computer
or video game playing, are also impor-
tant. It is important that public health ef-
forts be aimed at promoting physical ac-
tivity and providing opportunities for
non-sedentary activities, particularly for
female, older, and minority adolescents.
REFERENCES
1. Troiano RP, Flegal KM, Kuczmarski
RJ, Campbell SM, Johnson CL.
Overweight prevalence and trends for
children and adolescents. Arch Pediatr
Adolesc Med 1995;149:1085-91.
2. Serdula MK, Ivery D, Freedman
RJC, Williamson DF, Byers T. Do
obese children become obese adults? A
review of the literature. Prev Med
1993;22:167-77.
3. Siervogel RM, Roche AF, Guo SM,
Mukherjee D, Chumlea WC. Patterns
of change in weight/stature 2 from 2 to
18 years: findings from long-term seri-
al data for children in the Fels Longi-
tudinal Growth Study. Int J Obes
1991;15:479-85.
4. National Institutes of Health. Physical
activity and cardiovascular health.
NIH Consensus Statement Online,
December 18-20. 1995;13:1-33.
5. Must A, Jacques PF, Dallal GE, Baje-
ma CJ, Dietz WH. Long-term mor-
bidity and mortality of overweight
adolescents: a follow-up of the Har-
vard Growth Study of 1922 to 1935. N
Engl J Med 1992;327:1350-5.
6. Gortmaker SL, Sobal AM, Peterson
K, Colditz CA, Dietz WH. Television
viewing as a cause of increasing obesi-
ty among children in the United States.
Arch Pediatr Adolesc Med 1996;150:
536-62.
7. Raitakari OT, Porkka KVK, Taimela S,
Telama R, Rasanen L, Viikari JSA. Ef-
fects of persistent physical activity and
inactivity on coronary risk factors in
children and young adults. The Cardio-
vascular Risk in Young Finns Study.
Am J Epidemiol 1994;140:195-205.
8. Epstein LH, Smith JA, Vara LS,
Rodefer JS. Behavioral economic
analysis of activity choice in obese chil-
dren. Health Psychol 1991;10:311-6.
9. Epstein LH, Saelens BE, O’Brien JG.
Effects of reinforcing increases in ac-
tive versus decreases in sedentary be-
havior in obese children. Int J Behav
Med 1995;2:41-50.
10. National Institutes of Health. Consen-
sus development conference on diet
and exercise in non-insulin dependent
diabetes mellitus. Diabetes Care 1987;
10:639-43.
11. US Department of Health and Human
Services. Physical activity and health:
a report of the Surgeon General. At-
lanta (GA): US Department of Health
and Human Services, Centers for Dis-
ease Control and Prevention, National
Center for Chronic Disease Prevention
and Health Promotion; 1996.
12. Martinek TJ, Cheffers JTF,
Zaichkowsky LD. Physical activity,
motor development and self concept:
race and age differences. Percept Mot
Skills 1978;46:147-54.
13. Ross CE, Hayes DE. Exercise and psy-
chologic well-being in the community.
Am J Epidemiol 1988;127:762-71.
14. Heath GW, Pratt M, Warren CW,
Kann L. Physical activity patterns in
American high school students. Re-
sults from the 1990 Youth Risk Behav-
ior Survey. Arch Pediatr Adolesc Med
1994;148:1131-6.
15. Andersen RE, Crespo CJ, Bartlett SJ,
Cheskin LJ, Pratt M. Relationship of
physical activity and television watching
with body weight and level of fatness
among children: results from the Third
National Health and Nutrition Exami-
nation Survey. JAMA 1998;279:938-42.
16. Wolf AM, Gortmaker SL, Cheung L,
Gray HM, Herzog DB, Colditz CA.
Activity, inactivity, and obesity: racial,
ethnic, and age differences among
schoolgirls. Am J Public Health 1993;
83:1625-7.
17. Sallis JF, Zakarian JM, Hovell MF,
Hofstetter CR. Ethnic, socioeconomic,
and sex differences in physical activity
among adolescents. J Clin Epidemiol
1996;49:125-34.
18. Popkin BM, Udry JR. Adolescent
obesity increases significantly in sec-
ond and third generation US immi-
grants: The National Health and Lon-
gitudinal Study of Adolescent Health.
J Nutr 1998;128:701-6.
19. Cole TJ. Weight-stature indices to mea-
sure underweight, overweight, and obe-
sity. In: Himes JH, editor. Anthropomet-
ric assessment of nutritional status. New
York: Wiley and Sons; 1991. p. 83-112.
20. Himes JH, Dietz WH. Guidelines for
overweight in adolescent preventive
services: recommendations from an ex-
pert committee. Am J Clin Nutr 1994;
59:307-16.
21. World Health Organization Expert
Committee. Physical status: the use
and interpretation of anthropometry.
Geneva (Switzerland): World Health
Organization; 1995. WHO Technical
Report Series. No. 854.
22. Must A, Dallal GE, Dietz WH. Refer-
ence data for obesity 85th and 95th
percentiles of body mass index
(wt/ht2) and triceps skinfold thick-
ness. Am J Clin Nutr 1991;53:839-46.
23. Baranowski T. Validity and reliability of
self-report measures of physical activity:
an information processing perspective.
Res Q Exerc Sport 1988;59:314-27.
24. Pate RR, Health GW, Dowda M, Trost
SG. Associations between physical activ-
ity and other health behaviors in a repre-
sentative sample of US adolescents. Am
J Public Health 1996;86:1577-81.
25. Sallis JF, Buono MJ, Roby JJ, Mi-
cale FG, Nelson JA. Seven-day recall
and other physical activity self-reports
in children and adolescents. Med Sci
Sports Exerc 1993;25:99-108.
26. Ainsworth B, Haskell WL, Leon AS,
Jacobs DR Jr, Montoye HJ, Sallis
JF, et al. Compendium of physical ac-
tivities: classification of energy costs of
human physical activities. Med Sci
Sports Exerc 1993;25:71-80.
27. Dietz WH. The role of lifestyle in
health: the epidemiology and conse-
quences of inactivity. Proc Nutr Soc
1996;55:829-40.
28. StataCorp. Stata Statistical Software:
Release 5.0. College Station (TX):
Stata Corporation; 1997.
29. Sallis JF, Patrick K. Physical activity
guidelines for adolescents: consensus
statement. Pediatr Exerc Sci 1994;6:
302-14.
30. Dietz WH, Strasburger VC. Children,
adolescents, and television. Curr Prob
Pediatr 1991;1:8-31.
31. DuRant RH, Baranowski T, Johnson
M, Thompson WO. The relationship
among television watching, physical ac-
tivity, and body composition of young
children. Pediatrics 1994;94:449-55.
306