Comparison of Self-reported versus

Accelerometer-Measured Physical Activity

SINDRE M. DYRSTAD1, BJKRGE H. HANSEN2, INGAR M. HOLME2, and SIGMUND A. ANDERSSEN2
1
Department of Education and Sport Science, University of Stavanger, Stavanger, NORWAY; and 2Department of Sports
Medicine, Norwegian School of Sport Sciences, Oslo, NORWAY

ABSTRACT
DYRSTAD, S. M., B. H. HANSEN, I. M. HOLME, and S. A. ANDERSSEN. Comparison of Self-reported versus Accelerometer-
Measured Physical Activity. Med. Sci. Sports Exerc., Vol. 46, No. 1, pp. 99–106, 2014. Introduction: The International Physical Activity
Questionnaire (IPAQ) is one of the most widely used questionnaires to assess physical activity (PA). Validation studies for the IPAQ
have been executed, but still there is a need for studies comparing absolute values between IPAQ and accelerometer in large population
studies. Purpose: To compare PA and sedentary time from the self-administered, short version of the IPAQ with data from ActiGraph
accelerometer in a large national sample. Methods: A total of 1751 adults (19–84 yr) wore an accelerometer (ActiGraph GT1M) for
seven consecutive days and completed the IPAQ–Short Form. Sedentary time, total PA, and time spent in moderate to vigorous activity
were compared in relation to sex, age, and education. Results: Men and women reported, on average, 131 minIdj1 (SE = 4 minIdj1) less
sedentary time compared with the accelerometer measurements. The difference between self-reported and measured sedentary time
and vigorous-intensity PA was greatest among men with a lower education level and for men 65 yr and older. Although men re-
ported 47% more moderate to vigorous physical activity (MVPA) compared with women, there were no differences between sexes
in accelerometer-determined MVPA. Accelerometer-determined moderate PA was reduced from 110 to 42 minIdj1 (62%) when ana-
lyzed in blocks of 10 min (P G 0.0001) compared with 1-min blocks. The main correlation coefficients between self-reported variables
and accelerometer measures of physical activity were between 0.20 and 0.46. Conclusions: The participants report through IPAQ–Short
Form more vigorous PA and less sedentary time compared with the accelerometer. The difference between self-reported and
accelerometer-measured MVPA increased with higher activity and intensity levels. Associations between the methods were affected
by sex, age, and education, but not body mass index. Key Words: IPAQ, ADULTS, MET-MINUTES, MONITORING, NATIONAL
SAMPLE, PUBLIC HEALTH

uestionnaires are the most commonly used method
to assess physical activity (PA) at population level,
and a diversity of questionnaires is available for
this purpose. A review of PA questionnaires for adults states
that although no conclusion could be drawn regarding the
best available questionnaire (32), the International Physical
Activity Questionnaire (IPAQ) appeared to be the most
widely used PA questionnaire. The IPAQ was developed by
a multinational working group as a tool suitable for as-
sessing population levels of PA across countries. There are
two versions of the IPAQ—the IPAQ–Short Form (IPAQ-S)
and the IPAQ–Long Form. Both can be administered by
telephone interview or self-administered. The forms were
Address for correspondence: Sindre M. Dyrstad, PhD, Department of Edu-
cation and Sport Science, University of Stavanger, 4036 Stavanger, Norway;
E-mail: sindre.dyrstad@uis.no.
Submitted for publication November 2012.
Accepted for publication June 2013.
0195-9131/14/4601-0099/0
MEDICINE & SCIENCE IN SPORTS & EXERCISEÒ
Copyright Ó 2013 by the American College of Sports Medicine
DOI: 10.1249/MSS.0b013e3182a0595f
EPIDEMIOLOGY
developed to be used among 18- to 65-yr-old adults in
diverse settings. In the IPAQ-S, participants report the fre-
quency and duration of vigorous, moderate, and walking
activities as well as the time spent sitting during the last 7 d.
The IPAQ instrument has gained wide acceptance, and the
short version in particular has been used in many interna-
tional studies (3,4,28).
The first reliability and validity study of the IPAQ-S was
conducted across 12 countries in year 2000. It demonstrated
reasonable test–retest reliability and intermethod validity (9).
The criterion validity against accelerometer had a pooled
correlation coefficient of 0.30, but large differences were
found between countries. Some studies have reported that the
IPAQ-S may overestimate total PA (7,27). Low test–retest
reliability for the IPAQ-S telephone interview was also found
by the European Physical Activity Surveillance System (26).
Fogelholm et al. (11) validated the IPAQ-S against fitness
and found that almost 10% of the young men participating
in the study had poor fitness and apparently low PA, but
they reported very high PA on the IPAQ. They concluded
that there was an evident need to develop the IPAQ further
to solve the apparent overreporting by a considerable pro-
portion of sedentary individuals.
A total of 23 validation studies were included in a review
of the validity of the IPAQ-S (21). The correlation between

99

Copyright © 2013 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
 total PA level measured by the IPAQ-S and objective
measurements varied widely and was found to be lower than
the acceptable standard. Because few studies had evaluated
the concordance of the absolute values between the IPAQ
and an accelerometer, the authors recommended further val-
idation studies. Investigations of the validity of the IPAQ
in different populations as well as further explorations of sex,
age, socioeconomic and regional differences were called for
by Craig et al. (9). Despite a large number of evaluation
studies of the IPAQ in the interim, Lee et al. (22) repeated
this call, and van Poppel et al. (32) stated the importance of
researchers assessing the measurement properties of the ques-
tionnaire in their own language and in their own target pop-
ulation. To examine some of these questions, researchers
require a large number of participants from a representative
sample of the population. To date, none of the IPAQ-S reli-
ability and validity studies has involved such a large sample,
and the Norwegian version of IPAQ-S has not been validated
except for men age 20–39 yr (20). The purpose of the pres-
ent study was therefore to compare PA and sedentary time
from the self-administered short version of the IPAQ with
an objective measure of PA using an accelerometer, sup-
plemented with data on sex, age, and education in a large
Norwegian sample.
METHODS
The study design, dropouts, and assessment of PA have
been described in detail elsewhere (17). In brief, this multi-
center study involved 10 regional test centers throughout
Norway. A representative sample of 11,515 adults (age
EPIDEMIOLOGY
20–84 yr) from the areas surrounding each test center was
drawn from the Norwegian population registry. Written in-
formed consent was obtained from 3867 participants (34%).
The participants received a preprogrammed accelerometer
and the questionnaire by mail and wore the accelerometer
for seven consecutive days. After the registration period, the
participants returned the accelerometer and questionnaire
by prepaid express mail. Of the 3238 participants with
IPAQ data and valid accelerometer recordings, 2462 an-
swered most of the IPAQ questionnaire, whereas a total of
1751 subjects had a complete set of IPAQ data. Partici-
pants 65 yr and older were overrepresented among those
with missing IPAQ data. The study was approved by
the Regional Ethics Committee for Medical Research, the
Norwegian Social Science Data Services AS, and the
Norwegian Tax Department.
Objective PA measurement. The ActiGraph GT1M
(ActiGraph, LLC, Pensacola, FL) was used to assess par-
ticipants’ PA levels. Participants with at least 4 d of at least
10 h of daily recordings were included in the analysis. Data
were collected in 10-s epochs, which were collapsed into
60-s epochs for comparison with other studies. The data
were reduced using an SAS-based macro (SAS Institute Inc.,
Cary, NC). Wear time was defined by subtracting nonwear
time from 18 h because all data between 12:00 a.m. and
100 Official Journal of the American College of Sports Medicine
Copyright © 2013 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
6:00 a.m. were excluded to avoid potential bias due to par-
ticipants forgetting to remove the monitor when going to bed
at night. Nonwear time was defined as intervals of at least
60 consecutive minutes with zero counts, with allowance
for 1 min with counts greater than zero (30).
Participants accumulating a minimum average of 30 min
of daily moderate-intensity PA in bouts of 10 min or more
(with allowance for interruptions of 1–2 min) were catego-
rized as being sufficiently active (5). This definition allowed
participants to have longer bouts of activity on certain days
and to be less active on other days, but to still be categorized
as sufficiently active.
Counts per minute is a measure of total PA and was ex-
pressed as the total number of registered counts for all
valid days divided by wearing time. To identify PA of dif-
ferent intensities, count thresholds corresponding to the en-
ergy cost of the given intensity were applied to the data set.
Sedentary time was defined as all activities less than 100 cpm,
a threshold that corresponds with sitting, reclining, or lying
down (18,24). Low-intensity PA was defined as between
100 and 759 cpm. Time in moderate-intensity was defined as
between 760 and 5998 cpm (16,23) (tasks, 3–6 METs) and
5999 cpm or more for vigorous intensity (Q6 METs) (30).
Because 760 cpm provides a useful cut point of moderate-
intensity activities in daily life, included walking, the self-
reported walking and the moderate-intensity PA from IPAQ
were merged when compared with moderate-intensity PA
from the accelerometer. Mean minutes per day at different
intensities was determined by summing all minutes where the
count met the criterion for that intensity, divided by the
number of valid days.
Questionnaire. Self-reported PA over the previous 7 d
was obtained by a Norwegian version of the short, self-
administered version of the IPAQ. Additional questions in-
cluding age, anthropometry, exercise habits, health status,
income, and education were included in the questionnaire.
Participants also reported the type of PA they most com-
monly participated in. Age was categorized into four levels:
20–34, 35–49, 50–64, and 65–84 yr. Body mass index
(BMI) was computed as body weight divided by height
squared (kgImj2). Overweight and obesity were defined as
a BMI of 25–29 and Q30 kgImj2, respectively (34). Edu-
cational attainment was categorized into four groups: less
than high school, completed high school, less than 4 yr of
university/college, and university/college education lasting
4 yr or more.
The IPAQ addresses PA performed for at least 10 min
and time spent at three intensities: walking, moderate, and
vigorous. Examples of activity commonly performed in the
different intensities were mentioned in the items. Sitting
time, which also included lying down to watch television,
was expressed as minutes per day. Data within the different
intensities were summed to estimate the total amount of
time spent in PA per week or day. Total daily PA in MET-
minutes per day was estimated by summing the product of
reported time within each intensity by a MET value specific
http://www.acsm-msse.org
 to each category of PA and expressed as a daily average

T 14.9
T 14.6

(36.9)
(11.8)

(13.5)
(37.3)
(23.6)
(25.6)
(100)

(100)
T 9.0
T 3.9
MET score according to the official IPAQ scoring protocol

All
(19). Vigorous-intensity PA was assumed to correspond to

369

433

760
824
3238
49.4
76.5

25.5
1151

3215

1198
173.0
8 METs, moderate-intensity activity to 4 METs, and walk-

All Included, N = 3238

ing to 3.3 METs. The participants were categorized into
three PA levels according to the IPAQ scoring criteria:

T 14.9
T 12.1
(53.6)

(29.2)
(10.6)

(13.0)
(35.5)
(25.0)
T 6.1
T 4.2

(26)
Women
1734
48.5
69.1

24.9
166.9

485
176

225
615
434
451
Low: meets neither moderate nor high criterion.
Moderate: meets any of the following three criteria: (a) three
or more days of vigorous-intensity activity of at least
20 minIdj1, (b) five or more days of moderate-intensity
T 14.8
T 12.5
(46.4)

(45.6)
(13.2)

(13.8)
(38.8)
(21.7)
(24.8)
T 6.5
T 3.5
activity and/or walking of at least 30 minIdj1, and (c)
Men
1504

180.0

666
193

208
583
326
373
50.3
84.9

26.2

five or more days of any combination of walking,

moderate-intensity, or vigorous-intensity activities achiev-
ing a minimum of at least 600 METIminIwkj1.
T 15.6
T 14.5

(36.0)
(12.6)

(18.7)
(43.0)
(19.9)
(18.4)
(100)

(100)
T 9.0
T 4.1

High: meets any one of the following two criteria: (a)

All

vigorous-intensity activity on at least 3 d and accumu-

Participants with Incomplete IPAQ Data

1487

1473
172.0

515
181

275
634
293
271
50.7
75.7

25.5

lating at least 1500 METIminIwkj1 and (b) seven or

more days of any combination of walking, moderate-
intensity, or vigorous-intensity activities accumulating
Set, n = 1487

T 15.7
T 12.0
(57.9)

(29.8)
(11.1)

(18.1)
(40.2)
(21.9)
(19.8)
T 6.2
T 4.4

at least 3000 METIminIwkj1.

Women

92
861

166.7

246

155
344
187
169
50.4
69.2

25.0

Those categorized as moderate or high activity were clas-

sified as being sufficiently active according to PA guidelines.
Data cleaning and processing were carried out in accordance
T 15.6
T 12.9
(42.1)

(44.5)
(14.7)

(19.4)
(46.9)
(17.2)
(16.5)
T 6.7
T 3.5

with the guidelines published by the IPAQ Research Com-

Men

mittee, and the methods used to score the IPAQ are described
626

269

120
290
106
102
51.0
84.5

26.2

89
179.3

in the IPAQ scoring protocol. Of those with valid acceler-

ometer data, 75% answered most of the IPAQ questions,
whereas 54% had valid answers to all IPAQ questions.
T 14.1
T 14.7
Participants with Complete IPAQ Data Set, n = 1751

(37.7)
(11.1)

(32.4)
(26.8)
(31.7)
(100)

(100)
T 9.1
T 3.8

(9.1)

Statistics. Variable values are presented as proportion,

All

mean T SD, or mean T SE. Independent-sample t-test or one-

636
188

158
564
467
553
48.2

25.4
1751

77.2

1742
173.8
TABLE 1. Characteristics of participants, classified into those who have complete and incomplete data sets.

EPIDEMIOLOGY
way ANOVA with the Tukey post hoc tests were used to
test associations of anthropometric or other continuous var-
iables with sex and age. A multivariate general linear model,
T 13.9
T 12.1
(49.1)

(28.7)
(10.1)

(31.1)
(28.4)
(32.4)
T 5.9
T 4.1

(8.0)
Women

adjusted for age and sex, was used to assess educational

873

239

271
247
282
46.8
69.0

24.7

84

70
167.1

differences between the self-reported and the objectively

measured PA. Spearman’s correlation coefficient, Q (rho),
was used to assess associations between the two methods.
T 14.2
T 12.3
(50.1)

(46.5)
(12.2)

(10.1)
(33.6)
(25.2)
(31.1)

The strength of agreement between the two methods was

T 6.3
T 3.4
Men

assessed using the Bland–Altman technique (6). Sedentary time

49.7
85.2

26.1

88
180.5
878

397
104

293
220
271

and walking + moderate to vigorous physical activity (MVPA)

from the IPAQ were compared with sedentary time and
MVPA from the ActiGraph in these analyses. The ability of the
Overweight, n (valid %) (BMI = 25.0j29.9 kgImj2)

IPAQ to identify not sufficiently active individuals was used

to measure the questionnaires sensitivity (10). A P value of
less than 0.05 was regarded as statistically significant. All sta-
Data are presented as mean T SD or n (%).
Obesity, n (valid %) (BMI Q 30 kgImj2)

tistical analyses were performed using PASW Statistics 18 for

Windows (IBM Corporation, Somers, NY).
Educational level, n (valid %)

University or college G4 yr
University or college Q4 yr
Less than high school

RESULTS
The physical characteristics and education levels of
High school

participants are presented in Table 1. The age span was

BMI, kgImj2
Height, cm
Sex, n (%)

Weight, kg

20–84 yr. No sex difference in level of education was

Variable

Age, yr

observed. Overall, 36% of the participants were overweight

and 11% were defined as obese.

SUBJECTIVE AND OBJECTIVE MEASURE OF PA Medicine & Science in Sports & Exercised 101

Copyright © 2013 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
 The descriptive data for the IPAQ and the accelerometer

(95% Confidence Interval)

are shown in Table 2. There were no significant differences
All, N = 1751, Mean

(8072–8340)
between men and women in total PA measured by acceler-

(878–884)

(547–555)
(214–220)
(108–112)

(339–352)

(412–429)

(315–346)
(2.4–3.0)

(28–31)
(24–28)
(15–17)
(69–75)
ometer counts per minute, but men reported a total activity
level (METIminIdj1) 25% higher than women. The self-
8206

30
26
16
72
2.7
881

551
217
110

345

420

330
reported vigorous-intensity PA were 67% higher in men
compared with women, whereas no significant difference in
accelerometer-measured vigorous PA was found (Table 2).
96**

Accelerometer-measured moderate PA where reduced from

3**
2**

6**
1**
1**
1**
2**
9**
0.2
Women,
n = 873

4 110 to 42 min (95% confidence interval = 41–43), a re-

T
T
T
T
T
T
T

T
T
T
T
T
T
33
22
12
67
2.4
878

535
233
108

347

401

293
8462

duction by 62% (P G 0.0001), when only time in 10-min

blocks with an allowance for interruptions of 2 min was
included. Using 2020 cpm, and not 760 cpm, as the cut point
0.2
97

12
n = 878

3
2
2

7
1
1
1
2
Men,

for moderate-intensity PA would have excluded 77 min of

T
T
T
T
T
T
T

T
T
T
T
T
T
26
30
20
76
2.9
882

565
202
112

344

439

367
7952

PA from the accelerometer data. When analyzing these data

in 10-min blocks, self-reported moderate-intensity (walking
College/University,

not included) PA was 70% higher than corresponding ac-

0.2e

10e
n = 1020

90

celerometer data.
2e
2e
1e

6e
6e
1e

2e
2

1
T
T
T
T
T
T
T

T
T
T
T
T
T

Analyses comparing the concordance of absolute values

2.9
883

564
210
106

344

455

300
8234

28
22
15
65
Education Levelb

between the IPAQ and accelerometer data showed that both

men and women in all age groups reported less sedentary
time, less moderate-intensity, and higher level of vigorous-
High School or
Less, n = 722

107

0.2

intensity PA compared with the accelerometer data (Table 3;

12
2

3
2
2

7
1
1
1
3
TABLE 2. Descriptive PA data (mean T SE) from objective measures and self-reported variables for participants with valid data in all categories.

T
T
T
T
T
T
T

T
T
T
T
T
T

P G 0.0001 for all tests, analyzed in 1-min blocks). The

877

533
227
115

348

372

372
8181

32
32
17
82
2.2

Bland–Altman plot (Figs. 1 and 2) illustrates the agreement

between the two methods and shows that the mean differ-
ence was j54 and 13 minIdj1 for moderate- and vigorous-
184d

12d
0.4

21
n = 234

4d

3d

9d
5
3

2
3
2
4

intensity PA, respectively. The corresponding data for

65+,

T
T
T
T
T
T
T

T
T
T
T
T
T

sedentary time were j131 minIdj1 (SD = 166 minIdj1; figure

1.1
850

552
202

302

342

368
7242

94

39
29
15
84

not shown). Participants age 20–34 yr reported 25% less

sedentary time than measured by the accelerometer. The
119

0.3

14
2d
n = 564
50–64,

8
2
2
1
3
EPIDEMIOLOGY

corresponding value for men and women older than 65 yr

T
T
T
T
T
T
T

T
T
T
T
T
T
Age Groups (yr)a

34
29
15
78
2.0
893

556
219
116

349

431

348
8491

was 61% (P G 0.0001; Table 3). There were no significant

differences in accelerometer-measured sedentary time in
any age groups. The difference between self-reported and
113

13d
0.2
2d
n = 619
35–49,

7
1
2
1
3

objectively measured sitting time was 98% larger among

Different from participants with an education level of high school or less and, P G 0.05.
T
T
T
T
T
T
T

T
T
T
T
T
T
24
24
15
63
3.1
890

548
225
114

354

431

295
8394

men with high school or less compared with men with college/
university (Table 3). BMI was not associated with the dif-
157

ference between the self-reported and the accelerometer-

0.3

10

18
3d

1d
n = 324
20–34,

4
3

2
2

Assessed by ActiGraph GT1M accelerometer, analyzed in 1-min blocks.

T
T
T
T
T
T
T

T
T
T
T
T
T

measured sedentary time or MVPA.

26
24
20
69
3.8
864

548
210
102

352

440

339
8049

The association between variables for self-reported PA

and objectively measured PA are shown in Table 4. Sex and
Activity time; walking + moderate + vigorous (minIdj1)

age affected the correlation coefficients. Self-reported sitting

time and accelerometer-determined sedentary time had a
Moderate intensity (minIdj1 of 760–5998 cpm)

correlation coefficient between 0.45 and 0.55 for both sexes

Vigorous intensity (minIdj1 of 9 5999 cpm)

Different from all other age groups, P G 0.05.

and all age groups (P G 0.0001) except those 65 yr and older,

Low intensity (minIdj1 of 100–759 cpm)

which had a correlation coefficient of 0.06 for women and

Variables are adjusted for sex and age.
Sedentary (minIdj1 of G100 cpm)

0.24 for men. For moderate-intensity PA, the correlation

Self-reported PA from the IPAQ-S.

coefficient varied between 0.14 (P = 0.052) for women 20–

Total activity (METIminIdj1)
**Different from men, P G 0.01.
Variables are adjusted for sex.
Objective PA measurementsc
Registered time (minIdj1)
Steps (per day), n = 1741

34 yr and 0.40 (P G 0.0001) for men 65 yr and older. For

Total activity (cpmIdj1)
Self-reported variablesf

Moderate (minIdj1)
Vigorous (minIdj1)

vigorous-intensity PA, the correlation coefficient varied be-

Walking (minIdj1)
Sitting (minIdj1)

tween 0.25 (P G 0.0001) for men 50–64 yr and 0.47 (P G

0.0001) for women age 20–34 yr. Education level did not
affect the correlation coefficients significantly.
Variable

Twenty-two percent of the participants were found to

meet the guidelines of a minimum of 30 min of at least
b

d
a

e
f

102 Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

Copyright © 2013 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
 moderate-intensity activity per day (in bouts of 8–10 min)

N = 1020

12 (1)c
–108 (5)c

j57 (2)
as determined by the accelerometer. Of these, 86% were
also captured by the IPAQ. However, the sensitivity of

All,
the IPAQ in capturing those who were categorized as in-
University

10 (1)***
College/
sufficiently active individuals by the accelerometer was

–121 (7)c*

j54 (3)*
n = 529
Women, only 39%.
n = 491

DISCUSSION
Education Level

–87 (8)c

c
15 (1)
j62 (3)
Men,

The main finding of the study was that education level,

sex, and age, but not BMI, affected the differences between
N = 722

15 (1)
j51 (2)
–161 (6)

self-reported (IPAQ-S) and accelerometer-measured

All,

(ActiGraph) PA and sedentary time. This finding is in con-

trast to studies carried out in other countries and with fewer
High School

8 (2)***
or Less

n = 341

j48 (4)
–159 (9)

participants (10,22). In the present study, participants with

Women,

an education level of high school or less reported 21%

higher total activity in minutes per day (walking + MVPA)
n = 381
–172 (8)

j50 (4)

20 (2)

than participants with a college/university degree, whereas

Men,

no differences in accelerometer-determined total activity

(cpmIdj1) was observed between the groups. Furthermore,
9 (1)***

men reported 47% more moderate to vigorous physical ac-

N = 873

j52 (2)
–134 (5)
Women,

tivity (MVPA) than women, although there was no differ-

ence between the sexes in accelerometer-determined MVPA.
Difference (IPAQ-ActiGraph)

All

As suggested by Hagstromer et al. (14), it is possible that

TABLE 3. Mean T SE difference in minutes per day between self-reported and objectively measured PA in relation to age and education level.

N = 878
–127 (6)

j56 (2)

17 (1)

less-educated individuals, and men in general, engage in

Men,

manual activities that might not be recorded efficiently using

an accelerometer or that there could be a differential bias
–203 (14)b

7 (2)*
j33 (7)f
n = 97
Women,

between educational groups or sex in how they answered the

IPAQ. Although participants in the present study reported
65+

walking as the most frequently performed activity, we found

–214 (13)b

j22 (5)b
n = 138

20 (3)

that men carried out more running than women (data not
Men,

EPIDEMIOLOGY
shown). It is therefore possible that the sex difference in the
choice of activity may affect the divergence between the
9 (1)***

two methods. On the other hand, an overestimation of self-

–134 (10)
n = 266
Age Groups (yr)

j53 (4)
Women,

reported PA may also be caused by social desirability re-

IPAQ: self-reported time in walking and moderate activity; accelerometer: 760–5998 cpm.
50–64

sponse bias (1), which could be larger in men and less

Different from participants with an education level of high school or less and, P e 0.01.
A negative value occurs when the IPAQ value is smaller than the accelerometer value.

educated people. Finally, more educated participants and

–119 (10)
n = 303

j52 (4)

17 (2)

women in general may also be more thorough in their esti-

Men,

mation of their PA level.

IPAQ: self-reported time in vigorous activity; accelerometer: Q5999 cpm.
9 (1)***

The divergence of self-reported and objectively measured

IPAQ: self-reported time sitting/lying down; accelerometer: G100 cpm.
j61 (3)*
n = 327
–125 (9)
Women,

PA could also be due to the IPAQ asking about time spent

in moderate and vigorous PA, which is subjective and
Different from the age groups 35–49 and 50–64 yr, P G 0.05.
35–49

dependent on the participant’s physical fitness. At the same

–107 (10)
n = 295

j73 (3)e

15 (1)

time, the accelerometer measures absolute movement inde-

Different from the age group 50–64 yr, P = 0.001.
Men,

pendent of the individual’s fitness. Lower fitness among older

Different from all other age groups, P e 0.001.

people could explain why they report more moderate-intensity

–113 (12)

12 (1)*
j46 (5)*
n = 183
Women,

PA but are measured with less moderate-intensity PA than

participants age 20–34 yr. A higher volume of self-reported
***Different from men, P G 0.0001.
20–34

vigorous-intensity PA compared with accelerometer-

*Different from men, P G 0.05.
–102 (15)
n = 142

Walking+ moderate j60 (4)

20 (3)

measured found in the present study has also been found

Men,

in other studies (14,25). Possible factors explaining this di-

vergence are (a) the inability of accelerometers to measure
activities involving no vertical acceleration, such as cycling
(minIdj1)g
(minIdj1)a

(minIdj1)d
Sedentary

and upper-body movement, and (b) that high-intensity ac-

Vigorous
Variable

tivities are easier to remember due to their association with

the feeling of exhaustion and could easily be overestimated
b

g
a

e
f

SUBJECTIVE AND OBJECTIVE MEASURE OF PA Medicine & Science in Sports & Exercised 103

Copyright © 2013 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.

FIGURE 1—Bland–Altman plot for minutes per day of walking +
moderate-intensity PA assessed by IPAQ and moderate-intensity PA
from accelerometer, defined as 760–5998 cpm (N = 1751). The differ-
ence between the IPAQ and the accelerometer is plotted against the
mean of the two estimates. Mean difference and limits of agreement (SD
= 1.96) are shown in the figure.
if warm-up and cool-down are included. The Bland–Altman
plots (Figs. 1 and 2) showed that the difference between the
two methods increased with higher activity levels, a finding
EPIDEMIOLOGY
in the accelerometer measurement increase the difference
between the two methods because the IPAQ only asks about
time spent sitting/lying and not standing.
Although the IPAQ is not designed for older people,
a study of Japanese men and women aged 65–89 yr found
that the validity for the IPAQ was adequate (29). However,
the present study shows large differences between self-
reported and measured PA in older people. Bauman et al.
(2) also called attention to that IPAQ often is inappro-
priate used with older adults. Reducing the overestimation
of PA using questionnaires seems to require highlighting
important points for participants in general and older peo-
ple in particular.
The choice of accelerometer cut points and analyze
method have large influence on the comparison of the ab-
solute results. Several studies have set the cut point of
moderate-intensity PA at 1952 or 2020 cpm (13,30). How-
ever, Matthews (23) claimed that a cut point of 760 cpm
provided the most accurate group level estimate of time
spent in moderate-intensity activity in daily life, a finding
also supported by other studies (31,33). This cut point cap-
tured most time spent in the activities that were Q3.6 METs
and included common daily activities of a moderate inten-
sity, in addition to ambulatory activities traditionally cap-
tured by the 1952/2020 cpm cut points. Because walking
in the IPAQ-scoring protocol is defined as 3.3 METs,
merging self-reported walking and moderate-intensity PA
in the present study and comparing it to accelerometer-
measured moderate-intensity PA (760–5998 cpm) seem

consistent with other validation studies of different IPAQ

versions (14,15,35). This indicates that either people with
high activity levels overreport their amount of PA or the
accelerometer is less suitable for validating the PA level in
highly active people.
The highest correlation between the IPAQ and accelerom-
eter items was found between total sitting and accelerom-
eter counts per minute G100 (Q = 0.46). Although this
correlation is moderate, the mean self-reported sitting time
was 2.18 hIdj1 lower compared with accelerometer data, with
significant variations between age groups and education level
(Table 3). The oldest age group, 65–84 yr, reported 3.5 hIdj1
less sedentary time compared with accelerometer measure-
ments. Although the accelerometer measurement in the oldest
age group showed around 5 minIdj1 more with sedentary
time compared with the younger participants, the oldest
participants reported approximately 90 minIdj1 less sed-
entary time than the younger age groups. Participants without
a college/university degree reported around 18% less sitting
time than participants with a college/university degree, but the
difference between the groups according to the accelerometer
was only 6% (Table 2). The cut point for sedentary time FIGURE 2—Bland–Altman plot for minutes per day of vigorous-
is G100 cpm and includes sitting, standing, and lying down to intensity PA assessed by IPAQ and accelerometer, defined as 95998 cpm
(N = 1751). The difference between the IPAQ and the accelerometer is
watch television. Because the accelerometer do not differen- plotted against the mean of the two estimates. Mean difference and
tiate between sitting and standing, the standing time included limits of agreement (SD = 1.96) are shown in the figure.

104 Official Journal of the American College of Sports Medicine http://www.acsm-msse.org

Copyright © 2013 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
TABLE 4. Correlation coefficients (Spearman) between questionnaire-derived variables (IPAQ) and objective measures (ActiGraph) of PA, n = 1751.
Sitting
Objective Variables (minIdj1)a
Sedentary, G100 cpm (minIdj1) 0.46***
Moderate, 760–5998 cpm (minIdj1) j0.20***
Vigorous, Q5999 cpm (minIdj1) 0.10***
Moderate and vigorous, Q760 cpm (minIdj1) j0.20***
Total activity (average cpmIdj1) j0.16***
***P G 0.0001.
a
Self-reported time sitting.
b
Self-reported time in walking and moderate activity.
c
Self-reported time in vigorous activity.
d
Self-reported time in activity (walking+ moderate+ vigorous).
fair. We found that the participants reported 49% less
moderate PA than measured by the accelerometer (Table 3),
with no large differences between sex and education level.
A lower self-reported level of moderate-intensity PA could
be explained by the fact that the accelerometer measures all
minutes spent at this intensity, whereas the participants are
asked only to report 10-min blocks of activity. The abso-
lute difference between self-reported and objectively mea-
sured moderate PA was reduced from 54 to 12 minIdj1
when only 10-min blocks of these accelerometer data were
included.
Overall, the main correlation coefficients in the present
study ranged from 0.23 to 0.46 and corresponded to the
range found in the review study of the IPAQ-S (21). In three
systematic reviews of the content and measurement proper-
ties of PA questionnaires, effect sizes higher than 0.5 were
considered acceptable for correlations between objective
activity-measuring devices and questionnaires (8,12,32).
However, when the correlation criterion of 0.5 was used,
the conclusions in the reviewed studies were overly opti-
mistic in almost all cases. A reason for this could be that the
researchers validating the questionnaires are the same re-
searchers who want to use the questionnaire later on. None
of the reviewed IPAQ-S studies reached the minimal ac-
ceptable correlation standard recommended in the litera-
ture of 0.5 for objective activity measures (21). On the
other side, Lee et al. (22) concluded that a correlation of
0.3–0.4 is perhaps as close as can be expected for criterion
validity of a physical-activity questionnaire with 10 ques-
tions against a mechanical device that detects body move-
ment. In the present study, 67% of the participants were
categorized as sufficiently active by the IPAQ, whereas the
corresponding number for the accelerometer was 22% (data
not shown). This indicates that the IPAQ ability to capture
inactive people is limited, which could result in an over-
reporting of physically active people if PA were measured
by the IPAQ only. However, this finding should be inter-
preted with care because the definitions of sufficiently
physically active is different between IPAQ and acceler-
ometer (see Methods section).
Strengths and limitations of the study. The major
strength of the study is the large population sample size
recruited from a wide age range throughout Norway and
SUBJECTIVE AND OBJECTIVE MEASURE OF PA
Copyright © 2013 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
Questionnaire-Derived Variables
Walking + Moderate Vigorous Total activity time Total Activity
(minIdj1)b (minIdj1)c (minIdj1)d (METIminIdj1)
j0.18*** j0.08 j0.19*** j0.18***
0.23*** 0.18*** 0.27*** 0.27***
j0.02 0.36*** 0.11*** 0.19***
0.22*** 0.21*** 0.27*** 0.28***
0.25*** 0.25*** 0.32*** 0.33***
the random inclusion of rural and urban populations. Be-
cause of the large number of participants, we could obtain a
sizable number of participants from every subgroup based
on sex, age, education level, and BMI. Limitations of the
study include the large number of participants not answer-
ing all the IPAQ questions. There was no particular question
causing a large dropout number, but many participants had
not reported either how many days per week they carried
out the specific PA or how much time they usually spent
doing the PA. One explanation could be that some partici-
pants found the questionnaire too long because we included
other questions than the IPAQ. The mean age was higher
in the group with incomplete IPAQ answers, so it seems
that older people were more inclined to skip or misunder-
stand some of the items. Those with an incomplete set of
IPAQ data had no significant difference in total PA level
(9100 cpm), 4.0% less MVPA (9760 cpm; P = 0.011), and
2.7% less sedentary time compared with those with a com-
plete set of IPAQ data (P G 0.0001). These differences in
EPIDEMIOLOGY
measured PA-level were relatively small. Thus, the exclu-
sion of the group with an incomplete set of IPAQ data does
not seem to have a large effect on the results in the present
study. Because both the accelerometer and the IPAQ were
sent to the participants by mail, another limitation of the
study is that the questionnaire data and the accelerometer
data do not necessarily represent the same 7 d. However, a
comparison between the two methods provides valuable in-
formation, as they both provide snapshots of habitual phys-
ical activity level.
CONCLUSIONS
The present study shows large variations between self-
reported and accelerometer-measured PA and sedentary time.
Sex, age, and educational level, but not BMI, influenced these
variations, and the diversity between the two methods in both
sedentary time and vigorous-intensity PA was greatest among
men with a lower education level and at the higher end of the
age spectrum. The difference between the self-reported and
objectively measured PA increased with higher activity and
intensity levels. The general agreement between self-reported
and accelerometer-measured PA was poor, and their correla-
tion coefficients were lower than what is recommended. The
Medicine & Science in Sports & Exercised 105
 choice of accelerometer cut points and data reduction method
have large influence on the comparison of the absolute values.
The sum of self-reported walking and moderate PA seems
comparable with accelerometer-measured activity using a cut
point of 760–5998 cpm and analyzed in blocks of 10 min.
The authors thank all the test personnel at the 10 institutions in-
volved in the study for their work during the data collection: Finnmark
REFERENCES
1. Aadahl M, Jorgensen T. Validation of a new self-report instrument
for measuring physical activity. Med Sci Sports Exerc. 2003;35(7):
1196–202.
2. Bauman A, Ainsworth BE, Bull F, et al. Progress and pitfalls in
the use of the International Physical Activity Questionnaire (IPAQ)
for adult physical activity surveillance. J Phys Act Health. 2009;
6:S5–8.
3. Bauman A, Ainsworth BE, Sallis JF, et al. The descriptive epide-
miology of sitting. A 20-country comparison using the Interna-
tional Physical Activity Questionnaire (IPAQ). Am J Prev Med.
2011;41(2):228–35.
4. Bauman A, Bull F, Chey T, et al. The International Prevalence
Study on Physical Activity: results from 20 countries. Int J Behav
Nutr Phys Act. 2009;6(1):21.
5. Becker W, Lyhne N, Pedersen A, et al. Nordic Nutrition Recom-
mendations 2004—Integrating Nutrition and Physical Activity. 4th ed.
Copenhagen (Denmark): Nordic Concil of Ministers; 2004. p. 436.
6. Bland JM, Altman DG. Applying the right statistics: analyses of
measurement studies. Ultrasound Obstet Gynecol. 2003;22(1):85–93.
7. Brown W, Bauman A, Chey T, Trost S, Mummery K. Comparison
of surveys used to measure physical activity. Aust N Z J Public
Health. 2004;28(2):128–34.
8. Chinapaw MJM, Mokkink LB, van Poppel MNM, van Mechelen
W, Terwee CB. Physical activity questionnaires for youth. A sys-
tematic review of measurement properties. Sports Med. 2010;
40(7):539–63.
9. Craig CL, Marshall AL, Sjostrom M, et al. International Physical
EPIDEMIOLOGY
Activity Questionnaire: 12-country reliability and validity. Med Sci
Sports Exerc. 2003;35(8):1381–95.
10. Ekelund U, Sepp H, Brage S, et al. Criterion-related validity of
the last 7-day, short form of the International Physical Activity
Questionnaire in Swedish adults. Public Health Nutr. 2006;9(2):
258–65.
11. Fogelholm M, Malmberg J, Suni J, et al. International Physical
Activity Questionnaire: validity against fitness. Med Sci Sports
Exerc. 2006;38(4):753–60.
12. Forsen L, Loland NW, Vuillemin A, et al. Self-administered
physical activity questionnaires for the elderly. A systematic re-
view of measurement properties. Sports Med. 2010;40(7):601–23.
13. Freedson PS, Melanson E, Sirard J. Calibration of the Computer
Science and Applications, Inc. accelerometer. Med Sci Sports
Exerc. 1998;30(5):777–81.
14. Hagstromer M, Ainsworth BE, Oja P, Sjostrom M. Comparison of
a subjective and an objective measure of physical activity in a
population sample. J Phys Act Health. 2010;7(4):541–50.
15. Hagstromer M, Oja P, Sjostrom M. The International Physical
Activity Questionnaire (IPAQ): a study of concurrent and con-
struct validity. Public Health Nutr. 2006;9(6):755–62.
16. Hagstromer M, Troiano RP, Sjostrom M, Berrigan D. Levels and
patterns of objectively assessed physical activity—a comparison
between Sweden and the United States. Am J Epidemiol. 2010;
171(10):1055–64.
17. Hansen BH, Kolle E, Dyrstad SM, Holme I, Anderssen SA.
Accelerometer-determined physical activity in adults and older
people. Med Sci Sports Exerc. 2012;44(2):266–72.
18. Healy GN, Dunstan DW, Salmon J, et al. Objectively measured
light-intensity physical activity is independently associated with
2-h plasma glucose. Diabetes Care. 2007;30(6):1384–9.
106 Official Journal of the American College of Sports Medicine
Copyright © 2013 by the American College of Sports Medicine. Unauthorized reproduction of this article is prohibited.
University College, Hedmark University College, NTNU Social Re-
search AS, Sogn og Fjordane University College, University of Agder,
University of Nordland, University of Stavanger, Telemark University
College, Vestfold University College, and Norwegian School of Sport
Sciences. The study was funded by the Norwegian Directorate of
Health and the Norwegian School of Sport Sciences.
The authors declare no conflict of interests.
The results of the present study do not constitute endorsement by
the American College of Sports Medicine.
19. IPAQ. Guidelines for data processing and analysis of the Inter-
national Physical Activity Questionnaire (IPAQ). Short and long
forms. Available from http://www.ipaq.ki.se/scoring.pdf. 15 pp.
20. Kurtze N, Rangul V, Hustvedt BE. Reliability and validity of the
International Physical Activity Questionnaire in the Nord–
Trondelag Health Study (HUNT) population of men. BMC Med
Res Methodol. 2008;8:63.
21. Lee PH, Macfarlane DJ, Lam TH, Stewart SM. Validity of
the International Physical Activity Questionnaire Short Form
(IPAQ-SF): a systematic review. Int J Behav Nutr Phys Act. 2011;
8:115.
22. Lee PH, Yu YY, McDowell I, Leung GM, Lam TH, Stewart SM.
Performance of the International Physical Activity Questionnaire
(Short Form) in subgroups of the Hong Kong Chinese population.
Int J Behav Nutr Phys Act. 2011b;8:81.
23. Matthews CE. Calibration of accelerometer output for adults. Med
Sci Sports Exerc. 2005;37(11):S512–22.
24. Matthews CE, Chen KY, Freedson PS, et al. Amount of time spent
in sedentary behaviors in the United States, 2003–2004. Am J
Epidemiol. 2008;167(7):875–81.
25. Prince SA, Adamo KB, Hamel ME, Hardt J, Gorber SC, Tremblay M.
A comparison of direct versus self-report measures for assessing
physical activity in adults: a systematic review. Int J Behav Nutr
Phys Act. 2008;5:56.
26. Rutten A, Vuillemin A, Ooijendijk WT, et al. Physical activity
monitoring in Europe. The European Physical Activity Surveil-
lance System (EUPASS) approach and indicator testing. Public
Health Nutr. 2003;6(4):377–84.
27. Rzewnicki R, Vanden AY, De BI. Addressing overreporting on the
International Physical Activity Questionnaire (IPAQ) telephone
survey with a population sample. Public Health Nutr. 2003;6(3):
299–305.
28. Sjöström M, Oja P, Hagströmer M, Smith BJ, Bauman A.
Health-enhancing physical activity across European Union
countries: the Eurobarometer study. J Public Health. 2006;
14(5):291–300.
29. Tomioka K, Iwamoto J, Saeki K, Okamoto N. Reliability and va-
lidity of the International Physical Activity Questionnaire (IPAQ)
in elderly adults: the Fujiwara-kyo Study. J Epidemiol. 2011;21(6):
459–65.
30. Troiano RP, Berrigan D, Dodd KW, Masse LC, Tilert T,
McDowell M. Physical activity in the United States measured
by accelerometer. Med Sci Sports Exerc. 2008;40(1):181–8.
31. van der Ploeg HP, Merom D, Chau JY, Bittman M, Trost SG,
Bauman AE. Advances in population surveillance for physical
activity and sedentary behavior: reliability and validity of time use
surveys. Am J Epidemiol. 2010;172(10):1199–206.
32. van Poppel MNM, Chinapaw MJM, Mokkink LB, van Mechelen
W, Terwee CB. Physical activity questionnaires for adults: a sys-
tematic review of measurement properties. Sports Med. 2010;
40(7):565–600.
33. Welk GJ, McClain JJ, Eisenmann JC, Wickel EE. Field validation
of the MTI ActiGraph and BodyMedia armband monitor using
the IDEEA monitor. Obesity. 2007;15(4):918–28.
34. World Health Organization. Obesity: Preventing and Managing
the Global Epidemic. Report of a WHO consultation. Geneva:
World Health Organization; 2000.
35. Wolin KY, Heil DP, Askew S, Matthews CE, Bennett GG. Vali-
dation of the International Physical Activity Questionnaire–Short
among blacks. J Phys Act Health. 2008;5(5):746–60.
http://www.acsm-msse.org