Journal of Sports Sciences

ISSN: 0264-0414 (Print) 1466-447X (Online) Journal homepage: http://www.tandfonline.com/loi/rjsp20

Validity of the “Samsung Health” application

to measure steps: A study with two different
samsung smartphones

Vicente J. Beltrán-Carrillo, Alejandro Jiménez-Loaisa, Miriam Alarcón-López &

Jose L. L. Elvira

To cite this article: Vicente J. Beltrán-Carrillo, Alejandro Jiménez-Loaisa, Miriam Alarcón-

López & Jose L. L. Elvira (2018): Validity of the “Samsung Health” application to measure
steps: A study with two different samsung smartphones, Journal of Sports Sciences, DOI:
10.1080/02640414.2018.1527199

To link to this article: https://doi.org/10.1080/02640414.2018.1527199

Published online: 17 Oct 2018.

Submit your article to this journal

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=rjsp20
JOURNAL OF SPORTS SCIENCES
https://doi.org/10.1080/02640414.2018.1527199

PHYSICAL ACTIVITY, HEALTH AND EXERCISE

Validity of the “Samsung Health” application to measure steps: A study with two
different samsung smartphones
Vicente J. Beltrán-Carrillo , Alejandro Jiménez-Loaisa , Miriam Alarcón-López and Jose L. L. Elvira
Sport Research Centre, Miguel Hernández University of Elche, Alicante, Spain

ABSTRACT KEYWORDS
The purpose of this study was to examine the validity of a highly popular pedometer application Pedometer; physical activity;
(Samsung Health). Sixteen adults (28.8 ± 8.9 years of age) wore two Samsung smartphone models, walking; running; health
Samsung Galaxy Core Prime (SGCP) and Samsung Galaxy S4 (SGS4), at three body locations (waist, arm,
and hand) while walking and running over a 50-m test. All trials were recorded using a video as a gold
standard measure of step counts. Results indicated that the validity of Samsung Health varied depend-
ing on the smartphone model, its body location, and the type of gait (walking and running). Samsung
Health showed acceptable validity when the SGCP was located on the hand (Bias = −8.3%; RMSE = 5.6),
and especially on the arm (Bias = −7.2%; RMSE = 4.9) while running, and when the SGS4 was located on
the arm (Bias = −7.5%; RMSE = 5.4), and especially on the waist (Bias = 5.4%; RMSE = 3.7) while walking.
Samsung Health only showed good validity when the SGS4 was located on the arm (Bias = 2.9%;
RMSE = 3.6), and especially on the hand (Bias = 0.5%; RMSE = 2.5) while running. This application
showed unacceptable validity in the remaining options.

Introduction social obstacles entailed by this type of physical activity in

comparison to other forms of exercise (Alexander, Cowburn, &
Physical inactivity increases the risk of premature mortality
Foster, 2006; Pillay et al., 2015).
and many chronic diseases, including cardiovascular disease,
Given the importance of walking for public health, and
thromboembolic stroke, hypertension, Type 2 diabetes, osteo-
considering the high rates of physical inactivity in today’s
porosis, obesity, colon cancer, breast cancer, anxiety and
society (WHO, 2018), the tools that objectively measure this
depression (Lee & Buchner, 2008; World Health Organization
type of physical activity have gained special relevance in
[WHO], 2018).
recent years. Pedometers have been one of the most popular
Whereas the health benefits of vigorous activities like run-
devices for self-monitoring the number of daily steps, as they
ning have traditionally been recognized, moderate-intensity
are relatively inexpensive, small, and easy to use (Gardner &
physical activity, including walking, began to receive attention
Campagna, 2011).
as a healthy form of physical activity in the 1990s (Lee &
With the advances in smartphone technology and the
Buchner, 2008). Nowadays, there is strong evidence about
inclusion of motion sensors such as accelerometers and gyro-
the health benefits of walking. For instance, Caspersen and
scopes in these devices, smartphones can be used as ped-
Fulton (2008) indicated that walking can reduce the risk of
ometers via pedometer applications. Nowadays, more than
Type 2 diabetes by between approximately 20% and 40%.
250 pedometer applications can be downloaded from differ-
Dwyer et al. (2015) revealed that people who change from a
ent mobile platforms to measure steps and physical activity
sedentary state to take about 10,000 steps per day have more
(Leong & Wong, 2016).
than a 40% lower risk of all-cause mortality. Other studies have
If these applications were valid tools for the measurement
shown that walking is a healthy behaviour which can reduce
of steps, the population could use them to set their own goals,
cardiovascular risk markers such as systolic and diastolic blood
self-monitor their own step counts and, in this way, increase
pressure, body fat, cholesterol and waist circumference
their motivation for physical activity and compliance with
(Bravata et al., 2007; Murtagh, Murphy, & Boone-Heinonen,
physical activity recommendations (Boyce, Padmasekara, &
2010; Yuenyongchaiwat, 2016).
Blum, 2012; Gu, Chen, Jackson, & Zhang, 2018). Moreover,
Moreover, walking is the most commonly reported physical
researchers could use these tools to evaluate interventions
activity for the general population both in developed and
for the promotion of physical activity, replacing higher cost
developing countries (Eyler, Brownson, Bacak, & Housemann,
devices (Orr et al., 2015).
2003; Owen, Humpel, Leslie, Bauman, & Sallis, 2004; Pillay
Despite the popularity of pedometer applications, there is a
et al., 2015). This is, to a large extent, due to the low cost
lack of research about their validity as tools for physical activity
and easy accessibility of this activity for the majority of the
assessment. To date, the studies which have explored the validity
population, along with the lower physical, psychological and

CONTACT Vicente J. Beltrán-Carrillo vbeltran@umh.es Centro de Investigación del Deporte, Universidad Miguel Hernández de Elche, Avenida de la
Universidad s/n, 03202 Elche, Alicante, Spain
© 2018 Informa UK Limited, trading as Taylor & Francis Group
2 V. J. BELTRÁN-CARRILLO ET AL.
of Android apps (Åkerberg, Lindén, & Folke, 2012; Leong & Wong,
2016; Orr et al., 2015) or iPhone apps (Åkerberg et al., 2012;
Bergman, Spellman, Hall, & Bergman, 2012; Boyce et al., 2012;
Orr et al., 2015) have shown inconsistent results. Only two of
these studies have found an acceptable validity in two ped-
ometer applications, the Android-based application Tayutau
(Leong & Wong, 2016), and the iPhone-based application
Pedometer Ultimate GPS+ (Åkerberg et al., 2012).
In addition, previous research has warned that the validity
of pedometer applications depends on the body location of
the smartphone, the intensity with which a person walks, and
the quality of the movement sensors of a particular model of
smartphone (Åkerberg et al., 2012; Boyce et al., 2012; Leong &
Wong, 2016). In this regard, pedometer applications are often
more valid when the smartphone is placed in the front pants
pocket than when placed on the arm (Åkerberg et al., 2012;
Bergman et al., 2012; Leong & Wong, 2016). The validity of
pedometer applications also improves when the walking
speed increases (Dondzila, Swartz, Miller, Lenz, & Strath,
2012; Leong & Wong, 2016). Moreover, a pedometer applica-
tion may differ in its measurement depending on the model of
smartphone used (Åkerberg et al., 2012).
In view of the previous evidence and the increasing
popularity of pedometer applications, Bergman et al.
(2012) declared that further research with non-analysed
applications is needed to assess their validity at different
locations and speeds. Therefore, the aim of this study was
to examine the validity of a highly popular pedometer
application based on an Android platform (Samsung
Health; Samsung Electronics Co., Ltd), comparing two differ-
ent Samsung smartphones (Samsung Galaxy Core Prime,
SGCP; and Samsung Galaxy S4, SGS4), three body locations
(waist, arm and hand) and two different gaits (walking and
running). This application has been downloaded more than
100 million times in Google Play by 12 February 2018 and it
is already being pre-installed on all modern Samsung smart-
phones. However, to our knowledge, only one previous
study has analysed the validity of the Samsung Health appli-
cation for measuring steps (Johnson, Turek, Dornfeld, Drews,
& Hansen, 2016). Moreover, our study includes important
methodological innovations regarding the study of Johnson
et al. (2016). First, our study included video analysis as a
criterion or gold standard measure of step counts, whereas
Johnson et al. (2016) used the pedometer StepWatch 3,
which indirectly estimates step counts. Second, our study
compared the validity of two different Samsung
Smartphones (SGCP and SGS4), whereas Johnson et al.
(2016) only analysed the validity of SGS4. Third, the tests
used in our study to examine the validity of the Samsung
Health application (to walk and run a 50-m track) were
different from the test used by Johnson et al. (2016) (to
walk a 200-m track and run on a treadmill). Fourth, our
study used different body locations of the smartphones
(waist, arm, hand) from those used by Johnson et al.
(2016) (pants pocket and hand). Fifth, our study was carried
out with the 5.7.1.0003 version of the Samsung Health
application, whereas Johnson et al. (2016) used the
2.0.0.009 version, and different versions may include differ-
ent algorithms affecting step counting.
Material and methods
Participants and ethics
Sixteen adults participated in this study (eight men and eight
women). A purposive sampling was used to recruit partici-
pants from the university setting. All participants were healthy
and injury-free in the moment of the study. Subjects with
known medical problems or episodes of musculoskeletal
pain were excluded. Each participant signed and provided a
written informed consent before participating. This study was
approved by the Research Ethics Committee of the first
author’s university. Participants’ characteristics are presented
in Table 1.
Instruments and procedure
This study used two different models of Samsung smart-
phones based on the Android operating system, SGCP and
SGS4. Both models had already pre-installed the application
Samsung Health 5.7.1.0003. According to its website, this appli-
cation was designed to track users’ activity, nutrition and
sleep, and to promote an active and healthy lifestyle among
users (see http://www.samsung.com/uk/apps/samsung-
health/). This application includes a pedometer function.
Measurements of each participant were performed on a
flat, straight running track. Participants were instructed to
perform the tests over a distance of 50 m. This distance was
previously marked with two lines which perpendicularly
crossed the running track. Participants stood still before the
starting line and after passing the finish line. We considered
the effect of the acceleration and deceleration in the 50-m
section to be negligible because the maximum speeds were
relatively low. Participants had to walk at the pace they would
normally walk for health in their daily life. With regard to
running, participants were instructed to run at a moderate
intensity (around 5–6 of the Borg CR10 Physical Exertion
Scale; Borg, 1998). In order to guarantee a correct understand-
ing of moderate intensity, all participants were previously
instructed to report appropriately perceived exertion with
Borg Scale. After the explanations and examples, participants
could raise doubts which were solved by the researcher in
charge of the measurements.
Participants had to perform a total of six trials, three for
walking and, subsequently, three for running. In the first trial,
the smartphones were placed laterally on the right side of the
waist, in vertical orientation, using a waistband. Concretely,
the smartphones were placed on the most lateral position of
the waist, with their upper part aligned with the participants’
iliac crest. In the second trial, the smartphones were placed
laterally on the right arm, in vertical orientation, using an
armband (mid-point between the shoulder and the elbow).
Table 1. Participants’ characteristics (Mean ± SD).
Men (n = 8) Women (n = 8) All participants (n = 16)
Age (years) 28.5 ± 10.8 29.1 ± 7.2 28.8 ± 8.9
Height (cm) 180.62 ± 0.04 170.87 ± 0.05 175.70 ± 0.07
Weight (kg) 80.75 ± 2.91 62.75 ± 7.19 71.75 ± 10.70
BMI (kg/m2) 24.76 ± 1.97 21.46 ± 1.97 23.11 ± 2.24
BMI = Body Mass Index.

In the third trial, participants had to wear the smartphones on
their right hand (the smartphones were placed along the
longitudinal axis). Both models of Samsung phones were
placed one together with the other, always in the same posi-
tion, and with their screens facing away from the participant’s
body. Thus, both smartphones would provide step counts
concerning the same trials.
Simultaneously, the trials were video recorded with an
iPhone SE (30 Hz), held at a fixed central point about 15 m
from the track to appropriately follow the participants with
the smartphone camera along the 50-m test. After each trial,
the researcher registered the steps counted by the Samsung
Health application.
The videos were later analysed to count the steps, and used
as the gold standard when compared with the smartphones
measurements. The video analysis was carried out by the same
researcher who did the measurements. The videos were repro-
duced with the Brain Craft Ltd application (version 1.5) in slow-
motion option to favour the count of steps. Each video was
analysed two times to check that the counting was correct. If
the two step counts did not coincide, the researcher had to
analyse the video a third time. One step was defined oper-
ationally as the interval between two consecutive counter-
footsteps. To minimize the error, the count began when either
foot touched the ground after the starting line, and finished
when either foot touched the ground after the finish line.
Statistical analysis
The Shapiro-Wilk test was calculated, and all the data showed
a normal distribution. Paired sample t-tests were used to
compare mean differences between the step counts of the
two models of smartphones (SGCP, SGS4) with the steps
recorded in the video. Statistical differences were established
Figure 1. Differences (Mean ± SD) in step counts between Samsung Health with each smartphone and the video (gold standard).
JOURNAL OF SPORTS SCIENCES 3
at a p-value of less than 0.05. In addition, the corrected effect
size (ES) or Hedges’ g for related samples with pooled stan-
dard deviation was calculated between the video and each
smartphone measurement. ES 95% confidence intervals were
calculated as described in Nakagawa and Cuthill (2007), and
used to determine statistical differences when the confidence
interval did not contain the zero value. Threshold values for
Cohen’s ES statistic were < 0.2 (small), around 0.5 (moderate)
and > 0.8 (large) (Cohen, 1988). Additional comparisons
between each smartphone and the video included: relative
bias and root mean square error (RMSE) to account for preci-
sion and accuracy, respectively; Pearson’s correlation (with its
95% confidence interval) to test the association between mea-
surements; and Bland-Altman plots (Bland & Altman, 1986) to
establish the agreement between measurements. All the cal-
culations were made using custom Python 2.7 scripts.
Results
The mean differences in step counts between Samsung Health
with each smartphone (SGCP, SGS4) and the video (gold stan-
dard) are presented in Figure 1. These differences are specified
for the three body locations of the smartphones (waist, arm,
and hand) and the two types of gait (walking and running).
Statistical comparisons are presented in Table 2. Paired
sample t-tests showed significant statistical differences
(p < 0.001), with large effect sizes, in most of the comparisons
between the two Samsung smartphones and the video,
regardless of body location and type of gait. However, no
significant differences with respect to the video were found
when the smartphone SGS4 was placed on the arm (p = 0.066;
ES = 0.57 [−0.11, 1.25], moderate) or on the hand (p = 0.700;
ES = 0.06 [−0.30, 0.43], small) while running.
4 V. J. BELTRÁN-CARRILLO ET AL.

Table 2. Statistical differences between Samsung Health with each smartphone placed on three different body positions and the video (gold standard).
SGCP vs. Video SGS4 vs. Video
p ES [95% CI] Bias (%) RMSE r [95% CI] p ES [95% CI] Bias (%) RMSE r [95% CI]
Walking
Waist < .001 −6.60 [−9.18, −4.02] −28.1 18.5 0.91 [0.74, 0.97] < .001 1.12 [0.64, 1.60] 5.4 3.7 0.93 [0.80, 0.98]
Arm < .001 −1.88 [−2.71, −1.05] −10.3 7.3 0.73 [0.33, 0.91] < .001 −1.53 [−2.24, −0.83] −7.5 5.4 0.75 [0.37, 0.92]
Hand < .001 −3.34 [−4.71, −1.97] −16.0 11.1 0.67 [0.21, 0.88] < .001 −2.43 [−3.42, −1.44] −12.9 8.9 0.86 [0.59, 0.95]
Running
Waist < .001 2.25 [1.24, 3.25] 14.1 9.0 0.56 [0.04, 0.84] < .001 4.26 [2.57, 5.94] 28.3 17.1 0.84 [0.56, 0.95]
Arm < .001 −1.70 [−2.49, −0.91] −7.2 4.9 0.67 [0.21, 0.88] 0.066 0.57 [−0.11, 1.25] 2.9 3.6 0.27 [−0.31, 0.70]
Hand < .001 −1.69 [−2.44, −0.94] −8.3 5.6 0.77 [0.40, 0.92] 0.700 0.06 [−0.30, 0.43] 0.5 2.5 0.76 [0.39, 0.92]
SGCP, Samsung Galaxy Core Prime; SGS4, Samsung Galaxy S4; p, t-test p-value; ES, Effect Size; CI, Confidence Interval; RMSE, Root-Mean-Square Error in step counts;
r, correlation coefficient.

Figure 2 presents the Bland-Altman plots of the compar-
isons. They showed no trend in the differences, which means
that the agreement between methods did not vary over the
range of measurements. The limits of agreement were about
10 steps in all conditions except for walking with the smart-
phones on the waist, where it was about 5 steps.
For the SGCP, the Samsung Health application was more
precise and accurate when the smartphone was placed on the
arm, both while walking (Bias = −10.3%, RMSE = 7.3 steps) and
running (Bias = −7.2%, RMSE = 4.9 steps), but less so when the
device was placed on the waist, both while walking
(Bias = −28.1%, RMSE = 18.5 steps) and running (Bias = 14.1%,
RMSE = 9.0 steps). All measures underestimated step counts
except when participants ran with the SGCP placed on the waist.
The r-values concerning SGCP showed that the highest
association between the SGCP and the video step counts
was found when the smartphone was placed on the waist
while walking (r = 0.91 [0.74, 0.97]). Conversely, the lowest
association between the SGCP and the video step counts was
found when the smartphone was placed on the waist while
running (r = 0.56 [0.04, 0.84]).
Regarding the SGS4, the Samsung Health application was
more precise and accurate when participants ran with the
smartphone placed on the arm (Bias = 2.9%, RMSE = 3.6
steps) and especially when they ran with the smartphone
placed on the hand (Bias = 0.5%, RMSE = 2.5 steps). Again,
wearing the smartphone on the waist was the least precise
and accurate location when running (Bias = 28.3%,

Figure 2. Bland-Altman plots of step counts comparing Samsung Health with each smartphone and the video. The thick line represents the mean difference and the
dashed line its 95% limits of agreement. SGCP, Samsung Galaxy Core Prime; SGS4, Samsung Galaxy S4; VI, video. Note. Some dots overlap.

RMSE = 17.1 steps). However, this phenomenon was reversed
when participants had to walk, as the measure with the
smartphone on the waist was the most precise and accurate
(Bias = 5.4%, RMSE = 3.7 steps), whereas the measure with the
smartphone on the hand was the least precise and accurate
(Bias = −12.9%, RMSE = 8.9 steps).
The r-values concerning SGS4 showed that the highest
association between the SGS4 and the video step counts
was found when the smartphone was placed on the waist
while walking (r = 0.93 [0.80, 0.98]). The lowest association
was found at running gait, when participants wore the smart-
phone on the arm (r = 0.27 [−0.31, 0.70]).
Discussion
In view of the increasing popularity of pedometer-based appli-
cations to measure and monitor physical activity, and the lack
of knowledge about their validity, the aim of this study was to
analyse the validity of the highly popular application Samsung
Health to measure steps. The results of this study generally
indicated that the validity of Samsung Health varied depend-
ing on the smartphone model, the body location of the
smartphone and the type of gait (walking or running).
With regard to the smartphone model, the different validity
results between SGCP and SGS4 are in line with previous
research. Åkerberg et al. (2012), who evaluated the validity
of six different pedometer applications for three different
smartphones, also found that the same pedometer applica-
tions registered broad differences in the number of steps
depending on the smartphones in which they were installed.
Leong and Wong (2016), when analysing the validity of three
pedometer applications in free-living settings, also confirmed
that each concrete application provided clearly different step
counts with different smartphones. Hence, the different valid-
ity results observed between SGCP and SGS4 could be due to
the quality of the movement sensors incorporated in each
different smartphone. In general, Samsung Health showed
better validity with SGS4, probably because this model of
smartphone possesses newer and better technology.
The validity of Samsung Health also varied depending on
the body location of the smartphone (waist, arm and hand)
and the type of gait (walking and running) (see Table 2). In
contrast, the only previous study with the Samsung Health
application (Johnson et al., 2016) found no significant differ-
ences between Samsung Health estimated steps and the mea-
sures of the pedometer StepWatch 3 while walking or running,
regardless of the body location of the smartphone (pants
pocket and hand). These results are understandable consider-
ing the methodological differences between the study of
Johnson et al. (2016) and our study (see last paragraph of
the introduction). Concretely, the differences could be due to
the fact that Johnson et al. (2016) used an indirect and less
accurate criterion measure of step counts. Another possible
explanation is that the front pants pocket could be as valid a
body location as the hand, although our study did not analyse
this specific issue because the pants pocket location was not
included. Nevertheless, the findings of our study are congru-
ent with other previous research. Åkerberg et al. (2012) found
more valid step counts with the smartphone located in the
JOURNAL OF SPORTS SCIENCES 5
front pants pocket than in the chest pocket of the jacket or on
the arm. Leong and Wong (2016), when testing the validity of
three pedometer applications in laboratory, confirmed that
the pedometer application improved in step counting validity
when the walking speed on a treadmill increased and the
smartphone was placed in the front pants pocket (compared
to the waist and the arm). This increase in validity as walking
speed increases was also found in studies with electronic
pedometers (Crouter, Schneider, Karabulut, & Bassett, 2003;
Dondzila et al., 2012). It seems that faster movements gener-
ate higher accelerations that can trigger step detection by
different smartphone applications (Bergman et al., 2012;
Leong & Wong, 2016).
In our study, Samsung Health showed more valid results
with both smartphones located on the arm or the hand while
running, when compared to the waist location (Table 2). This
finding could be due to the larger arm swinging action at
faster speeds, generating higher accelerations to be identified
as steps (Åkerberg et al., 2012; Leong & Wong, 2016; Meyns,
Bruijn, & Duysens, 2013).
Our results concerning waist location while walking were
inconsistent. The waist location was the least valid with SGCP,
while it was the most valid with SGS4 (see Figure 1). In line
with this last result, Leong and Wong (2016) found more valid
step counts when a SGS4 (same model of smartphone as in
our study) was located on the waist than on the arm at low
walking speed. Hence, our results with SGS4 may be due to a
light arm swinging action at walking compared to running,
influencing the underestimation of step counts at arm/hand
location. Future studies with Samsung Health and other mod-
els of smartphones are necessary to clarify which is the most
recommended body location to measure steps while walking.
In view of Leong and Wong’s (2016) results, these studies
should also include the location in the front pants pocket to
confirm that this location may be the most recommendable
for users when walking. Moreover, future studies should follow
more similar methodologies to favour comparison of results.
Previous research has established that instruments provid-
ing step counts below 10% of absolute percentage error show
acceptable validity, regardless of a tendency to over- or under-
count, although they would not be valid enough for research
purposes (Leong & Wong, 2016; Tudor-Locke et al., 2006).
Therefore, Samsung Health users should consider that this
application showed good validity only when the SGS4 was
located on the arm, and especially on the hand, while running,
and acceptable validity when located on the arm, and espe-
cially on the waist, while walking. On another hand, Samsung
Health showed acceptable validity only when the SGCP was
located on the hand, and especially on the arm, while running
but not while walking. However, when the SGCP was placed
on the waist while walking, the correlation was high (r = 0.91
[0.74, 0.97]) and the limits of agreement of the Bland-Altman
plot were quite narrow (about 5 steps). This leads us to
assume that adding the bias or mean raw difference (18
steps) to the steps measured with SGCP could provide a
good approximation for this smartphone while walking. This
application showed unacceptable validity in the remaining
options (see Table 2 and Figure 2). Therefore, these results
lead to the following use recommendations. Populations who
6 V. J. BELTRÁN-CARRILLO ET AL.
cannot run or walk fast, like some older adults, should use the
SGS4 and wear it on the waist for a valid step count. Runners
should use the SGS4 and wear it on the hand. If this position is
uncomfortable for long distances, they could wear it on
the arm.
The results of this study also indicated that correlations (r)
can be misleading statistical values when assessing the validity
of pedometer applications. For instance, step counts with the
SGCP on the waist while walking showed the second highest
correlation (r = 0.91) with the video step counts, although this
measure showed the second highest Bias % (−28.14). In this
sense, step counts with the SGS4 on the arm while running
showed the lowest correlation (r = 0.27), although this mea-
sure showed the second smallest Bias % (2.86). This highlights
the importance of analysing the validity from different view-
points using different statistics that provide complementary
information.
Lastly, more research is needed about the validity of the most
common pedometer applications for smartphones. This informa-
tion may be useful to consumers and researchers attempting to
identify the most cost-effective options for physical activity mea-
surement (Orr et al., 2015). Given the importance of pedometer
self-monitoring for the promotion of physical activity (Aittasalo,
Miilunpalo, Kukkonen-Harjula, & Pasanen, 2006), pedometer
applications should not be diffused for use by the general public
until an acceptable validity is demonstrated. Perhaps, the inher-
ent competitiveness of a free market economy forces a quick
development and inclusion of new apps and smartphones in the
market without the necessary concern about quality standards.
In this regard, it would be recommendable for pedometer appli-
cation developers, smartphone manufacturers and physical
activity and sport scientists to collaborate to improve the validity
of pedometer applications. This collaboration could be profit-
able, taking into account that validity of smartphone applications
will depend on different factors, such as the quality of the move-
ment sensors of the smartphone, the quality of the application
and its algorithm, the body location of the smartphone or the
user’s pace while walking/running (Åkerberg et al., 2012;
Bergman et al., 2012; Leong & Wong, 2016). Manufacturers
should provide their customers with recommendations of use
based on this knowledge to improve the validity of the
measurements.
Some limitations of our study must be acknowledged.
First, our study should have included the front pants
pocket location to better compare results with the study
of Johnson et al. (2016). Second, our study examined the
validity of the Samsung Health application with a very
concrete test. Future studies should examine its validity
in other settings with different conditions, like on a tread-
mill or in natural settings. Third, the fixed central point
where the researcher should be located to record the
videos was not standardized. The consideration of this
aspect would have been desirable, although it does not
seem very relevant for step counting. Fourth, the inclusion
of another researcher to analyse the videos and a good
inter-rater reliability would have improved the rigor of our
protocol. Nevertheless, counting steps in a short slow-
motion video seems to be quite an easy task, with reduced
risk of error.
Disclosure statement
No potential conflict of interest was reported by the authors.
ORCID
Vicente J. Beltrán-Carrillo http://orcid.org/0000-0001-8794-5901
Alejandro Jiménez-Loaisa http://orcid.org/0000-0003-0033-2374
Miriam Alarcón-López http://orcid.org/0000-0001-9900-5335
Jose L. L. Elvira http://orcid.org/0000-0003-0821-9210
References
Aittasalo, M., Miilunpalo, S., Kukkonen-Harjula, K., & Pasanen, M. (2006).
A randomized intervention of physical activity promotion and
patient self-monitoring in primary health care. Preventive Medicine,
42, 40–46.
Åkerberg, A., Lindén, M., & Folke, M. (2012). How accurate are pedometer cell
phone applications? A comparison between six applications. Procedia
Technology, 5, 787–792.
Alexander, S., Cowburn, G., & Foster, C. (2006). Understanding participa-
tion in sport and physical activity among children and adults: A
review of qualitative studies. Health Education Research, 21(6), 826–
835.
Bergman, R. L., Spellman, J. W., Hall, M. E., & Bergman, S. M. (2012). Is there
a valid app for that? Validity of a free pedometer iPhone application.
Journal of Physical Activity and Health, 9(5), 670–676.
Bland, J. M., & Altman, D. G. (1986). Measuring health and disease assessing
agreement between methods of clinical measurement. Lancet, i, 307–310.
Borg, G. (1998). Perceived exertion and pain scales. Champaign: Human Kinetics.
Boyce, G., Padmasekara, G., & Blum, M. (2012). Accuracy of mobile phone
pedometer technology. Journal of Mobile Technology in Medicine, 1(2), 23–27.
Bravata, D. M., Smith-Spangler, C., Sundaram, V., Gienger, A. L., Lin, N., Lewis, R.,
. . . Sirard, J. R. (2007). Using pedometers to increase physical activity and
improve health: A systematic review. JAMA, 298(19), 2296–2304.
Caspersen, C. J., & Fulton, J. E. (2008). Epidemiology of walking and Type 2
diabetes. Medicine & Science in Sports & Exercise, 40(7), 519–528.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences.
Hillsdale, NJ: Lawrence Erlbaum Associates.
Crouter, S., Schneider, P., Karabulut, M., & Bassett, D. R. (2003). Validity of
10 electronic pedometers for measuring steps, distance, and energy
cost. Medicine & Science in Sports & Exercise, 35(8), 1455–1460.
Dondzila, C., Swartz, A., Miller, N., Lenz, E., & Strath, S. (2012). Accuracy of
uploadable pedometers in laboratory, overground, and free-living condi-
tions in young and older adults. International Journal of Behavioral Nutrition
and Physical Activity, 9, 143.
Dwyer, T., Pezic, A., Sun, C., Cochrane, J., Venn, A., Srikanth, V., . . .
Ponsonby, A. L. (2015). Objectively measured daily steps and subse-
quent long term all-cause mortality: The Tasped prospective cohort
study. PLoS One, 10(11), e0141274.
Eyler, A. A., Brownson, R. C., Bacak, S. J., & Housemann, R. A. (2003). The
epidemiology of walking for physical activity in the United States.
Medicine & Science in Sports & Exercise, 35(9), 1529–1536.
Gardner, P. J., & Campagna, P. D. (2011). Pedometers as measurement
tools and motivational devices: New insights for researchers and practi-
tioners. Health Promotion Practice, 12, 55–62.
Gu, X., Chen, Y., Jackson, A. W., & Zhang, T. (2018). Impact of a pedometer-
based goal-setting intervention on children’s motivation, motor com-
petence, and physical activity in physical education. Physical Education
and Sport Pedagogy, 23, 54–65.
Johnson, M., Turek, J., Dornfeld, C., Drews, J., & Hansen, N. (2016). Validity
of the Samsung Phone S health application for assessing steps and

energy expenditure during walking and running: Does phone place-
ment matter? Digital Health, 2, 1–8.
Lee, I. M., & Buchner, D. M. (2008). The importance of walking to public
health. Medicine & Science in Sports & Exercise, 40(7), 512–518.
Leong, J. Y., & Wong, J. E. (2016). Accuracy of three Android-based pedometer
applications in laboratory and free-living settings. Journal of Sports Sciences,
35, 14–21.
Meyns, P., Bruijn, S. M., & Duysens, J. (2013). The how and why of arm
swing during human walking. Gait & Posture, 38(4), 555–562.
Murtagh, E. M., Murphy, M. H., & Boone-Heinonen, J. (2010). Walking – The
first step in cardiovascular disease prevention. Current Opinion in
Cardiology, 25(5), 490–496.
Nakagawa, S., & Cuthill, I. C. (2007). Effect size, confidence interval and
statistical significance: A practical guide for biologists. Biological
Reviews, 82(4), 591–605.
Orr, K., Howe, H. S., Omran, J., Smith, K. A., Palmateer, T. M., Ma, A. E., &
Faulkner, G. (2015). Validity of smartphone pedometer applications.
BMC Research Notes, 8, 733.
JOURNAL OF SPORTS SCIENCES 7
Owen, N., Humpel, N., Leslie, N., Bauman, A., & Sallis, J. E. (2004).
Understanding environmental influences on walking; Review and
research agenda. American Journal of Preventive Medicine, 27, 67–
76.
Pillay, J. D., van der Ploeg, H. P., Kolbe-Alexander, T. L., Proper, K. I., van
Stralen, M., Tomaz, S. A., . . . Lambert, E. V. (2015). The association
between daily steps and health, and the mediating role of body
composition: A pedometer-based, cross-sectional study in an employed
South African population. BMC Public Health, 15, 174.
Tudor-Locke, C., Sisson, S. B., Lee, S. M., Craig, C. L., Plotnikoff, R. C., &
Bauman, A. (2006). Evaluation of quality of commercial pedometers.
Canadian Journal of Public Health, 97, S10–S15.
World Health Organization [WHO]. (2018). Physical activity. Retrieved from:
http://www.who.int/mediacentre/factsheets/fs385/en/
Yuenyongchaiwat, K. (2016). Effects of 10,000 steps a day on physical and
mental health in overweight participants in a community setting: A pre-
liminary study. Brazilian Journal of Physical Therapy. doi:10.1590/bjpt-
rbf.2014.0160