Amount of time spent in sedentary behaviors in the United States, 2003-2004.

Abstract

Sedentary behaviors are linked to adverse health outcomes, but the total amount of time spent in these behaviors in the United States has not been objectively quantified. The authors evaluated participants from the 2003-2004 National Health and Nutrition Examination Survey aged >/=6 years who wore an activity monitor for up to 7 days. Among 6,329 participants with at least one 10-hour day of monitor wear, the average monitor-wearing time was 13.9 hours/day (standard deviation, 1.9). Overall, participants spent 54.9% of their monitored time, or 7.7 hours/day, in sedentary behaviors. The most sedentary groups in the United States were older adolescents and adults aged >/=60 years, and they spent about 60% of their waking time in sedentary pursuits. Females were more sedentary than males before age 30 years, but this pattern was reversed after age 60 years. Mexican-American adults were significantly less sedentary than other US adults, and White and Black females were similarly sedentary after age 12 years. These data provide the first objective measure of the amount of time spent in sedentary behavior in the US population and indicate that Americans spend the majority of their time in behaviors that expend very little energy.

Full text

Page 1

Original Contribution
Amount of Time Spent in Sedentary Behaviors in the United States, 2003–2004
Charles E. Matthews1, Kong Y. Chen2, Patty S. Freedson3, Maciej S. Buchowski4, Bettina M.
Beech1, Russell R. Pate5, and Richard P. Troiano6
1Institute for Medicine and Public Health, Department of Medicine, School of Medicine, Vanderbilt University, Nashville, TN.
2Metabolic Research Core, Division of Intramural Research, National Institute of Diabetes and Digestive and Kidney Diseases,
Bethesda, MD.
3Department of Kinesiology, School of Public Health and Health Sciences, University of Massachusetts, Amherst, MA.
4Division of Gastroenterology, Hepatology and Nutrition, Department of Medicine, School of Medicine, Vanderbilt University,
Nashville, TN.
5Department of Exercise Science, Norman J. Arnold School of Public Health, University of South Carolina, Columbia, SC.
6Risk Factor Monitoring and Methods Branch, National Cancer Institute, Bethesda, MD.
Received for publication August 20, 2007; accepted for publication December 11, 2007.
Sedentary behaviors are linked to adverse health outcomes, but the total amount of time spent in these behav-
iors in the United States has not been objectively quantified. The authors evaluated participants from the 2003–
2004 National Health and Nutrition Examination Survey aged ?6 years who wore an activity monitor for up to
7 days. Among 6,329 participants with at least one 10-hour day of monitor wear, the average monitor-wearing time
was 13.9 hours/day (standard deviation, 1.9). Overall, participants spent 54.9% of their monitored time, or 7.7
hours/day, in sedentary behaviors. The most sedentary groups in the United States were older adolescents and
adults aged ?60 years, and they spent about 60% of their waking time in sedentary pursuits. Females were more
sedentary than males before age 30 years, but this pattern was reversed after age 60 years. Mexican-American
adults were significantly less sedentary than other US adults, and White and Black females were similarly sed-
entary after age 12 years. These data provide the first objective measure of the amount of time spent in sedentary
behavior in the US population and indicate that Americans spend the majority of their time in behaviors that expend
very little energy.
energy metabolism; monitoring, ambulatory; motor activity; obesity; population surveillance
Abbreviations: IDEEA, Intelligent Device for Energy Expenditure and Activity; MET, metabolic equivalent; NHANES, National
Health and Nutrition Examination Survey; SD, standard deviation.
The amount of time spent in sedentary behaviors has been
independently associated with lower levels of physical-
activity energy expenditure (1, 2), increased risk of weight
gain (2, 3), and increased risk of metabolic syndrome, di-
abetes, and heart disease (3–5). In light of these links to
adverse health outcomes and the continued increase in the
prevalence of overweight and obesity in the United States
(6), sedentary behaviors have emerged as an important tar-
get of health promotion and obesity and disease prevention
efforts (7–9), complementing efforts to increase levels of
moderate-to-vigorous physical activity (10–12). Sedentary
behaviors have been defined as a range of human endeavors
that result in an energy expenditure of no more than 1.5
times’ resting energy expenditure (8). These behaviors are
typically associated with time spent sitting, reclining, or
lying down during waking hours (7, 8).
Correspondence to Dr. Charles E. Matthews, Institute for Medicine and Public Health, Vanderbilt University Medical Center, 2525 West End
Avenue, Suite 600, Nashville, TN 37232-1738 (e-mail: charles.matthews@vanderbilt.edu).
1
American Journal of Epidemiology
ª The Author 2008. Published by the Johns Hopkins Bloomberg School of Public Health.
All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org.
DOI: 10.1093/aje/kwm390

American Journal of Epidemiology Advance Access published February 25, 2008

Page 2

Current estimates of amounts of time spent in sedentary
behaviors in the United States have been derived from self-
reportsofalimitednumberofdiscretionarybehaviorsthought
to be indicative of a more sedentary lifestyle (e.g., television
viewing, video game-playing, and computer use). However,
assessment of these behaviors provides only a partial picture
of overall levels of sedentary behavior in a typical waking
day. For example, the American Time Use Survey, a popula-
tion-based survey of adolescents and adults aged ?15 years,
indicated that respondents reported an average of 2.6 hours/
day of televisionviewing and less than 0.5 hours/day of com-
puter use at home (13), which represents only a modest pro-
portion of the average 15.4-hour waking day in this sample.
Similarly, 80 percent of adults in the United States reported
only three or fewer hours of television viewing or non-work-
related computer use per day (4). The average amount oftime
spent in overall sedentary behaviors during waking hours in
the United States is unknown and may be substantially under-
estimated by surveillance systems that have been used to
assess only a limited number of discretionary behaviors.
Incorporation of an objective measurement of physical
activity in the National Health and Nutrition Examination
Survey (NHANES) enabled the direct measurement of
time spent in sedentary behaviors in a representative sample
of youth and adults in the United States. Accordingly, the
purposeofthisinvestigationwastodescribe,forthefirsttime,
the amount of time spent inoverall sedentary behaviors in the
United States, by gender, age, and racial/ethnic group.
MATERIALS AND METHODS
Study population and design
NHANES 2003–2004 included a representativesample of
the US civilian, noninstitutionalized population, selected
with a complex multistage probability design. Participants
aged ?6 years who were seen in NHANES mobile exami-
nation centers were asked to wear an Actigraph accelerom-
eter (model 7164; Actigraph, LLC, Fort Walton Beach,
Florida) for 7 consecutive days.
The Actigraph is a small (2.0 3 1.6 3 0.6 inches (5.1 3
4.1 3 1.5 cm)), light (1.5 ounces (0.4 kg)) instrument that
records integrated acceleration information as an activity
count, which provides an objective estimate of the intensity
of bodily movement. The device was programmed to record
in 1-minute epochs. Participants were instructed to wear the
monitor on their right hip attached by an elastic belt during
all waking hours; however, because the monitors were not
waterproof, participants were asked to take them off while
bathing or swimming. After the monitors had been worn,
they were returned by mail to the NHANES contractor,
where data were downloaded and the instruments were
checked for calibration. Units that did not meet the manu-
facturer’s specifications for calibration (?5 percent) were
recalibrated or repaired prior to reuse.
The high sensitivity of the Actigraph results in wide
variability of activity count values over a range of human
ambulatory movements (e.g., walking at 2–3 miles/hour
(3.2–4.8 km/hour), about 1,000–3,000 counts/minute; run-
ning at 6–7 miles/hour (9.6–11.2 km/hour), about 10,000
counts/minute) (14). Activity counts recorded while sitting
and working at a desk are very low (?50 counts/minute),
and counts recorded while driving an automobile are
typically below 100 counts/minute (unpublished observa-
tions). The sensitivity and well-described measurement
properties of this device facilitated the development of
an automated approach to objective estimation of monitor
wear by scanning of the minute-by-minute data. Nonwear
time was defined as intervals of at least 60 consecutive
minutes of zero counts, with allowance for up to 2 consec-
utive minutes of observations of 1–100 counts/minute.
Periods of nonwear were defined as ending when count
levels exceeded 100 counts/minute or when 3 consecutive
minutes of observation were between 1 and 100 counts/
minute. Monitor-wearing time was estimated by subtracting
nonwearing time from the total observation time for the day.
We estimated time spent in sedentary behaviors as the
amountoftime accumulated below 100 counts/minute during
periodswhenthemonitorwasworn,expressedasaproportion
of monitor-wearing time (percent) and as total duration
(hours/day). This threshold was identified in a calibration
study carried out among adolescent girls (15) and was used
to identify an independent association between sedentary be-
havior and 2-hour postload glucose measures in adults (16).
In a small validation study in adults (n ¼ 19; 53 percent
female; mean age ¼ 40.1 years (standard deviation (SD),
10.0); body mass index (weight (kg)/height (m)2) 27.2
(SD, 4.4)), we tested the 100-counts/minute Actigraph
threshold against data from the Intelligent Device for
Energy Expenditure and Activity (IDEEA) monitor, which
is a near-gold-standard measure of amounts of time spent
sitting, reclining, and lying down. Compared with direct
observation, the IDEEA monitor has been shown to
correctly identify these body positions with greater than
98 percent accuracy (17). In 2 days of observation in the
validation study, average wear time for both devices was
13.20 hours/day (SD, 2.15), and participants spent approx-
imately 65 percent of their time in sedentary behaviors.
The Actigraph (<100 counts/minute) and the IDEEA re-
corded similar amounts of time spent in sedentary behav-
iors (8.63 hours/day (SD, 1.90) and 8.53 hours/day (SD,
1.86), respectively (p ¼ 0.82)), and correlations between
the measures were moderately high (r ¼ 0.59, p < 0.01;
unpublished observations). These findings suggest that
the 100 counts/minute Actigraph threshold provides a useful
measure of time spent in sedentary behaviors among adults.
There were 7,176 persons who participated in the Acti-
graph component of NHANES. Initial evaluation and clean-
ing of these data resulted in exclusion of the records of
participants whose monitors were not in calibration upon
return (n ¼ 346). In addition, implausible minute-by-minute
values that were consistent with monitor malfunction (e.g.,
32,767 counts/minute) were replaced by imputed values that
were the averages of the counts recorded in the minute
immediately before and the minute immediately after the
implausible value(s). After exclusion of participants with no
recorded monitor-wearing time, 6,830 persons remained.
We then excluded records with fewer than 10 hours/day of
monitor wear, leaving 6,329 persons with at least one day of
observation available for analyses.
2Matthews et al.

Page 3

Statistical analyses
All analyses were conducted with SAS 9.0 (SAS Institute,
Inc., Cary, North Carolina) and SUDAAN 9.0 (Research
Triangle Institute, Research Triangle Park, North Carolina)
to account for the multistage probability design used in
NHANES. Descriptive analyses were initially conducted
to identify possible outliers and to describe the distributions
of our primary data. These analyses revealed significant
differences (p < 0.01) in monitor-wearing time by age,
gender, and racial/ethnic group. Since wear time was posi-
tively associated with sedentary time, we controlled for
wear time in our analysis to enhance our ability to describe
differences between groups. Men, middle-aged adults, and
African Americans tended to wear the monitor for longer
periods of time. Thus, we used analysis of variance to esti-
mate amounts of time spent in sedentary behaviors (hours/
day) after adjustment for average monitor-wearing time.
Standard errors derived from SUDAAN were calculated
via Taylor series linearization. Tests for linear trend were
conducted by constructing a continuous variable that con-
tained the median value in each category of the variable and
fitting the resulting variable in the regression model as a
continuous variable. Pairwise statistical comparisons of
the conditional marginal means were employed to test the
hypothesisthatgroupswerenotdifferentfromoneanother(18).
RESULTS
Among the 6,329 participants with at least one 10-hour
day of monitor wear, the average number of days of obser-
vation was 5.0 (SD, 1.9). The average monitor-wearing time
was 13.9 hours/day (SD, 1.9), and 79.2 percent of partici-
pants provided 3 or more days of observation. Table 1
presents the frequency distribution of the population sample
by gender, age, and racial/ethnic group. Overall, children
and adults in the United States spent 54.9 percent of their
wakingtime,or7.7hours/day,insedentarybehaviors(figure1
and table 2).
Differences by age and gender
Across the age range of the study population, there were
two striking peaks in the amount of time spent in sedentary
behavior. Children aged 6–11 years were the least sedentary
group in the United States (males: 41.4 percent, 6.0 hours/
day; females: 43.4 percent, 6.1 hours/day), but by age 16–19
years, time spent in sedentary pursuits had increased by
about 2.0 hours/day (males: 55.8 percent, 7.9 hours/day;
females: 59.0 percent, 8.1 hours/day). Young adults (ages
20–29 years) were less sedentary than older adolescents, but
sedentary time increased by about 2 hours/day between the
ages of 30 and 39 years (men: 50.8 percent, 7.2 hours/day;
women: 53 percent, 7.3 hours/day) and the ages of 70 and 85
years (table 2). Adults aged 70–85 years were the most
sedentary group in the population (men: 67.8 percent, 9.5
hours/day; women: 66.3 percent, 9.1 hours/day). Tests for
linear trend by age were significant (p < 0.001) overall
(table 2) and when evaluated separately in youth and adult-
hood (data not shown).
There was also an interaction between age and gender
(figure 1 and table 2). Females were more sedentary than
males throughout youth and early adulthood, but this pattern
was reversed after age 60 years, when the level of sedentary
behavior in men surpassed that of women (p for interaction
< 0.01). However, the interaction between age and gender
in adulthood was significant in Whites (p < 0.01) but not in
Blacks or Mexican Americans.
Differences by race/ethnicity
In youths (ages 6–19 years), there were few significant
differences by race/ethnicity, except that Black girls aged
6–11yearswerelesssedentarythaneitherWhitesorMexican
Americans (table 3). In addition, when males and females
were examined together, Mexican-American adolescents
were less sedentary than Whites. In adulthood (ages 20–
85 years), Mexican Americans were less sedentary than
Whites and Blacks, both before and after age 60 years.
Differences between White and Black adults were few, ex-
cept that White men aged 40–59 years were more sedentary
than Black men of the same age. White and Black females
TABLE 1.
in sedentary behaviors, by gender, age, and racial/ethnic group,
United States, 2003–2004
Sample size for a study of the amount of time spent
Gender and age
group (years)
Total*
Non-
Hispanic
Whites
Non-
Hispanic
Blacks
Mexican
Americans
Males
All3,1201,334828 782
6–11 386104 138 118
12–15466 121 180146
16–19429113162137
20–29 300 1386960
30–39282131 6162
40–49 309 1547272
50–59 238137 5035
60–69299157 5179
70–85411 2794573
Females
All3,209 1,413812 796
6–11425117 149138
12–15420 105145 151
16–19 405126 142113
20–29 336169 6473
30–39322 1745969
40–49 3191449064
50–59 255 15053 34
60–69 3331597087
70–85 394269 4067
* The total sample included persons of all racial/ethnic groups in
the study; therefore, numbers for the individual racial/ethnic groups in
the table do not add up to those of the total sample.
Time Spent in Sedentary Behaviors3

Page 4

weresimilarlysedentaryafterage12years.Mexican-American
men aged 20–39 years were the least sedentary group of
adults (about 6 hours/day). In general, there were consistent
positive associations between age and time spent in seden-
tary behaviors in all racial/ethnic groups.
DISCUSSION
To our knowledge, this is the first quantification of the
amount of time spent in overall sedentary behaviors in
a large representative population sample in the United
States. Time spent in these behaviors reflects the accumu-
lated amount of time spent sitting, reclining, or lying down
at home, at work, at school, in transit, and during leisure
time. These US children and adults spent approximately
55 percent of their waking hours, or 7.7 hours/day, in be-
haviors that result in only very low levels of energy expen-
diture. Older adolescents (ages 16–19 years) and older
adults (ages 60–85 years) spent nearly 60 percent of their
time, or more than 8 hours/day, in sedentary behaviors. In
contrast, children aged 6–11 years spent only 6 hours/day in
these behaviors. Mexican-American adults were the least
sedentary group of adults, while Black and White adults
tended to be similarly sedentary. Given the amount of time
spent in these behaviors on a daily basis, particularly for
older adolescents and adults, reducing the amount of time
spent being sedentary representsan important opportunity for
increasing the level of physical activity in the population.
Our study differs significantly from other population-
based estimates of sedentary behaviors, in terms of both
methodology and critical findings, and thus provides unique
insight into the current level of sedentary behavior in the
United States. In contrast to previous estimates of sedentary
behavior derived from self-reports of a limited number of
behaviors (i.e., television viewing and computer use), we
directly quantified the total amount of time spent at very
low activity levels over an average of 13.9 hours/day. Given
our method of estimating time spent in sedentary behavior,
it is not surprising that our results showed that children and
adults are twice as sedentary as indicated by previous esti-
mates derived only from assessment of media-related dis-
cretionary behaviors (13, 19). These results highlight the
idea that media use accounts for only about half of the over-
all time spent in sedentary behavior in the US population.
The second important finding from our study is that pre-
viously quantified differences in television viewing and
computer use that have been described by age, gender,
and ethnicity may not translate to differences in levels of
overall sedentary behavior. For example, prevalence esti-
mates of sedentary behavior, defined as spending 3 or more
hours per day watching television or using a computer, from
the Youth Risk Behavior Surveillance System in 2005 in-
dicated that males were more sedentary than females,
Blacks were more sedentary than Hispanics or Whites,
and time spent in these behaviors decreased between grades
9 and 11 (20). Our study revealed a markedly different
picture among youth. We found that girls were more seden-
tary than boys, Blacks were similar to both Whites and
Mexican Americans, and the level of overall sedentary be-
havior increased substantially during adolescence. Perhaps
the most striking variation in our results from previous stud-
ies was seen among Black and White children. Previous
reports have found that approximately 40 percent of Black
children reported watching 4 or more hours of television per
day, while only about 20 percent of White children reported
this much televisionviewing (21, 22). We found no evidence
that this difference in television viewing translated into dif-
ferences in objectively determined levels of overall seden-
tary behavior between White and Black children.
TABLE 2.
hours/day), by gender and age, United States, 2003–2004
Amount of time* spent in sedentary behavior (mean
Age group
(years)
All participantsMales Femalesp valuey
Total 7.67 (0.04)z
7.63 (0.04)7.70 (0.03)0.001
Youths
6–11 6.07 (0.06)6.00 (0.10)6.14 (0.05)0.002
12–157.53 (0.10) 7.37 (0.15) 7.70 (0.08) 0.003
16–198.03 (0.08)7.91 (0.13) 8.13 (0.10) 0.028
Adults
20–29 7.48 (0.11) 7.27 (0.16)7.68 (0.09)0.001
30–397.25 (0.10)7.17 (0.17)7.34 (0.08)0.061
40–497.55 (0.08) 7.57 (0.11)7.53 (0.11)0.252
50–59 7.87 (0.09) 7.93 (0.14)7.82 (0.08) 0.515
60–69 8.41 (0.09)8.80 (0.14)8.08 (0.10)0.003
70–85 9.28 (0.06)9.52 (0.06)9.11 (0.08)0.015
p-trend
<0.001
<0.001
<0.001
* Estimates of amounts of time spent in sedentary behaviors are
least-square means adjusted for monitor-wearing time and age, as
well as for gender as appropriate.
yt test for the difference between males and females in each age
group.
zNumbers in parentheses, standard error.
FIGURE 1.
and gender, United States, 2003–2004. Bars, standard error.
Percentage of time spent in sedentary behaviors, by age
4 Matthews et al.

Page 5

Close evaluation of time spent in sedentary behaviors
across the wide age range in these data revealed two dra-
matic peaks that merit comment. Over the approximate de-
cade from ages 6–11 years to ages 16–19 years, there was
about a 2-hour/day increase in sedentary behavior, and this
was approximately the same magnitude of increase as that
observed across the four decades from young adulthood
(ages 20–39 years) to older age (ages 70–85 years). Previous
studies have also reported that the rate of decline in physical
activity levels is greater in adolescence than in adulthood
(23). Studies using objective measures of sedentary behav-
ior to describe age effects in youth have reported results that
are similar to ours (24, 25). For example, Treuth et al. (24)
described nearly a 2-hour/day increase in time spent in sed-
entary behaviors when comparing elementary, middle, and
high school students (proportion of monitored time in sed-
entary behavior: 46.7 percent, 51.4 percent, and 56.0 per-
cent, respectively), and this increase was predominantly due
to a reduction in time spent in light-intensity activities (i.e.,
100–899 counts/minute by Actigraph). Pate et al. (26) re-
cently reported that sixth-grade girls (ages 11–12 years)
spent 55.6 percent of their time, or 7.7 hours/day, in seden-
tary behaviors, which is virtually identical to results from
the present study. Fewer studies have estimated the amount
of time adults spend in sedentary behaviors, but results from
the American Time Use Survey indicate that a large pro-
portion of the average 15.4-hour waking day is spent in
activities that expend very little energy (13). In concordance
with our study, Healy et al. (16) employed the Actigraph and
reported that Australian adults (average age 53.3 years)
spent 57 percent of their monitored time in activities below
100 counts/minute. The reduction in time spent being sed-
entary between adolescence (ages 16–19 years) and early
adulthood (ages 20–39 years) may reflect entrance into the
workforce and domestic responsibilities related to raising
families (13). The increase in time spent in sedentary be-
haviors after age 60 years, particularly among men, may
reflect an increase in leisure time following retirement
and/or the development of comorbid health conditions that
may increasingly limit activity in later life. Troiano et al.
(27) reported a reduction in overall activity levels with in-
creasing age among men in this NHANES sample, which
suggests that the increase in time spent in sedentary behav-
ior we observed may not be offset by increases in leisure-
time physical activity as reported by previous investigators
(28, 29).
Time spent in pursuits that require only low levels of
energy expenditure have been linked to increased risk of
weight gain (2, 30), an adverse metabolic profile (4, 16,
31, 32), and type 2 diabetes (3). Westerterp (1) has described
the importance of lower-intensity (e.g., light-to-moderate)
activity on physical activity energy expenditure levels, and
Levine et al. (2) suggested that an additional 2 hours/day
spent sitting conserved approximately 350 kcal/day in
a small sample of obese adults, relative to lean adults. In
the context of the current study, wherewe observedtwo age-
related increases in sedentary behaviors of approximately 2
hours/day in adolescence and adulthood, we estimate that
approximately 2 metabolic equivalent (MET)-hours/day (or
2 kcal/kg/day) of energy would be conserved by shifting 2
hours/day from light-intensity behaviors (2.5 METs) to sed-
entary behaviors (1.5 METs). For a person with a resting
energyexpenditure of 67 kcal/hour, this would equate to 134
kcal/day. This amount of energy is slightly greater than that
associated with 30 minutes/day of brisk walking (0.5 hours/
day 3 3.5 METs ¼ 1.75 MET-hours/day, or 117 kcal/day)
and is similar to the level of energy imbalance that has been
associated with the current obesity epidemic (33). Hence,
even modest reductions in time spent in sedentary behaviors
have the potential to increase energy expenditure and alter
energy balance, if the increase in expenditure is not offset by
a compensatory increase in energy intake.
Strengths of this study include the use of a large repre-
sentative sample of children and adults in the United States,
TABLE 3.
hours/day), by age, gender, and racial/ethnic group, United
States, 2003–2004
Amount of time* spent in sedentary behavior (mean
Gender and age
group (years)
Non-Hispanic
Whites
Non-Hispanic
Blacks
Mexican
Americans
All participants 7.74 (0.06)y
6.03 (0.07)
7.61 (0.05) 7.18 (0.05)z
6.03 (0.09) 6–116.10 (0.10)
12–157.47 (0.13) 7.62 (0.12) 7.55 (0.07)
16–19 8.03 (0.10)8.10 (0.10) 7.62 (0.08)§
20–397.42 (0.12)7.81 (0.08)6.47 (0.10)z
7.00 (0.07)z
8.31 (0.11)z
<0.001
40–59 7.79 (0.08)7.59 (0.08){
9.04 (0.11)?60 8.85 (0.08)
p-trend
<0.001
<0.001
Males 7.73 (0.08)7.54 (0.06)6.89 (0.07)z
6.08 (0.13)6–11 5.90 (0.13) 6.30 (0.12)
12–15 7.36 (0.18)7.44 (0.19)7.35 (0.11)
16–19 7.91 (0.19)8.00 (0.12)7.42 (0.13)
20–39 7.36 (0.19)7.81 (0.13)5.98 (0.12)z
6.82 (0.12)§40–59 7.83 (0.10)7.57 (0.17){
9.32 (0.17)?60 9.14 (0.11)8.42 (0.20)z
0.001p-trend
<0.001
<0.001
Females7.74 (0.04)7.67 (0.06)7.47 (0.05)z
6.02 (0.08)6–116.18 (0.06) 5.88 (0.14){,#
7.83 (0.12)12–157.60 (0.12) 7.78 (0.09)
16–198.15 (0.12) 8.20 (0.10)7.89 (0.08)
20–397.49 (0.09)7.80 (0.08)6.97 (0.13)z
7.18 (0.10)§40–59 7.74 (0.11) 7.60 (0.13)
?60 8.60 (0.10)8.83 (0.15)8.26 (0.13)
p-trend
<0.001
<0.001
<0.001
* Estimates of amounts of time spent in sedentary behaviors
(hours/day) are least-square means adjusted for monitor-wearing
time, as well as for age and gender as appropriate.
yNumbers in parentheses, standard error.
zMexicans were significantly less sedentary than Whites and
Blacks (p < 0.05).
§ Mexicans were significantly less sedentary than Whites (p <
0.05).
{ Blacks were significantly less sedentary than Whites (p < 0.05).
# Blacks were significantly less sedentary than Mexicans (p <
0.05).
Time Spent in Sedentary Behaviors5

Page 6

oversampling of Blacks and Mexican Americans to improve
precision of the estimates in these groups, and the use of an
objective measurement of overall sedentary behaviors. We
supported our accelerometer cutpoint selection with a vali-
dation study in adults using a near-gold standard. This same
cutpoint was identified in a calibration study among adoles-
cent girls (15). An additional strength of our study was that
the majority of participants were quite compliant with the
measurement procedures. While the minimum number of
days of observation required to be included in our analyses
was only 1, nearly 80 percent of the sample wore the device
for at least 10 hours/day on 3 or more days. There has been
some debate regarding the minimal number of days of ob-
servation required to achieve useful estimates from acceler-
ometers in field-based studies (34, 35) and issues associated
with missing data (36). Our decision to use data for persons
with at least one day of observation differs from decisions
made by other investigators, but our results are virtually
identical to those of Healy et al. (16), who required at least
5 days of observation, including oneweekend day, and those
of Pate et al. (26), who employed a minimum of 6 days of
measured and imputed observation. This suggests that our
methods provide useful estimates of the population mean for
amount of time spent in sedentary behavior.
Limitations of our study should also be considered. The
amount of monitor-wearing time in our study (13.9 hours/
day) was approximately 1.5 hours/day less than the average
waking time reported in other national surveys (13). Thus,
our estimates of time spent in sedentary behaviors are likely
to have underestimated the actual amount of time spent in
these behaviors.If we normalized the population average for
the proportion of monitored time spent being sedentary
(54.9 percent) to a 15.4-hour waking day, our estimate of
time spent in sedentary behavior would increase from 7.7
hours/day to 8.5 hours/day. In addition, the assessment of
sedentary behaviors by accelerometer is relatively new, and
our estimates of time spent in these behaviors will have been
sensitive to the cutpoints employed. That said, it is unlikely
that cutpoint selection would have affected our results with
respect to age, gender, and race/ethnicity. Future studies that
use more direct estimates of body position (e.g., see the
paper by Grant et al. (37)) are needed to confirm the present
results. Additionally, our results provide estimates of the
average amount of time spent per day in sedentary behaviors
and reflecttheinfluence ofbothweekend daysand weekdays;
patterns of behavior may vary on these days. Finally, our
cross-sectional study design limits causal inferences, partic-
ularly for the age differences reported, and our results are
generalizable only to the noninstitutionalized population of
the United States. Results from this report should be inter-
preted accordingly.
In conclusion, children and adults in the United States
spend the majority of their waking time in sedentary behav-
iors. The most sedentary groups in the population are older
adolescents (ages 16–19 years) and older adults (ages ?60
years), while the least sedentary groups are children (ages
6–11 years) and Mexican-American males. Given the large
amount of time spent each day in sedentary behaviors in the
United States, efforts to reduce the amount of time spent in
low-energy-expenditure pursuits are warranted.
ACKNOWLEDGMENTS
The authors thank the staff of the Risk Factor Monitoring
and Methods Branch of the National Cancer Institute for
assisting with incorporation of the Actigraph measures into
the NHANES survey.
Conflict of interest: none declared.
REFERENCES
1. Westerterp KR. Pattern and intensity of physical activity.
Nature 2001;410:539–40.
2. Levine JA, Lanningham-Foster LM, McCrady SK, et al.
Interindividual variation in posture allocation: possible role in
human obesity. Science 2005;307:584–6.
3. Hu FB, Li TY, Colditz GA, et al. Television watching and other
sedentary behaviors in relation to risk of obesity and type 2
diabetes mellitus in women. JAMA 2003;289:1785–91.
4. Ford ES, Kohl HW III, Mokdad AH, et al. Sedentary behavior,
physical activity, and the metabolic syndrome among U.S.
adults. Obes Res 2005;13:608–14.
5. Morris JN, Heady JA, Raffle PA, et al. Coronary heart-disease
and physical activity of work. Lancet 1953;265:1111–20.
6. Ogden CL, Carroll MD, Curtin LR, et al. Prevalence of over-
weight and obesity in the United States, 1999–2004. JAMA
2006;295:1549–55.
7. Dietz WH. The role of lifestyle in health: the epidemiology
and consequences of inactivity. Proc Nutr Soc 1996;55:
829–40.
8. Owen N, Leslie E, Salmon J, et al. Environmental determi-
nants of physical activity and sedentary behavior. Exerc Sport
Sci Rev 2000;28:153–8.
9. Epstein LH, Roemmich JN. Reducing sedentary behavior: role
in modifying physical activity. Exerc Sport Sci Rev 2001;29:
103–8.
10. Pate RR, Pratt M, Blair SN, et al. Physical activity and public
health. A recommendation from the Centers for Disease
Control and Prevention and the American College of Sports
Medicine. JAMA 1995;273:402–7.
11. Nelson ME, Rejeski WJ, Blair SN, et al. Physical activity and
public health in older adults: recommendation from the
American College of Sports Medicine and the American Heart
Association. Med Sci Sports Exerc 2007;39:1435–45.
12. Haskell WL, Lee IM, Pate RR, et al. Physical activity and
public health: updated recommendation for adults from the
American College of Sports Medicine and the American Heart
Association. Med Sci Sports Exerc 2007;39:1423–34.
13. Bureau of Labor Statistics, US Department of Labor. American
Time Use Survey—2004 results announced by BLS. (News
release). (USDL publication no. 05-1766). Washington, DC:
Bureau of Labor Statistics, 2005. (http://www.bls.gov/news.
release/archives/atus_09202005.pdf).
14. Matthews CE. Calibration of accelerometer output for adults.
Med Sci Sports Exerc 2005;37(suppl):S512–22.
15. Treuth MS, Schmitz K, Catellier DJ, et al. Defining acceler-
ometer thresholds for activity intensities in adolescent girls.
Med Sci Sports Exerc 2004;36:1259–66.
16. Healy GN, Dunstan DW, Salmon J, et al. Objectively mea-
sured light-intensity physical activity is independently asso-
ciated with 2-h plasma glucose. Diabetes Care 2007;30:
1384–9.
17. Zhang K, Werner P, Sun M, et al. Measurement of human daily
physical activity. Obes Res 2003;11:33–40.
6 Matthews et al.

Page 7

18. Research Triangle Institute. SUDAAN user’s manual, release
8.0. Research Triangle Park, NC: Research Triangle Institute,
2001.
19. Roberts DF, Foehr UG, Rideout V. Generation M: media in the
lives of 8–18 year olds. Menlo Park, CA: Henry J. Kaiser
Family Foundation, 2005. (http://www.kff.org/entmedia/
entmedia030905pkg.cfm).
20. Eaton DK, Kann L, Kinchen S, et al. Youth risk behavior
surveillance—United States, 2005. MMWR Surveill Summ
2006;55:1–108.
21. Andersen RE, Crespo CJ, Bartlett SJ, et al. Relationship of
physical activityand televisionwatching withbody weight and
level of fatness among children: results from the Third
National Health and Nutrition Examination Survey. JAMA
1998;279:938–42.
22. Bowman SA. Television-viewing characteristics of adults:
correlations to eating practices and overweight and health
status. Prev Chronic Dis 2006;3:A38. (Electronic article).
23. Sallis JF. Age-related decline in physical activity: a synthesis
of human and animal studies. Med Sci Sports Exerc
2000;32:1598–600.
24. Treuth MS, Hou N, Young DR, et al. Accelerometry-measured
activity or sedentary time and overweight in rural boys and
girls. Obes Res 2005;13:1606–14.
25. Janz KF, Burns TL, Levy SM, et al. Tracking of activity and
sedentary behaviors in childhood: The Iowa Bone Develop-
ment Study. Am J Prev Med 2005;29:171–8.
26. Pate RR, Stevens J, Pratt C, et al. Objectively measured
physical activity in sixth-grade girls. Arch Pediatr Adolesc
Med 2006;160:1262–8.
27. Troiano RP, Berrigan D, Dodd KW, et al. Physical activity in
the United States measured by accelerometer. Med Sci Sports
Exerc 2008;40:181–8.
28. Schnohr P, Scharling H, Jensen JS, et al. Changes in leisure-
time physical activity and risk of death: an observational study
of 7,000 men and women. Am J Epidemiol 2003;158:639–44.
29. Caspersen CJ, Pereira MA, Curran KM, et al. Changes in
physical activity patterns in the United States, by sex and
cross-sectional age. Med Sci Sports Exerc 2000;32:1601–9.
30. Levine JA, Eberhardt NL, Jensen MD. Role of nonexercise
activity thermogenesis in resistance to fat gain in humans.
Science 1999;283:212–4.
31. Dunstan DW, Salmon J, Owen N, et al. Physical activity
and television viewing in relation to risk of undiagnosed
abnormal glucose metabolism in adults. Diabetes Care 2004;
27:2603–9.
32. Dunstan DW, Salmon J, Healy GN, et al. Association of tele-
vision viewing with fasting and 2-h postchallenge plasma
glucose levels in adults without diagnosed diabetes. Diabetes
Care 2007;30:516–22.
33. Hill JO, Wyatt HR, Reed GW, et al. Obesity and the environ-
ment: where do we go from here? Science 2003;299:853–5.
34. Trost SG, McIver KL, Pate RR. Conducting accelerometer-
based activity assessments in field-based research. Med Sci
Sports Exerc 2005;37(suppl):S531–43.
35. Masse LC, Fuemmeler BF, Anderson CB, et al. Accelerometer
data reduction: a comparison of four reduction algorithms
on select outcome variables. Med Sci Sports Exerc 2005;
37(suppl):S544–54.
36. Catellier DJ, Hannan PJ, Murray DM, et al. Imputation of
missing data when measuring physical activity by acceler-
ometry. Med Sci Sports Exerc 2005;37(suppl):S555–62.
37. Grant PM, Ryan CG, Tigbe WW, et al. The validation of
a novel activity monitor in the measurement of posture and
motion during everyday activities. Br J Sports Med 2006;40:
992–7.
Time Spent in Sedentary Behaviors7