Content uploaded by John R Sirard
Author content
All content in this area was uploaded by John R Sirard
Content may be subject to copyright.
Physical Activity Assessment in
Children and Adolescents
John R. Sirard and Russell R. Pate
University of South Carolina, Columbia, South Carolina, USA
Contents
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 439
1. Criterion Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 440
1.1 Direct Observation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 441
1.2 Doubly Labelled Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443
1.3 Indirect Calorimetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443
2. Objective Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443
2.1 Heart Rate Monitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443
2.2 Motion Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445
2.2.1 Pedometers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445
2.2.2 Accelerometers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445
3. Subjective Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 448
3.1 Self-Report Questionnaires . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 448
3.2 Interviewer-Administered Questionnaires . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 448
3.3 Proxy-Reports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 450
3.4 Diaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 450
4. Future Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 451
5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 451
Abstract
Chronic disease risk factors, including a sedentary lifestyle, may be present
even in young children, suggesting that early prevention programmes may be
critical to reducing the rates of chronic disease. Accurate assessment of physical
activity in children is necessaryto identifycurrent levels of activityand to assess
the effectiveness of intervention programmes designed to increase physical ac-
tivity. This article summarises the strengths and limitations of the methods used
to evaluatephysical activityinchildren andadolescents. MEDLINEsearchesand
journal article citations were used to locate 59 articles that validated physical
activity measurement methods in children and adolescents. Only those methods
that were validated against a morestringent measurewere included in the review.
Based on the definition of physical activity as any bodily movement resulting
in energy expenditure (EE), direct observation of the individual’s movement should
be used as the gold standard for physical activity research. The doubly labelled
water technique and indirect calorimetry can also be considered criterion meas-
ures for physical activity research, because they measure EE, a physiologic con-
sequence closely associated with physical activity. Devices such as heart rate
monitors, pedometers and accelerometers have become increasingly popular as
measurement tools for physical activity. These devices reduce the subjectivity
REVIEW ARTICLE
Sports Med 2001; 31 (6): 439-454
0112-1642/01/0006-0439/$22.00/0
© Adis International Limited. All rights reserved.
inherent in survey methods and can be used with large groups of individuals.
Heart rate monitoring is sufficiently valid to use in creating broad physical activ-
ity categories (e.g. highly active, somewhat active, sedentary) but lacks the spec-
ificity needed to estimate physical activity in individuals. Laboratory and field
validations of pedometers and accelerometers yield relatively high correlations
using oxygen consumption (r = 0.62 to 0.93) or direct observation (r = 0.80 to
0.97) as criterion measures,although,they may not be able tocaptureall physical
activity.
Physical activity has traditionally been measured with surveys and recall in-
struments. These techniques must be used cautiously in a paediatric population
that has difficulty recalling such information. Still, some studies have reported
73.4% to 86.3% agreement between these instruments and direct observation.
Future investigations of physical activity instruments should validate the novel
instrument against a higher standard.Additional studies are neededto investigate
thepossibility ofimprovingtheaccuracyofmeasurementbycombining2ormore
techniques. The accurate measurement of physical activity is critical for deter-
mining current levels of physical activity, monitoring compliance with physical
activity guidelines, understanding the dose-response relationship between phys-
ical activity and health and determining the effectiveness of intervention pro-
grammes designed to improve physical activity.
Physical activity is defined as ‘any bodily move-
ment produced by skeletal muscle that results in en-
ergy expenditure’.
[1]
It is now well established that
an inverse relationship exists between physical activ-
ityandriskfordevelopingseveralchronicdiseases,
including obesity, coronary heart disease (CHD),
diabetesandcolon cancer.
[2-4]
Sinceobesityandthe
risk factors for CHD and diabetes can be present
even in young children,
[5-7]
it is important that pri-
mary prevention programmes involving physical ac-
tivitybeginearlyinlife.Toassesslevelsofphysical
activityand determinethe effectivenessofphysical
activity intervention programmes, accurate meas-
ures of physical activity are required.
[8,9]
Measure-
ment techniques used for research and programme
evaluation purposes must be valid, reliable, practi-
cal and nonreactive.
[8,10]
This article will review the strengths, limita-
tions and validity of the subjective and objective
techniques that have been developed to assess phys-
ical activity in children and adolescents. MEDLINE
searches were used to identify studies of physical
activity measurement in children and adolescents;
keywords included physical activity, children, youth,
adolescent and energy expenditure (EE). Other
sources were identified by journal article citations.
Onlystudiesthatreportedthe validity of the instru-
mentwereincludedinthereview. Reliabilityof the
instrumentis alsopresented ifthatinformationwas
provided along with the validation data.
1. Criterion Standards
This review considered 3 types of measures of
physical activity in children and adolescents: pri-
mary measures, secondary measures and subjec-
tive measures. Figure 1 represents the 3 levels of
physicalactivitymeasures usedforthisreview.For
the purpose of this review, direct observation, dou-
bly labelled water (DLW) and indirect calorimetry
are considered the primary standards for assess-
ment of physical activity in children and adolescents.
DLW is well recognised as a criterion measure for
field evaluations of EE. This technique assesses to-
tal caloric expenditure by estimating carbon diox-
ide production using isotope dilution during a min-
imum of 3 days. EE is a physiologic consequence
of physical activity and is directly linked to health
and disease prevention. Thus, DLW and indirect
440 Sirard & Pate
Adis International Limited. All rights reserved. Sports Med 2001; 31 (6)
calorimetry can be used as criterion measures for
physical activity assessment. However, it should
be noted that EE and physical activity are distinct
constructs, which may limit attempts to validate
physical activity measures against EE. Cardiore-
spiratory fitness measured by indirect calorimetry
during progressive exercise tests has been used to
indirectly validate physical activity surveys. This
association, however, is weak or unclear in chil-
dren and adolescents.
[11,12]
Therefore, studies us-
ing this indirect validation were not included in
this review.
Direct observation is a more practical and com-
prehensive criterion measure for physical activity
research.Basedonthe abovedefinitionof physical
activity, direct observation of movement seems to
be the most appropriate standard for physical ac-
tivity assessment. Subsequent sections will explore
the strengths and limitations of these techniques in
more detail.
Heart rate monitors,pedometersandaccelerom-
eters will be considered secondary measures be-
cause they provide an objective assessment of phys-
ical activity.Validating oneof these measures against
another secondary measure provides little insight
to the instruments’ true validity. For this reason
only the results from studies that validated a sec-
ondary method against a primary measure are in-
cluded in this review. These secondary measures
may be used, however, as criterion standards to
validate subjective measures of physical activity
behaviour (see fig. 1).
Surveysandother subjectivetechniques used as
criterion measures carry the least compelling vali-
dation results and should not be used in this capac-
ity. Therefore, only those subjective measures that
were validated against a more stringent standard
areincludedinthisreview.
1.1 Direct Observation
Direct observation is the most practical and ap-
propriatecriterionmeasureofphysicalactivityand
patterns of activity. Seven observational systems
are reviewed in table I. While 2 of these systems
are specific for observation during physical educa-
tion classes,
[16,18,19]
the others can be used in a va-
rietyofsettings.
[13-15,17,20]
Evidencesupporting the
use of these instruments is available from studies
comparingdirectobservationscores withheartrate
or oxygen consumption. Correlations range from r
=0.61to0.91
[15,16,20]
and heartrate oroxygen con-
sumption were significantly different among the
observed physical activity intensity levels.
[13,14,18]
All 7 observational techniques attained satisfactory
inter-observer agreement (84% to 99%) among si-
multaneous observationsofthesamechild.
[13-17,19,20]
The total observation time required to attain ac-
ceptable day-to-day stability is not clear for most
observational instruments. Drawbacks of direct
observation include the relatively high experi-
menter burden and the potential reactivity of the
study participant. Puhl et al.
[13]
found that only
16.6% of the 5- to 6-year-olds observed in their
study reacted to the observers. The ability of ob-
servational techniques to capture short term pat-
terns and sudden changes in physical activity is
crucial for the study of young children. McKen-
zie
[21]
suggests using a previously tested instrument
rather than creating new techniques so that results
from future studies will be comparable to earlier
research.
Criterion standards
1. Direct observation
2. Doubly labelled water
3. Indirect calorimetry
Secondary measures
1. Heart rate
2. Pedometers
3. Accelerometers
Subjective measures
1. Self-report
2. Interview
3. Proxy-report
4. Diary
Fig. 1. Validation Schema. Arrows indicate acceptable criterion
standardsforthevalidationoftertiaryandsecondarylevelmeth-
ods.
Physical Activity Assessment in Children and Adolescents 441
Adis International Limited. All rights reserved. Sports Med 2001; 31 (6)
442 Sirard & Pate
© Adis International Limited. All rights reserved. Sports Med 2001; 31 (6)
Table I. Validation of direct observation techniques used to assess young people’s physical activity
Instrument Technique Participants Reliability Criterion
measure
Validity Reference
Children’s activity rating
scale (CARS)
1 minute partial time sampling with
5 categories during various
conditions
12 boys,
13 girls;
5-6y
84 ± 10% agreement between
observers;
n = 192, 3-4y
V
.
O
2
,HR V
.
O
2
and HR differed between treadmill speeds
designed to represent the 5 categories of
exercise intensity (p < 0.05)
13
Modified fargo activity
timesampling survey
(FATS)
3 second continuous time sampling
with 30 categories during free-
living
a
conditions
2 boys,
2 girls;
7-10y
91% agreement between
observers;
n = 15, 6-10y
V
.
O
2
Categories separated by intensity as measured
by
V
.
O
2
14
Activity patterns and energy
expenditure
(APEE)
15 second momentary time
sampling with 5 categories during
freeplay
19 girls;
5-8y
86-99% agreement between
observers
HR r = 0.72-0.91 15
Children’s physical activity
form
(CPAF)
1 minute partial time sampling with
4 categories during PE
18 boys,
18 girls;
8-10y
96-98% agreement between
observers
HR r = 0.61-0.72 16
Behaviours of eating and
activity for children’s health
evaluation system
(BEACHES)
1 minute momentary time sampling
with 5 categories during various
conditions
19;
4-9y
94-99% agreement among
observers; Kappa = 0.71-1.0;
n = 17 boys, 25 girls; 4-8y
HR HR increased for each of the 5 categories 17
System for observing
fitness instruction time
(SOFIT)
10 second momentary time
sampling with 5 categories during
PE class
173;
Grades 1-8
N/A HR HR increased for each of the 5 categories
except lying versus sitting categories
18
System for observing fit-
ness instruction time
(SOFIT)
10 second momentary time
sampling with 5 categories during
PE class
88;
Grades 3-5
88.3% agreement among
observers
lesson
context
(fitness)
r = -0.65 w/ standing;
r = 0.49 w/ walking;
r = 0.36 w/ very active;
r = 0.69 w/ MVPA
19
Fargo activity timesampling
survey
(FATS)
10 second momentary time
sampling with 8 categories during
various conditions
7 boys, 7
girls;
2-4y
91-98% agreement among
observers
LSI® r = 0.78-0.90 20
a normal daily life.
HR = heart rate; LSI
®
= large scale integrated physical activity monitor; MVPA = moderate to vigorous physical activity; n = sample size; PE = physical education; r = Pearson
product-moment correlation coefficient; V
.
O
2
= oxygen consumption; y = age of participants (years).
1.2 Doubly Labelled Water
With this method, a dose of a radio-labelled iso-
tope (
2
H
2
18O) is administered orally and the oxy-
gen atoms in expired CO
2
equilibrate with the ox-
ygen atoms in the body water. Over the next 5 to 14
days,
2
H is eliminated as water, while
18
O will be
eliminated as waterandCO
2
.Thedifferencebetween
theeliminationratesis proportionaltoCO
2
produc-
tion (i.e. EE).
[22]
The DLW method has been vali-
dated against whole room calorimetry in adults
[22-24]
and with periodic respiratory gas exchange in in-
fants.
[25]
Similar research with children was not
found, probably becauseofthedifficultyin obtain-
ing consent from children and their parents for mul-
tiple days of calorimeter confinement. One study
was identified that associated the DLW technique
with several biological markers in thirty 4- to 6-
year-oldchildren.
[26]
Total energyexpenditure(TEE)
was positively associated with fat-free mass (r =
0.86), body mass (r = 0.83), body surface area (r =
0.82), height (r = 0.74) and fat mass (r = 0.65).
Activity energy expenditure (AEE) [AEE = total –
resting EE] was significantly correlated with the
same variables (r = 0.56 to 0.74).
[26]
The DLW technique has several advantages for
evaluating EE. It can be easily used easily in free-
living (normal daily life) participants, has low re-
activity and is accurate to within 3 to 4% of calo-
rimeter values in adults.
[24]
Unfortunately, DLW
also has several major limitations.
[24]
First, the iso-
topesaredifficultto obtain,veryexpensiveandnot
suitable for large studies. Second, accurate dietary
records must be obtained during the measurement
period for EE calculations. Lastly, measurements
must be taken over at least a 3-day period
[23]
and
only TEE can be obtained. Therefore, daily or hourly
patterns of EE cannot be investigated. While TEE
is critical, it may be equally important to evaluate
other parameters associated with physical activity
such as the duration, intensity and frequency of
moderate-to-vigorous physical activity (MVPA),
vigorous physical activity (VPA),orsedentary be-
haviour.
1.3 Indirect Calorimetry
Open-circuit indirect calorimetry measures EE
from O
2
consumption and CO
2
production. Indi-
rect calorimetry during rest and exercise is used
extensively and considered an accurate and valid
measure of short term EE.
However, using indirect calorimetry to measure
physical activity is difficult because of the non-
portable gas analysis equipment required. There-
fore,thismethodisimpracticalforvalidatinga sur-
veythat measures‘usual’orweeklyphysicalactivity.
Indirect calorimetry has been used, however, to val-
idate heart rate monitors, pedometers and acceler-
ometersinlaboratorysettings.
[27-32]
Manufacturers
are now introducing portable, lightweight metabo-
licsystemsthatshouldimprovetheestimatesofEE
during physical activities under more natural set-
tings. Despite this advance, the equipment is still
too cumbersome to useunder longterm free-living
conditions, especially in young children.
2. Objective Techniques
Several objective techniques, such as heart rate
monitors, pedometersand accelerometers,arenow
widely available for the measurement of physical
activity.
[33]
This review includes only the results
from these secondary measures which have been
validatedagainstaprimarystandard.Thestrengths
and limitations of each technique are also consid-
ered.
2.1 Heart Rate Monitors
Heart rate monitoring as a means of estimating
EEorphysicalactivityhasbeenusedinbothyoung
people and adults and relies on the linear relation-
ship between heart rate and oxygen consumption
(V
.
O
2
). But this relationship is not as robust at the
low end of the physical activity spectrum. During
sedentary or light intensity activities, an individu-
al’s heart rate can be affected by factors other than
body movement.
[34]
Psychological and environmen-
tal stress, as well as caffeine and some medications
can significantly affect heart rate.
[28]
The FLEX
HR method has been employed to limit these ef-
Physical Activity Assessment in Children and Adolescents 443
Adis International Limited. All rights reserved. Sports Med 2001; 31 (6)
fects in young people
[27,28,34-36]
and adults.
[37,38]
Livingstone et al.
[34]
describe the FLEX HR as an
individually-determined heart rate, measured in con-
junction with V
.
O
2
, that can be used to distinguish
between resting and AEE. Resting metabolic rate
is substituted for periods when the heart rate falls
below the FLEX HR.
The FLEX HR technique was validated in the
studies summarisedin tableIIusing the DLWtech-
nique or whole room calorimetry as the criterion
measure. While the FLEX HR method assessed TEE
at the group level even in children with cerebral
palsy,
[36]
this was not the case for calculating indi-
vidual TEE. Bitar et al.
[27]
note that improvements
in estimating individual TEE may be obtained by
not only increasing the number of replicate heart
rate and V
.
O
2
calibration measures but by also in-
cluding typical activities performed by children dur-
ing these procedures. Also, Maffeis et al.
[35]
found
that TEE
FLEX HR
was equivalent to TEE
DLW
for non-
obese children but significantly overestimated TEE
in obese children. These differences may be due to
higher resting and submaximal heart rates and also
prolonged post-exerciseheart rateelevationsof the
obesechildreninthisstudy. Thesmallsample size,
however, limits the interpretation of these results.
Several studies have used absolute heart rate val-
uesto distinguishbetweenactivityintensities.
[39-43]
This method is based on using a percentage of the
maximumheartrate
[44]
and the recommendationby
Simons-Morton et al.
[45]
that an intensity of ≥140
beats per minute approximates MVPA. This may
be a useful method for large epidemiological stud-
ies when individual heart rate/V
.
O
2
curves are not
available.Allorand Pivarnik
[46]
recentlytestedthis
method using 6th grade girls. Their findings indi-
cate that heart rates of140and160beats per minute
were attained at approximately 46
±8% and 63±9%
of V
.
O
2max
, which would correspond to approximate-
ly 5.7 and 7.7 metabolic equivalents (MET; a mea-
sure of energy expenditure equivalent to 1.5 kcal
/kg/h in adults. Resting energy expenditure is con-
sidered 1 MET). Because of the limited age range
of the individuals in this study and the imprecise
nature of this method, it should only be used to
classify groups of individuals rather than to esti-
mate individual EE or physical activity levels.
Using heart rate monitors for the assessment of
physical activity and EE allows for the assessment
of patterns of activity as well as TEE. It is unobtru-
sive, requires minimal participant and experimenter
burdenandis costeffectiveforuseinsmalltomod-
erate size studies. Drawbacks of the FLEX HR
Table II. Validation of heart rate monitoring used to assess young people’s physical activity
Instrument Variables Participants Criterion measure Validity Reference
Heart rate monitor 1 day TEE 9 boys, 10 girls;
mean age = 8.5y
TEE from 1-day
whole-room
calorimeter;
2-week TEE
DLW
TEE
HR
10.4%
> TEEcalorimeter;
TEE
HR
12.3% > TEE
DLW
28
Sport tester PE3000
heart rate monitor
®
2-3 day TEE 23 boys, 21 girls;
7, 9, 12 and 15y
2-week TEE
DLW
95% CI for bias, –0.56-0.01 Mj/d 34
Polar sport tester
®
2-3 day TEE obese: 4 boys, 2 girls;
nonobese: 3 boys, 4
girls; mean age = 9y
1-week TEE
DLW
95% CI for bias;
obese: 0.04-0.92 Mj/d;
nonobese: –0.59-0.63 Mj/d
35
Sport tester heart rate
monitor
®
2-3 day TEE 5 boys, 4 girls; 8-13y
a
2-week TEE
DLW
Spearman r = 0.88;
relative bias = –0.07 Mj/d;
estimate of error = 1.09 Mj/d
36
Heart rate monitor 24 hour TEE 10 boys, 9 girls;
mean age = 10.5y
24-hour whole
room calorimetry
95% CI for bias, –0.15-1.21 Mj/d;
TEE
HR
> TEEcalorimeter;
7.6
± 20.6%
27
a TEE was assessed in children who had reduced physical activity i.e. spastic cerebral palsy.
CI = confidence interval; DLW = doubly labelled water; HR = heart rate; r = Pearson product-moment correlation coefficient; TEE = total
energy expenditure.
444 Sirard & Pate
Adis International Limited. All rights reserved. Sports Med 2001; 31 (6)
method include the need to calibrate individual heart
rate/V
.
O
2
relationships to avoid contamination from
psychologicalandenvironmentalstressors.Although
there are several limitations to heart rate monitor-
ing, the results indicate that this method is a valid
means of estimating EE and physical activity pat-
ternsingroupsoffree-living,nonobeseyoungpeo-
ple.
2.2 Motion Sensors
Consistent with the definition that physical ac-
tivity is bodily movement producing EE, motion
sensors detect that body movement and provide an
estimate of physical activity. Advancements in tech-
nology have increasedthesophistication and accu-
racy of these instruments. Results validating these
motion sensors are presented separately for pedome-
ters, the Caltrac
®
accelerometer (Hemokinetics, Inc.,
Madison, WI) and other accelerometers in the fol-
lowing sections.
2.2.1 Pedometers
Pedometers are relatively simple electronic de-
vices used to estimate mileage walked or the num-
ber of steps taken over a period of time. Studies
using adult participants wearing recent pedometer
models have shown favourable validity and reli-
ability.
[47-49]
Four pedometer validation studies were
identified that used children
[29,30,50,51]
and the re-
sults aresummarised in table III. Kilanowski et al.
[51]
observed a strong association (r = 0.80 to 0.97)
between a Digiwalker DW-200 pedometer and the
Children’s Activity Rating Scale (CARS) direct
observation system.
[13]
Correlations between pe-
dometer step counts and V
.
O
2
during treadmill lo-
comotionrangedfromr=0.62to0.93.
[29,30]
These findings indicate that several newer pe-
dometers may be suited for population-based as-
sessments of physical activity. They are relatively
inexpensive, re-useable, objective and nonreactive.
Pedometers detect only total counts or steps over
the observational period and cannot assess the in-
tensity or pattern of activities performed. Partici-
pants could be instructed to record the number dis-
played on the pedometer at regular intervals to better
capture patterns of activity, but this practice would
decrease objectivity by relying on accurate tran-
scription.
2.2.2 Accelerometers
Accelerometers are more sophisticated elec-
tronic devicesthatmeasureaccelerationsproduced
by bodymovement. In contrasttothe springmech-
anisms of pedometers, accelerometers use piezo-
electric transducers and microprocessors that con-
vert recorded accelerationsto a quantifiable digital
signal referred to as ‘counts’. Westerterp
[52]
recently
reviewed laboratory validations of various accel-
erometers using indirect calorimetry in adult par-
ticipants;pearsoncorrelationsrangedfrom r=0.25
to 0.91. This large variability is due to the use of
different monitors, their placement (e.g. hip, low
back, or ankle) and the specific activities performed
during the measurement protocols.
The Caltrac
®
monitorwasoneofthefirstcom-
mercially available accelerometers and has been
Table III. Validation of pedometers used to assess young people’s physical activity
Pedometer Variables Participants Criterion measure Validity Reference
Yamax Digiwalker
DW-200
®
Mean counts.min-1 10, 7-10y CARS DO
[13]
TriTrac
®
r = 0.80-0.97
r = 0.50-0.99
51
Yamax Digiwalker
DW-200
®
Total counts from hip
Total counts from ankle
Total counts from wrist
15 boys, 15 girls; 8-11y HR, V
.
O
2
r = 0.62-0.92
r = 0.59-0.91
r = 0.17-0.87
29
Yamax Digiwalker
DW-200
®
Total counts from hip
Total counts from ankle
Total counts from wrist
21 Chinese boys; 8-10y V
.
O
2
r = 0.77-0.93
r = 0.68-0.92
r = –0.45-0.82
30
Pedometer Number of steps 11, 4-6y DO (unspecified) r = 0.93 50
CARS = Children’s activity rating scale; DO = direct observation; HR = heart rate; r = Pearson product-moment correlation coefficient;
Tritrac
®
= tri-axial accelerometer; V
.
O
2
= oxygen consumption; y = age of participants (years).
Physical Activity Assessment in Children and Adolescents 445
Adis International Limited. All rights reserved. Sports Med 2001; 31 (6)
the most frequently studied. It is a single (vertical)
plane accelerometer that either provides ‘count’ val-
ues or can estimate EE if biodata (height, body-
weight, age, gender) are supplied. The Caltrac
®
monitor is small and unobtrusive (14 × 8 × 4cm,
400g) making it an attractive method for collecting
physicalactivitydata.Ninestudiesthatmettheinclu-
sion criteria were identified and are summarised in
tableIV.
[31,32,53-57,59,60]
Thesestudiesarepresented in
order of the strength of the criterion measure. Re-
search has found positive but variable associations
between the Caltrac
®
accelerometer and direct ob-
servationmethods(r=0.16to0.86)
[53-56]
or whole
room calorimetry (r = 0.80 to 0.85).
[57]
The wide
variation in correlations against direct observation
are most likely because of the young age of the
participants in several studies (2-6 years) and the
type of activity monitored. Because the Caltrac
®
is
a singleplaneaccelerometer,itis limited in itsabil-
ity to detect the wide variety of movements en-
gaged in by these young participants. Lower corre-
lations were also observed when the activity took
place outdoors compared with controlled labora-
tory conditions. Johnson et al.
[59]
used a previously
developed regression equation
[31]
to calculate 3-day
AEE from Caltrac
®
counts. Based on low correla-
tionswith14-dayAEE
DLW
, they concluded that the
Caltrac
®
accelerometer was not a useful predictor
of AEE. This equation,however, was developed in
a laboratory setting and applied to a free-living sit-
uation in this study. It may be more appropriate to
use just the accelerometer counts rather than at-
tempt to convert counts to units of EE.
Validation studies with newer accelerometers
primarily involve the CSA
®
(Computer Science
andApplications,Inc.,Shalimar,FL)ortheTritrac-
Table IV. Validation of the Caltrac
®
accelerometer used to assess young people’s physical activity
Variables Participants Reliability Criterion measure Validity Reference
Total counts.h free-play
–1
18 boys, 12 girls;
2-6y
N/A FATS DO
[20]
r = 0.39 53
Total counts.h
–1
,
total counts.day
–1
17 boys, 13 girls;
2-4y
N/A FATS DO
[20]
Spearman r = 0.54 54
kcal.h
–1
11 boys, 9 girls;
29 to 40 mo
N/A FATS DO
[20]
Total: r = 0.25-0.62; indoor:
r = 0.47-0.56; outdoor;
r = 0.16-0.48
55
Total counts.h free-play
–1
29 boys, 22 girls;
2-5y
N/A CARS DO
[13]
r = 0.86 56
24-h; counts, TEE, SEE,
WEE
40 girls; 10-16y N/A 24h; TEE, SEE
and WEE via
whole room
calorimeter
r = 0.80 w/ TEE;
r = 0.84 w/ SEE;
r = 0.85 w/ WEE; Caltrac
®
underestimates EE by 6.8%
to 30.4%
57
3-day counts,
calculatedAEE
[58]
22 boys, 14 girls;
mean age = 8.3y
N/A 14-day AEE
DLW
r = –0.09 w/ counts;
calculated AEE > AEE
DLW
(p < 0.01)
59
Calculated EE 10 boys, 10 girls;
mean age = 15.2y
N/A EE
video
EE
video
<EE
Caltrac
; (p < 0.05);
r = 0.95 w/ video
60
Counts at 3 treadmill
speeds
9 boys, 6 girls;
8-13y
Left hip vs right hip;
r = 0.89
V
.
O
2
r = 0.82 31
Mean
counts.min
–1
.activity
–1
(using normal and
‘cycling’modes)
16 boys, 15 girls;
10-16y
7-13 day test-retest;
cycling: R = 0.73-0.74;
treadmill: R = 0.76-0.80
V
.
O
2
r = 0.66 w/ cycling;
r = 0.93 w/ treadmill
32
AEE = activity energy expenditure; Caltrac
®
= uniaxial accelerometer; CARS = Children’s activity rating scale; DLW = doubly labelled water;
DO = direct observation; EE = energy expenditure; FATS = Fargo activity timesampling survey; r = Pearson product-moment correlation
coefficient; R = intraclass correlation coefficient; SEE = sedentary energy expenditure; TEE = total energy expenditure; V
.
O
2
= oxygen
consumption; WEE = waking energy expenditure.
446 Sirard & Pate
Adis International Limited. All rights reserved. Sports Med 2001; 31 (6)
R3D®(ProfessionalProducts,ReiningInternational,
Madison, WI) accelerometers. While the CSA
®
is
a single plane accelerometer, the Tritrac-R3D
®
is 3-
dimensional and may provide a more accurate as-
sessment of physical activity. Eight studies that used
these and other accelerometers with childen and
adolescents are presented in table V.
[29,30,32,53,61-64]
The studies are listed by the type of accelerometer
andthestrengthofthecriterionmeasure.Fairweather
et al.
[61]
reported a relatively high correlation be-
tween direct observation
[16]
and the CSA
®
acceler-
ometer during a preschool exercise class (r = 0.87).
Welk et al.
[63]
also reported promising results (r =
0.70 to 0.77) for the TriTrac-R3D
®
accelerometer
validatedagainstthe CARSdirectobservationsys-
tem.
[13]
Using whole room calorimetry as a crite-
rion measure of EE, Treuth et al.
[64]
assessed the
validity of simultaneously measuring heart rate and
leg accelerometry to estimate EE. Combining the
methodsresultedin valid estimatesofEE(percent-
age error; -2.9 to 5.1%, kJ/d) for both groups and
individual children.
Accelerometers provide an objective, nonreac-
tive and re-useable tool for assessing physical ac-
tivity. Nevertheless, they have a limited ability to
assess cycling, locomotion on a gradient or other
activities with limited torso movement. Also, con-
verting accelerometer counts to units of EE may
provide inaccurate estimates because of the addi-
tional measurement error. The Caltrac
®
device is a
first generation accelerometer and is limited by
possible participant tampering because of the easy
accessibility to its controls, and by its inability to
detect daily or hourly patterns of activity without
participant involvement. The CSA
®
, Tritrac-R3D
®
and other accelerometers are promising devices that
detect both the patterns of physical activity and
total activity, using internal memory with no exte-
rior controls. The benefit of the 2 extradimensions
of measurement in the Tritrac-R3D
®
compared
Table V. Validation of other accelerometers used to assess young people’s physical activity
Monitors Variables Participants Reliability Criterion measure Validity Reference
LSI
®
Counts.h
free-play
–1
18 boys, 12 girls;
2-6y
N/A FATS DO
[20]
r = 0.38 53
CSA
®
Counts.period
–1
11,
mean age = 4.0y
Left hip vs right hip
counts different
(p < 0.05)
CPAF DO
[16]
r = 0.87 61
CSA
®
Mean counts.min
–1
19 boys, 11 girls;
10-14y
Left hip vs right hip;
R=0.87
V
.
O
2
r = 0.77-0.87 62
CSA
®
Tritrac
®
Mean counts.min
–1
15 boys, 15 girls;
8-10y
N/A V
.
O
2
CSA
®
: r = 0.69-0.85;
Tritrac
®
: r = 0.74-0.93
29
CSA
®
Tritrac
®
Mean counts.min
–1
21 Chinese boys;
8-10y
N/A V
.
O
2
CSA
®
: r = 0.81-0.88;
Tritrac
®
: r = 0.93-0.94
30
Tritrac
®
Counts.period
–1
;
(classroom and
PE)
32, 10-12y N/A CARS DO
[13]
Classroom: r = 0.70;
PE: r = 0.77
63
Mini-logger
®
Mean
counts.min
–1
;
(ankle and hip
placements)
16 boys, 15 girls;
10-16y
7-13 day test-retest;
cycling: R = 0.05-0.75;
treadmill: R = 0.61-0.84
V
.
O
2
Cycling: r = 0.06-0.15;
treadmill: r = 0.37-0.67
32
Mini-mitter
2000
®
Mean counts.min
–1
10 boys, 10 girls;
8-12y
N/A 24-hour whole
room calorimetry;
(kJ.day
-1
)
–2.9
± 5.1% 64
CARS = Children’s activity rating scale; CPAF = children’s physical activity form; CSA
®
= uni-axial accelerometer; DO = direct observation;
FATS = Fargo activity timesampling survey; LSI
®
= large scale integrated physical activity monitor; Mini-logger
®
= uni-axial accelerometer;
Mini-mitter
®
=uni-axialaccelerometer;PE=physicaleducation;r=Pearsonproduct-momentcorrelationcoefficient;R=intraclasscorrelation
coefficient; Tritrac
®
= tri-axial accelerometer; V
.
O
2
= oxygen consumption.
Physical Activity Assessment in Children and Adolescents 447
Adis International Limited. All rights reserved. Sports Med 2001; 31 (6)
with the CSA
®
accelerometer has not been well
established.
3. Subjective Techniques
Survey methods of estimating physical activity
levels in children are considered subjective because
they rely on responses from the child. The sporadic
nature of children’s physical activity
[14]
makes these
activities difficult to recall, quantify and categorise.
Also, the lower cognitive functioning of children
compared with adults reduces their ability to accu-
rately recall intensity, frequency and especiallydu-
rationofactivities.
[65,66]
Surveytechniques should be
validated againstmore stringent measures of phys-
icalactivity (primary-orsecondary-levelmethods)
before extensive use. The techniques in this cate-
gory are classified into 4 groups: self-report ques-
tionnaires, interviewer-administered questionnaires,
proxy-report questionnaires and diaries.
3.1 Self-Report Questionnaires
Table VI summarises information about ques-
tionnaires listed by the strength of the criterion mea-
sure.
[58,65,68-75]
The ‘Time Frame’ column in tables
VI–VIII indicates the time frame for which physi-
cal activity was assessed. There is a wide range (r
= –0.10to 0.88) ofcorrelationcoefficientsbetween
these self-report measures and direct observation,
heart rate, or motion detection. Such wide variabil-
ity is indicative of the many different instruments
andcriterion measuresused.StudiesfromtheFam-
ilyHealthProject
[65,67]
aretheonlystudiesthatval-
idatedsurveys againstdirectobservationandfound
agreement ranging from 73.4% in 24 participants
[65]
to 86.3% in 812 participants.
[67]
Craig et al.
[68]
ob-
served a correlation of r = 0.47 between a 1-year
MVPArecall and 2-weeks of EE measured by DLW.
The inclusion of younger children in the Janz et al.
study
[71]
may have lowered the correlations be-
cause of their limited ability to accurately recall their
intensity and duration of physical activity. Also,
compared with adults, young children have lower
sweatratesgiventhesameenvironmentalstress.
[76]
Therefore, it may be inappropriate to use a sweat
recalltoestimate physicalactivityinpreadolescent
children.
Weston et al.
[73]
obtained the highest correla-
tions with the Previous Day Physical Activity Re-
call (PDPAR). The PDPAR was positively associ-
ated with both a pedometer (r = 0.77) and Caltrac
®
accelerometer(r= 0.88) in8th-11th grade students.
In contrast, Trost et al.
[70]
found associations be-
tween the PDPAR and the CSA
®
accelerometer
rangingfromr=0.19to0.39in5thgradechildren.
The lower correlations observed in the Trost et al.
study
[70]
may be because of the much smaller sam-
ple size and the younger age of the children.
Relatively inexpensive self-report measures of-
fer researchers a means of estimating physical ac-
tivity levels in large numbers of individuals while
maintaining low investigator and respondent bur-
den. The greatestlimitation with thesetypesofmeas-
ures is the subjectivity inherent when individuals
are asked to respond to questions about their behavi-
our. The issues of recall errors, deliberate misrep-
resentations,socialdesirabilityandotherbiasesare
particularly important when dealing with children.
Although the PDPAR and several others appear to
be promising tools, more research using primary
standards as criterion measures is needed to clarify
their full potential. Also, the number of administra-
tions needed to estimate ‘usual’physical activity is
not clear for most of the 1-day questionnaires.
3.2 Interviewer-Administered Questionnaires
The results from studies evaluating 7 interviewer-
administered surveys are presented in table
VII.
[31,58,77-79]
Although providing a trained ad-
ministrator may improve a child’s cognition and
accuracy, there is still a wide range of correlations
for these techniques. Wallace and McKenzie
[77]
used
1 week of direct observation as a criterion measure
andfound75%agreement between thisanda 7-day
physicalactivityrecall. Sallisetal.
[78]
obtained rel-
ativelyhigh correlationsusingtheGodin-Shephard
Survey and a simple activity rating compared with
heart rate recordings (r = 0.81 and 0.89, respec-
tively). These authors
[78]
also observed lower cor-
relations with a 7-day recall, which may indicate
448 Sirard & Pate
Adis International Limited. All rights reserved. Sports Med 2001; 31 (6)
the increased difficulty of remembering more than
thepreviousday’sactivities.Asawhole,thesestudies
indicate that 1-day or simpler measures of ‘usual’
physical activity provided greater correlations. The
greater correlations for the 5th graders compared
with 3rd graders in the Simons-Morton et al.
study
[79]
indicatesthat olderchildren may be better
able to complete these types of instruments. Fur-
ther evidence supporting the use of these and other
surveysshouldbeobtainedbyusingdirectobser-
vation or an accelerometer as the criterion mea-
sure.
Table VI. Validation of self-reports used to assess young people’s physical activity
Instruments Time
frame
Participants Reliability Criterion
measure
Validity Reference
6 different forms
a
1 day 24, 3rd-6th grade N/A DO 73.4% agreement across all
forms
65
MVPArecall 1 day 422 boys, 390 girls;
3rd and 4th grades
N/A DO 86.3% agreement between
reported and observed number
of MVPA bouts > 10 minutes
67
1-year physical activity
recall for MVPA
1 year 49 girls;
mean age = 10y
2-week test-retest;
r = 0.70
AEE
DLW
r = 0.47 68
Modifiable activity
questionnaire
1 week 48 boys, 53 girls;
mean age = 5.3y
N/A AEE
DLW
Nonsignificant correlations with
AEE
DLW
69
Previous day physical
activity recall (PDPAR)
1 day 18 boys, 20 girls;
5th grade
N/A CSA
®
r = 0.19-0.39 70
Activity rating Normative
scale
15 boys, 15 girls;
7-15y
1-month test-retest
R=0.85
CSA
®
r = –0.04-0.17 71
3-day aerobic recall 3 day R = 0.54 r = 0.46-0.51
3-day sweat recall 3 day R = 0.30 r = 0.05-0.39
Computerised activity
recall (CAR)
5, 1 day
recalls
20 boys, 25 girls;
6th-8th grade
1-2 week test-retest;
R=0.95forTEE;
R = 0.82 for AEE
1 day
Tritrac
®
r = 0.51 w/ TEE;
r = 0.20 w/ AEE
72
Previous day physical
activity recall (PDPAR)
1 day 119; 8th-11th grade 1-hour test-retest;
R=0.98
pedometer r = 0.77 73
Caltrac
®
r = 0.88
HR r = 0.37-0.63
Self-administered
physical activity
checklist (SAPAC)
1 day 55 boys, 70 girls;
5th grade
N/A Caltrac
®
HR
r = 0.28-0.60;
r = 0.02-0.32
58
Yesterday activity
checklist
1 day 34 boys, 35 girls;
4th grade
3-day test-retest
R=0.60
1 day
Caltrac
®
r = –0.22-0.33 74
Weekly activity sum 1 week R = 0.51 3 day
Caltrac
®
r = –0.15-0.40
Weekly activity
checklist
1 week R = 0.74 3 day
Caltrac
®
r = –0.26-0.34
7-day activity tally 1 week R = 0.68 3 day
Caltrac
®
r = –0.10-0.11
Physical activity
questionnaire for older
children (PAQ-C)
7 day 38 boys, 51 girls;
4th-8th grade
N/A Caltrac
®
r = 0.39 75
a 6 different forms: daily self-monitoring, daily, daily segmented, daily exact, daily dichotomous and daily trichotomous.
AEE = activity energy expenditure; Caltrac
®
= uniaxial accelerometer; CSA
®
= uni-axial accelerometer; DLW = doubly labelled water; DO =
directobservation;HR = heartrate;MVPA =moderatetovigorousphysicalactivity;normative scale=self-assessmentcomparedwithothers
of same age and gender; r = Pearson product-moment correlation coefficient; R = intraclass correlation coefficient; TEE = total energy
expenditure; Tritrac
®
= tri-axial accelerometer.
Physical Activity Assessment in Children and Adolescents 449
Adis International Limited. All rights reserved. Sports Med 2001; 31 (6)
Interviewer-administered surveyspossessmany
of the same strengths and limitations as self-report
measures. An interview format may slightly im-
prove results but the presence of the interviewer
may introduce additional bias. Any potential bene-
fits of this method must be weighed against the
increased cost and burden to the researcher, as well
as the potential for response bias.
3.3 Proxy-Reports
Validation studiesof proxy-reportsofchildren’s
physicalactivityare presentedintableVIII.
[56,80,81]
Noland et al.
[56]
observed little or no correlation
between direct observation and either a teacher’s
oraparent’srating ofthechild’sactivity.Twoother
studies, however,observed significant associations
using either a teacher report
[80]
or a parent report.
[81]
Overall, there is limited information for this type
of physical activity measure in children and ado-
lescents.
Althoughitistemptingtothinkthat parentswould
provide an accurate assessment of their child’s ac-
tivity, this does not always seem to be the case.
[56]
Part of the problem with proxy-reports is the type
of information sought. Questions that assess sub-
jective behaviours (e.g. physical activity) rather than
objectivefacts (e.g.eye colour) mayproducelower
agreement between the criterion measure and the
proxy respondent.
[82]
Also, the characteristics and
perceptions oftheproxy respondent mayintroduce
additionalsourcesofbias.
[83,84]
By using the parent
or teacher as a proxy respondent for young chil-
dren, however, researchers can avoid recall errors
caused by children’s cognitive limitations. Proxy
reports appear promising and wouldbe suitablefor
large study populations if a valid and reliable in-
strument can be developed.
3.4 Diaries
Because of the relatively high participant bur-
den, few studies have used the diary method for
estimating young people’s physical activity. Bou-
chard et al.
[85]
reported associations between a 3-
day activity log and several physiological meas-
ures in 150 children (mean age = 14.6
± 2.9 years)
and 150 adults. For the entire sample (n = 300),
correlations between EE from the diary (TEE
diary
)
and the Physical Work Capacity 170 cycle ergom-
etertestrangedfrom0.23to0.70.TheTEE
diary
was
weakly correlated with the sum of 6 skinfolds (r =
Table VII. Validation of interviewer-administered self-report measures used to assess young people’s physical activity
Instrument Time frame Participants Reliability Criterion
measure
Validity Reference
7-day physical activity recall 7 days 11 boys; 11-13y N/A DO 75% agreement for
intensity
77
Same day recall Previous 10
hours
20 boys, 15 girls; 8-13y 1-day test-retest;
r = 0.06
Caltrac
®
Day 1: r = 0.49;
day 2: r = 0.39
31
HR Day 1: r = 0.25;
day 2: r = 0.52
Physical activity checklist
interview (PACI)
7 days 55 boys, 70 girls; 5th
grade
N/A Caltrac
®
r = 0.22-0.54 58
HR r = 0.10-0.38
7-day recall interview (PAR) 7 days 36 5th grade, 36 8th
grade, 30 11th grade
R = 0.54-0.77 HR r = 0.44-0.53 78
Godin-Shepard survey (GS) 7 days 2-week test-retest
R=0.81
r = 0.81
Simple activity rating Normative
scale
R=0.89 r=0.89
Physical activity interview
(PAI)
1 day 34 3rd grade
30 5th grade
N/A HR r = 0.50-0.57
r = 0.72
79
Caltrac
®
= uniaxial accelerometer; DO = direct observation; HR = heart rate; normative scale = self-assessment compared with others of
same age and gender; r = Pearson product-moment correlation coefficient; R = intraclass correlation coefficient.
450 Sirard & Pate
Adis International Limited. All rights reserved. Sports Med 2001; 31 (6)
-0.08) and percent body fat (r = -0.13). Seliger et
al.
[86]
useda 24-hour diaryin 12-year-old boysand
found increasing heart rates for each of the 7 inten-
sity categories the boys used to evaluate their ac-
tivity. Bratteby et al.
[87]
found a mean difference of
1.2% between TEE
diary
and TEE
DLW
in 15-year-
olds. Garcia et al.
[88]
reported high test-retest reli-
ability (R = 0.94) using the Child/Adolescent Ac-
tivity Log (CAAT). These investigators state that
the CAAT was validated against the Caltrac
®
ac-
celerometer, but these data were not presented.
The activity diary is considered one of the most
accurate subjectivetechniquesforadults.Basedon
theparticipantburdenrequiredtomaintainanac-
tivity diary, however, this technique has limited uses
in a paediatric population. While adolescents may
be able to complete the diary, the accuracy of their
reports should be viewed with caution. It has been
noted that survey methods in children under the
age of 10 years are not advisable
[66]
and the same
limitation should be applied to the use of activity
diaries.
4. Future Research
Survey methods of assessing young people’s
physical activity are very cost effective but lack
objectivity. Many have not been validated against
direct observation or measured EE (DLW or V
.
O
2
),
although most have been validated against some
objective measure. Validation of these methods and
other new instruments against direct observation
would truly assess their validity. National surveys
are used to assess the physical activity of the entire
population and provide a basis for funding of ac-
tivity-related programmes and research. These in-
struments, however, may lack acceptable validity
when they are compared with a more stringent cri-
terion. Such surveys may produce erroneous val-
ues. Therefore, further work is needed to identify
valid and reliable items that are appropriate for in-
clusion into a national survey format. In addition,
there is a lack of valid proxy-report instruments
available for measuring physical activity in chil-
dren. Since this method would be an efficient means
of obtaining physical activity information for young
children, new proxy-report instruments need to be
developed and validated appropriately. The Digi-
walker DW-200pedometercorrelates wellwithshort
termdirectobservation and laboratorymeasuresof
oxygen consumption and heart rate. Further work
is needed to validate these pedometers in more re-
alistic settings.Accelerometersarean attractive tech-
nique for physical activity assessment because of
their objectivity and high validity. More research
is needed on the validity of accelerometers in free-
living children and the possibility of improving their
accuracy by combining accelerometry with either
heart rate or survey techniques.
5. Conclusion
To understand why some young peoplearemore
active than others and how to encourage them to
be more active, we need to measure physical activ-
ity accurately and reliably. Valid methods of esti-
mating physical activity in children and adolescents
are critical to understanding the dose-response re-
Table VIII. Validation of proxy reports used to assess young people’s physical activity
Instrument Time frame Participants Reliability Criterion
measure
Validity Reference
6-item parent survey 1-day 11 boys, 10 girls; 3-5y N/A 20-min video r = -0.19-0.06 56
6-item teacher survey 6-h DO r = -0.13-0.04
Teacher ratings of
activity
1-day 33 boys, 25 girls;
mean age = 2.5y
N/A activity recorder r = 0.41-0.66 80
Teacher report 5-day 17 boys, 22 girls; 6y 2-week test-retest
Spearman r = 0.84
HR Spearman
r = 0.07-0.59
81
Parent report 3-day Spearman r = 0.27-0.53 Spearman
r = 0.72-0.82
DO = direct observation; HR = hear; r = Pearson product-moment correlation coefficient.
Physical Activity Assessment in Children and Adolescents 451
Adis International Limited. All rights reserved. Sports Med 2001; 31 (6)
lationship between physical activity and chronic
diseases and associated risk factors. Accurate knowl-
edgeofphysicalactivitylevelsallowsustodevelop
physical activity intervention programmes and to as-
sess their effectiveness.
Although the idealmethod ofassessingphysical
activity in children (and adults) remains elusive,
direct observationiscurrentlythemost appropriate
criterion standard. When direct observation is not
possiblebecauseof longmeasurementtime periods
or personnel or monetary constraints, accelerome-
ters provide a promising alternative. When possi-
ble, new survey instruments should be validated
against a more stringent technique before they are
widely used. The goal of physical activity research
is to better understand the role of physical activity
in disease and health. Attainment of this goal de-
pends on the sensitivity of the measurement tools.
Technological advances suchasheart ratemonitor-
ing and accelerometry will make this goal increas-
ingly attainable in the near future.
References
1. Caspersen CJ, Powell KE, Christenson GM. Physical activity,
exercise, and physical fitness: definitions and distinctions for
health-related research. Public Health Rep 1985; 100: 126-31
2. Berenson GS, McMahon CA, Voors AW, et al. Cardiovascular
risk factors in children: the early natural history of atheroscle-
rosis and essential hypertension. New York: Oxford Univer-
sity Press, 1980
3. Pate RR, Pratt M, Blair SN, et al. A recommendation from the
Centers forDiseaseControl andPreventionandthe American
College of Sports Medicine. JAMA 1995; 273 (5): 402-7
4. U.S. Department of Health and Human Services. Physical ac-
tivity and health: a report of the surgeon general. Atlanta
(GA): U.S. Department of Health and Human Services, Cen-
ters for Disease Control and Prevention, National Center for
Chronic Disease Prevention and Health Promotion, 1996
5. Bao W, Srinivasan SR,WattigneyWA,etal. Persistence ofmul-
tiple cardiovascular risk clustering related to Syndrome X
from childhood to young adulthood. Arch Intern Med 1994;
154: 1842-7
6. Lauer RM, Connor WE, Learverton PE, et al. Coronary heart
disease risk factors in school children: the Muscatine Study.
J Pediatr 1975; 86 (5): 697-706
7. Williams CL, Carter BJ, Wynder EL. Prevalence of selected
cardiovascular and cancer risk factors in a pediatric popula-
tion: the ‘Know Your Body’ Project, New York. Prev Med
1981; 10: 235-50
8. Grundy SM, Blackburn G, Higgins M, et al. Physical activity
in the prevention and treatment of obesity and its comorbidit-
ies. Med Sci Sports Exerc 1999; 31 (11 Suppl.): 502-8
9. Hill JO, Trowbridge FL. Childhood obesity: future directions
and research priorities. Pediatrics 1998; 101: 570-4
10. Laporte RE, Montoye HJ, Caspersen CJ. Assessment of physi-
cal activity in epidemiologic research: problems and pros-
pects. Public Health Rep 1985; 100: 131-46
11. Katzmarzyk PT, Malina RM, Song TM, et al. Physical activity
and health-related fitness in youth: a multivariate analysis.
Med Sci Sports Exerc 1998; 30 (5): 709-14
12. MacAuley D, McCrum EE, Stott G, et al. Levels of physical
activity, physicalfitness and their relationshipin the Northern
Ireland Health and Activity Survey. Int JSports Med 1998; 19
(7): 503-11
13. Puhl J, Greaves KA, Hoyt M, et al. Children’s activity rating
scale (CARS): description and evaluation. ResQ Exerc Sport
1990; 61 (1): 26-36
14. Bailey RC, Olson J, Pepper SL, et al. The level and tempo of
children’s physical activities: an observationalstudy. Med Sci
Sports Exerc 1995; 27 (7): 1033-41
15. EpsteinLH, McGowanC,WoodallK. Abehavioralobservation
system for free play activity inyoung overweight female chil-
dren. Res Q Exerc Sport 1984; 55 (2): 180-3
16. O’Hara NM, Baranowski T, Simons-Morton BG, et al. Validity
of the observationof children’s physical activity. ResQ Exerc
Sport 1989; 60 (1): 42-7
17. McKenzie T, Sallis JF, Patterson T, et al. BEACHES: an obser-
vational system for assessing children’s eating and physical
activity behaviors and associated events. Appl Behav Anal
1991; 24 (1): 141-51
18. Rowe PJ, Schuldheisz JM, van der Mars H. Validation of SOFIT
for measuring physical activity of first- to eighth-grade stu-
dents. Pediatr Exerc Sci 1997; 9 (2): 136-49
19. McKenzie T, Sallis JF, Nader PR. SOFIT: system for observing
fitnessinstruction time. J Teach Phys Educ1992;62:195-205
20. Klesges RC, Coates TJ, Moldenhauer-Klesges L, et al. The
FATS: an observational system for assessing physicalactivity
in children and associated parent behavior. Behav Assess
1984; 6: 333-45
21. McKenzie T. Observational measures of children’s physical ac-
tivity. J Sch Health 1991; 61 (5): 220-3
22. Schoeller DA, Ravussin E, Schutz Y, et al. Energy expenditure
by doubly labelled water: validation in humans and proposed
calculation. Am J Physiol 1986; 250 (5 Pt 2): R823-30
23. KleinPD, James WPT,Wong WW,etal. Calorimetric validation
of the doubly-labelled water method for determination of en-
ergy expenditure in man. Human Nutr Clin Nutr 1984; 38:
95-106
24. Schoeller DA, Webb P. Five-day comparison of the doubly la-
belled water method with respiratory gasexchange. Am JClin
Nutr 1984; 40: 153-8
25. Jones PJ, Winthrop AL, Schoeller DA, et al. Validation of dou-
bly labelledwaterfor assessingenergy expenditurein infants.
Pediatr Res 1987; 21 (3): 242-6
26. Goran MI, Carpenter WH, Poehlman ET. Total energy expen-
diture in 4- to 6-yr-old children. Am J Physiol 1993; 264 (5
Pt 1): E706-11
27. Bitar A, Vermorel M, Fellmann N, et al. Heart rate recording
method validated by whole body indirect calorimetry in 10-
yr-old children. J Appl Physiol 1996 Sep; 81 (3): 1169-73
28. Emons HJ, Groenenboom DC, Westerterp KR, et al. Compari-
son of heart rate monitoring combined with indirect calorim-
etry and the doubly labelled water (2H2(18)O) method forthe
measurement of energy expenditure in children. Eur J Appl
Physiol 1992; 65 (2): 99-103
452 Sirard & Pate
Adis International Limited. All rights reserved. Sports Med 2001; 31 (6)
29. Eston RG, Rowlands AV, Ingledew DK. Validity of heart rate,
pedometry, and accelerometry for predicting the energy cost
of children’s activity. J Appl Physiol 1998; 84 (1): 362-71
30. Louie L, Eston RG, Rowlands AV, et al. Validity of heart rate,
pedometry, and accelerometry for estimating the energy cost
of activity in Hong Kong Chinese boys. Pediatr Exerc Sci
1999; 11 (3): 229-39
31. Sallis JF, Buono MJ, Roby J, et al. The Caltrac accelerometer
as a physical activity monitor for school-age children. Med
Sci Sports Exerc 1990; 22 (5): 698-703
32. TroutmanSR,AllorKM,HartmannDC,etal.MINI-LOGGER-
reliability and validity for estimating energy expenditure and
heart rate in adolescents. Res Q Exerc Sport 1999 Mar; 70
(1): 70-4
33. Freedson PS.Electronic motion sensors and heartrate as meas-
ures of physical activity in children. J Sch Health 1991; 61
(5): 215-9
34. Livingstone MB, Coward WA, Prentice AM, etal. Daily energy
expenditure in free-living children: comparison of heart-rate
monitoring with the doubly labelled water (2H2(18)O)
method. Am J Clin Nutr 1992 Aug; 56 (2): 343-52
35. Maffeis C, Pinelli L, Zaffanello M, et al. Daily energy expen-
diture in free-living conditions in obese and non-obese chil-
dren: comparison of doubly labelled water (2H2(18)O)
method and heart-rate monitoring. Int J Obes Relat Metab
Disord 1995 Sep; 19 (9): 671-7
36. Van den Berg-EmonsRJ, SarisWH, Westerterp KR, etal.Heart
rate monitoring to assess energy expenditure in children with
reduced physical activity. MedSciSports Exerc 1996Apr;28
(4): 496-501
37. Ceesay SM, Prentice AM, Day KC, et al. The use of heart rate
monitoring in the estimation of energy expenditure: a valida-
tion study using indirect whole-body calorimetry. Br J Nutr
1989; 61 (2): 175-86
38. Livingstone MBE, Prentice AM, Coward WA, et al. Simulta-
neous measurement of free-living energy expenditure by the
doubly labelled water method and heart-rate monitoring. Am
J Clin Nutr 1990; 52: 59-65
39. Armstrong N, Balding J, Gentle P, et al. Patterns of physical
activity among 11 to 16 year old British children. BMJ 1990;
301: 203-5
40. Gilbey H,GilbeyM. Thephysical activity ofSingapore primary
school children as estimatedby heartratemonitoring. Pediatr
Exerc Sci 1995; 7: 26-35
41. Gilliam TB, Freedson PS, Geenen DL, et al. Physical activity
patterns determined by heart rate monitoring in 6-7 year-old
children. Med Sci Sports Exerc 1981; 13 (1): 65-7
42. Janz KF, Golden JC, Hansen JR, et al. Heart rate monitoring of
physical activity in children and adolescents: the Muscatine
Study. Pediatrics 1992; 89: 256-61
43. Sallo M, Silla R. Physical activity with moderate to vigorous
intensity in preschool and first-grade schoolchildren. Pediatr
Exerc Sci 1997; 9: 44-54
44. Karvonen J, Vuorimaa T. Heart rate and exercise intensity dur-
ing sports activities: practical application. Sports Med 1988;
5 (5): 303-11
45. Simons-Morton BG, Parcel GS, O’Hara NM, et al. Health-
related physical fitness in childhood: status and recommen-
dations. Ann Rev Public Health 1988; 9: 403-25
46. Allor KM, Pivarnik JM. Use of heart rate cutpoints to assess
physical activity intensity in sixth-grade girls. Pediatr Exerc
Sci 2000; 12 (3): 284-92
47. BassettDRJ,AinsworthBE,Leggett SR, etal. Accuracy of five
electronic pedometers for measuring distance walked. Med
Sci Sports Exerc 1996 Aug; 28 (8): 1071-7
48. Bassey EJ, Dallosso HM, Fentem PH, et al. Validation of a
simplemechanicalaccelerometer (pedometer) for theestima-
tion of walking activity. Eur J Appl Physiol 1987; 56 (3):
323-30
49. Sequeira MM, Rickenbach M, Wietlisbach V, et al. Physical
activity assessment using a pedometer and its comparison
with a questionnaire in a large population survey. Am J
Epidemiol 1995 Nov 1; 142 (9): 989-99
50. Saris WH, Binkhorst RA. The use of pedometer and actometer
in studying daily physical activity in man. Pt II: validity of
pedometer and actometer measuring the daily physical activ-
ity. Eur J Appl Physiol 1977 Oct 31; 37 (3): 229-35
51. Kilanowski CK, Consalvi AR, Epstein LH. Validation of an
electronic pedometer for measurement of physical activity in
children. Pediatr Exerc Sci 1999; 11 (1): 63-8
52. Westerterp KR. Physical activity assessment with accelerome-
ters. Int J Obes Relat Metab Disord 1999 Apr; 23 Suppl. 3:
S45-9
53. Klesges RC, Klesges LM, Swenson AM, et al. A validation of
two motion sensors in the prediction of child and adult phys-
ical activity levels. Am J Epidemiol 1985 Sep; 122 (3):
400-10
54. Klesges LM, Klesges RC. The assessment of children’s physi-
cal activity: a comparison of methods. Med Sci Sports Exerc
1987 Oct; 19 (5): 511-7
55. Mukeshi M, Gutin B, Anderson WA, et al. Validation of the
Caltrac movement sensor using direct observation in young
children. Pediatr Exerc Sci 1990; 2 (3): 249-54
56. Noland M, Danner F, DeWalt K, et al. The measurement of
physical activity in young children. Res Q Exerc Sport 1990
Jun; 61 (2): 146-53
57. Bray MS, Wong WW, Morrow JRJ, et al. Caltrac versus calo-
rimeter determination of 24-h energy expenditure in female
children and adolescents.Med Sci SportsExerc1994Dec;26
(12): 1524-30
58. Sallis JF, Strikmiller PK, Harsha DW, et al. Validation of inter-
viewer- and self-administered physical activity checklists for
fifth grade students. Med Sci Sports Exerc 1996; 28 (7):
840-51
59. Johnson RK, Russ J,GoranMI.Physical activityrelated energy
expenditurein childrenby doubly labelledwater as compared
withthe Caltracaccelerometer. IntJObesRelatMetabDisord
1998 Nov; 22 (11): 1046-52
60. Ballor DL, Burke LM, Knudson DV, etal.Comparison of three
methods of estimatingenergy expenditure: caltrac,heart rate,
and video analysis. Res Q Exerc Sport 1989 Dec; 60 (4):
362-8
61. Fairweather SC, Reilly JJ, Grant S, et al. Using the Computer
Science andApplications (CSA)activitymonitorinpreschool
children. Pediatr Exerc Sci 1999; 11: 413-20
62. Trost SG, Ward DS, Moorehead SM, et al. Validity of the Com-
puter Science and Applications (CSA) activity monitor in
children. Med Sci Sports Exerc 1998; 30 (11): 629-33
63. Welk GJ, Corbin CB, Kampert JB. The validity of the Tritrac-
R3D activity monitor for the assessment of physical activity:
II. Temporal relationships among objective assessments. Res
Q Exerc Sport 1998 Dec; 69 (4): 395-9
64. Treuth MS, Adolph AL, Butte NF. Energy expenditure in chil-
dren predicted from heart rate and activity calibrated against
Physical Activity Assessment in Children and Adolescents 453
Adis International Limited. All rights reserved. Sports Med 2001; 31 (6)
respiration calorimetry. Am J Physiol 1998 Jul; 275 (1 Pt 1):
E12-8
65. Baranowski T, Dworkin RJ, Cieslik CJ, et al. Reliability and
validity of self report of aerobic activity: Family Health Proj-
ect. Res Q 1984; 55 (4): 309-17
66. Sallis JF. Self-report measures of children’s physical activity. J
Sch Health 1991; 61 (5): 215-9
67. Simons-Morton BG, O’Hara NM, Parcel GS, et al. Children’s
frequency of participation in moderate to vigorous physical
activities. Res Q Exerc Sport 1990; 61 (4): 307-14
68. Craig SB, Bandini LG, Lichtenstein AH, et al. The impact of
physical activity on lipids, lipoproteins, and blood pressure in
preadolescent girls. Pediatrics 1996 Sep; 98 (3 Pt 1): 389-95
69. Goran MI, Hunter G, Nagy TR, et al. Physical activity related
energyexpenditureand fat massin young children. Int JObes
1997; 21: 171-8
70. Trost SG, Ward DS, McGraw B, et al. Validity of the Previous
Day Physical Activity Recall (PDPAR) in fifth-grade chil-
dren. Pediatr Exerc Sci 1999; 11: 341-8
71. Janz KF, Witt J, Mahoney LT. The stability of children’s phys-
ical activity as measured by accelerometry and self-report.
Med Sci Sports Exerc 1995; 27 (9): 1326-32
72. McMurray RG, Harrell JS, Bradley CB, et al. Comparison of a
computerized physical activity recall with a triaxial motion
sensor in middle-school youth. Med Sci Sports Exerc 1998
Aug; 30 (8): 1238-45
73. Weston AT, Petosa R, Pate RR. Validity of an instrument for
measurement of physical activity in youth. Med Sci Sports
Exerc 1997; 29 (1): 138-43
74. Sallis JF, Condon SA, Goggin KJ, et al. The development of
self-administered physical activity surveys for 4th grade stu-
dents. Res Q Exerc Sport 1993; 64 (1): 25-31
75. Kowalski KC, Crocker PRE, Faulkner RA. Validation of the
physical activity questionnaire for older children. Pediatr Ex-
erc Sci 1997; 9 (2): 174-86
76. Falk B, Bar-Or O, Calvert R, et al. Sweat gland response to
exercise in the heat among pre-, mid-, and late-pubertal boys.
Med Sci Sports Exerc 1992; 24: 313-9
77. Wallace JP, McKenzie TL. Observed vs. recalled exercise be-
havior: a validation of a seven day exercise recall for boys 11
to 13 years old. Res Q Exerc Sport 1985; 56 (2): 161-5
78. Sallis JF, Buono MJ, Roby J, et al. Seven-day recall and other
physicalactivityself-reportsinchildrenandadolescents. Med
Sci Sports Exerc 1993; 25 (1): 99-108
79. Simons-Morton BG, Taylor WC, Wei Huang I. Validity of the
physical activity interview and Caltrac with preadolescent
children. Res Q Exerc Sport 1994; 65 (1): 84-8
80. Halverson Jr CF, Waldrop MF. The relations of mechanically
recorded activitylevel tovarieties ofpreschoolplaybehavior.
Child Develop 1973; 44: 678-81
81. Manios Y, Kafatos A, Markakis G. Physical activity of 6-year-
old children: Validation of two proxy reports. Pediatr Exerc
Sci 1998; 10 (2): 176-88
82. Whiteman D, Green A. Wherein lies the truth? Assessment of
agreement between parent proxy and child respondents. Int J
Epidemiol 1997 Aug; 26 (4): 855-9
83. Murphey JK, Alpert BS, Christman JV, et al. Physical fitness in
children: a survey method based on parental report. Am J
Public Health 1988; 78 (6): 708-10
84. Rajmil L, Fernandez E, Gispert R, et al. Influence of proxy
respondents in children’s health interview surveys. J
Epidemiol Community Health 1999 Jan; 53 (1): 38-42
85. Bouchard C, Tremblay A, LeBlanc C, et al. A method to assess
energy expenditure in children and adults. Am J Clin Nutr
1983 Mar; 37 (3): 461-7
86. Seliger V, Trefny Z, Bartunkova S, et al. The habitual activity
and physical fitness of 12 year old boys. Acta Paediatr Belg
1974; 28: 54-9
87. Bratteby LE, Sandhagen B, Fan H, et al. A 7-day activity diary
for assessment of daily energy expenditure validated by the
doubly labelled water method in adolescents. Eur J Clin Nutr
1997 Sep; 51 (9): 585-91
88. Garcia AW, Pender NJ, Antonakos CL, et al. Changes in phys-
ical activity beliefs and behaviors of boys and girls across the
transition to junior high school. J Adolesc Health 1998; 22:
392-402
Correspondence and offprints: John R. Sirard, Department
of Exercise Science, 1300 Wheat Street, Blatt PE Center, Uni-
versity of South Carolina, Columbia, SC 29208, USA.
E-mail: sirard@hotmail.com
454 Sirard & Pate
Adis International Limited. All rights reserved. Sports Med 2001; 31 (6)
