Reliability and validity of an accelerometric system for assessing vertical jumping performance

Abstract

The validity of an accelerometric system (Myotest©) for assessing vertical jump height, vertical force and power, leg stiffness and reactivity index was examined. 20 healthy males performed 3ד5 hops in place”, 3ד1 squat jump” and 3× “1 countermovement jump” during 2 test-retest sessions. The variables were simultaneously assessed using an accelerometer and a force platform at a frequency of 0.5 and 1 kHz, respectively. Both reliability and validity of the accelerometric system were studied. No significant differences between test and retest data were found (p<0.05), showing a high level of reliability. Besides, moderate to high intraclass correlation coefficients (ICCs) (from 0.74 to 0.96) were obtained for all variables whereas weak to moderate ICCs (from 0.29 to 0.79) were obtained for force and power during the countermovement jump. With regards to validity, the difference between the two devices was not significant for 5 hops in place height (1.8 cm), force during squat (-1.4 N · kg-1) and countermovement (0.1 N · kg-1) jumps, leg stiffness (7.8 kN · m-1) and reactivity index (0.4). So, the measurements of these variables with this accelerometer are valid, which is not the case for the other variables. The main causes of non-validity for velocity, power and contact time assessment are temporal biases of the takeoff and touchdown moments detection.

Full text

Biology of Sport, Vol. 31 No1, 2014 55
Validation of an accelerometric device
Reprint request to:
Mohamed-Amine Choukou
Université de Paris Sud, Bât 335
- 91 405 Orsay Cedex
Phone number:
33 (0)1 69 15 73 81
Email: choukouamine@gmail.com
Accepted
for publication
21.12.2013
INTRODUCTION
Accurate assessment of biomechanical properties of the human
lower limb in eld conditions interests not only sport scientists,
but also coaches and practitioners since it reects, for instance,
the efciency of training programmes. For that aim, sport experts
typically use valid laboratory-based instruments such as the
different types of force platforms (PF) [1,9,14,19,22-24,26,32],
photoelectric cells [6,10,21] and contact mats [5,16,34].
Nowadays, ever-expanding devices make it possible to assess
lower limb properties in eld conditions. One of these measurement
tools is the Myotest® (Myotest SA, Switzerland), which consists of
a transportable and autonomous 3D accelerometric system (AS).
AS is more involved than just acquiring and recording signals.
It is a data logger allowing one to instantaneously evaluate the
following variables from acceleration data:
a. jumping height (H),
b. vertical force (Fv) and power (P),
c. leg stiffness (kleg) and reactivity index (RI).
Accuracy of AS has been recently studied in the literature,
showing comparison with photoelectric cells for jump height
RELIABILITY AND VALIDITY OF AN ACCELE-
ROMETRIC SYSTEM FOR ASSESSING VERTICAL
JUMPING PERFORMANCE
AUTHORS: Choukou M.-A.1,2, Laffaye G.1, Taiar R.2
1 Laboratoire Contrôle Moteur et Perception, Université de Paris Sud
2 Laboratoire de Biomécanique, Université de Reims Champagne Ardenne
ABSTRACT: The validity of an accelerometric system (Myotest©) for assessing vertical jump height, vertical force
and power, leg stiffness and reactivity index was examined. 20 healthy males performed 3ד5 hops in place”,
3ד1 squat jump” and 3× “1 countermovement jump” during 2 test-retest sessions. The variables were
simultaneously assessed using an accelerometer and a force platform at a frequency of 0.5 and 1 kHz, respectively.
Both reliability and validity of the accelerometric system were studied. No signicant differences between test
and retest data were found (p<0.05), showing a high level of reliability. Besides, moderate to high intraclass
correlation coefcients (ICCs) (from 0.74 to 0.96) were obtained for all variables whereas weak to moderate
ICCs (from 0.29 to 0.79) were obtained for force and power during the countermovement jump. With regards
to validity, the difference between the two devices was not signicant for 5 hops in place height (1.8cm), force
during squat (-1.4N · kg-1) and countermovement (0.1N · kg-1) jumps, leg stiffness (7.8kN · m-1) and reactivity
index (0.4). So, the measurements of these variables with this accelerometer are valid, which is not the case for
the other variables. The main causes of non-validity for velocity, power and contact time assessment are temporal
biases of the takeoff and touchdown moments detection.
KEY WORDS: measurement, biomechanics, precision, leg stiffness, in-situ
assessment

[10,33], and with a force plate for assessing the
force and power during squat and bench press [15]. However,
comparison of AS and PF has never been done to demonstrate
the quality level of the AS measurements compared to PF. For that
aim, sport experts typically use valid laboratory-based instruments
such as the different types of force platforms

[1,9,12,17,20-
22,24,32]. Moreover, the reliability and validity of AS for
assessing leg stiffness and reactivity need to be investigated.
Basically, vertical jump performance corresponds to the difference
between the centre of mass position at the standing posture and its
position at the peak of the jump, which could be estimated using
the ight time (FT) method [5,20,29]. Fv corresponds to product of
body mass (m) and vertical acceleration (av) according to Newton’s
Second Law. Besides, power is equal to the product of force and
velocity, which are both measurable from acceleration data. As regards
leg stiffness, it corresponds to the ratio of Fv to the displacement(∆CoM)
of the centre of mass (CoM) according to the widely used spring-mass
model of McMahon et al. [30]. The latter considers the human
lower limb as a linear vertical spring supporting the whole body
Original Paper Biol. Sport 2014;31:55-62
DOI: 10.5604/20831862.1086733

56
Choukou M.-A. et al.
mass(i.e. m) and that the actions of the lower limb segments are
integrated once. Thus, the whole lower limb behaves like a linear
mechanical spring, that is, the spring constant (k) represents the
lower limb stiffness (i.e. kleg).
Before using the AS for scientic purposes, it would be essential
to verify its ability to reect what it is designed to measure [4].
Therefore, the aim of this study was to investigate the reliability and
validity of the accelerometric system for assessing H, F and P as well
as kleg and RI. Three types of standard vertical jump tasks were
proposed for examining the device: 5 maximal hopping in place (5H),
1 single countermovement jump (CMJ), and 1 single squat jump
(SJ). In this perspective, the different measurements obtained by the
AS were compared to those obtained by the PF.
MATERIALS AND METHODS
Participants. Twenty males took part in this study. The participants
were physical education students and physiotherapists (age: 27 ±
6

years, body mass: 74.52 ± 7.16 kg and height: 1.78 ± 0.06 m).
They were all amateur sportsman who train once or twice per week.
None of them was involved in a jump-based activity. Subjects refrained
from drinking alcohol or caffeine-containing beverages for 24 hours
before testing, to avoid any interference in the experiment. Each
subject completed all trials in the same time period of test days to
eliminate any inuence of circadian variation. The temperature of the
room was the same at each session (22°C). The experimental proto-
col was approved by the ethics committee of Université Paris-Sud
and according to the ethical principles laid out in the 2013 revision
of the Declaration of Helsinki. All participants gave their written con-
sent to the experiment after having been informed of the aims and
the risks of testing procedures. In addition, they kindly accepted to
wear the same clothes and shoes for both test and retest sessions.
Procedures
The experiment consists of two identical test and retest sessions
separated by 2-3 days. For both sessions, the participants were
tested by the same experimenters and at the same hour of the day
in order to control the circadian uctuation [3]. Each session consists
of three repetitions of each of the following tasks: 5H, a single SJ
and a single CMJ. Participants were equipped with a Myotest® device
(length × width × depth: 9.5 × 5 × 1 cm, mass: 60g). The device
was attached to a belt and vertically xed on the middle of the
lower back (Figure 1). The trials were simultaneously recorded by
the accelerometric system at a sampling frequency of 500 Hz and
by a 0.4×0.4 m force plate (AMTI OR 6-5, Watertown, MA, USA)
at a sampling frequency of 1000 Hz (Figure 1). Before each trial,
they were asked to stand over the PF assuming a vertical posture,
as well as to keep hands placed on their waist during the three
jumping conditions in order to avoid upper-body interference caused
by arm swing [27]. After the touch-down of each of the tasks, the
participants were instructed to reassume a vertical standing posture
and to wait for the nal acoustic signal.
The rest between two consecutive jump trials of the same set was
approximately 30 seconds and the rest between sets (5H, SJ, or
CMJ) was 3 minutes. After performing their standardized warm-up
and prior testing, the subjects completed familiarization trials for 5H,
SJ and CMJ by following instructions and feedback given by the
experimenters. Only successful trials were taken into account. The
participants were kindly asked to respect the protocol and to repeat
the trial if a jump was incorrectly performed. This validation protocol
respected the recommendations of Atkinson and Nevill [4].
Tasks
•
5H protocol: For the 5H test, the participants were asked
to hop in place 6 times as high as possible while reducing
the ground contact time [16]. The rst hop served as a CMJ
(impetus) and was consequently excluded from analysis. The
remaining 5 effective jumps were retained and averaged for
analysis (mean of the 5 hops). The instructions given before
the 5H test were as follows: “Upon the acoustic signal,
perform an initial countermovement jump (impetus), after
which perform 6 hops in place, with minimal knee exion
and a maximal jumping height. After the 6th jump, reassume
a vertical standing posture and wait for nal acoustic signal.”
Multiple trials were performed under researcher supervision
in order to familiarize the participants with this kind of
hopping task and to optimize the leg stiffness by reducing
the effect of technique. The recording of data began only if
the technique of bouncing was acquired.
• CMJ protocol: In order to perform a countermovement
jump, the participants were instructed to freely ex the
knees and to jump once as high as possible. This procedure
corresponds to the instructions advised by the manufacturer.
• SJ protocol: For the squat jump test, the participants were
asked to reach and hold a semi-squat position [~ 90° knee
FIG.
1.
STANDING POSITION AT THE BEGINNING OF ALL JUMP TASKS
(LEFT SIDE) AND SEMI-SQUAT POSITION REACHED AND HELD DURING
SQUAT JUMP TEST (RIGHT SIDE).
Note: The gure shows the set square used to control the knee angle during
SJ and the attachment of the accelerometric system to the lower back

Biology of Sport, Vol. 31 No1, 2014
57
Validation of an accelerometric device
exion controlled by a 0.4×0.4 m set square (maintained
by the experimenter) as biofeedback] (Figure 1) until an
acoustic signal was given, and to jump once as high as
possible without performing any countermovement before
jumping.
Jump height assessment
The vertical jump height was assessed using the FT data [5,20,
29], as follows:
(in cm) (Equation 1);
where g = acceleration due to gravity.
For PF measurements, FT corresponds to the lapse of time when
the vertical ground reaction force is equal to zero. However,
AS considers the FT as time duration that elapses between the mo-
ment of maximal vertical velocity (before take-off) and the moment
of minimal velocity after touch-down (tvmin afterpeak). Then the vertical
jump height is estimated by AS as follows:
(in cm) (Equation 2);
Vertical force and power assessment
Vertical force (Equation 3) and power (Equation 4) were assessed
using the following equations:
in N · kg-1 (Equation 3)
in W · kg-1 (Equation 4)
The vertical velocity (vv) was calculated from the integration of av
data as proposed by Cavagna for the force platform [11] and as
proposed by the device’s manufacturer for the accelerometer as fol-
lows:
For PF measurements: in cm · s-1. (Equation 5)
For AS: in cm · s-1.
To reduce the error due to the integration process, the frequency
of acquisition for both devices was calibrated on the highest possible
value: 1000Hz for the force platform and 500 Hz for the acceler-
ometer.
Leg stiffness and reactivity index
For PF measurements, leg stiffness (kN · m-1) was calculated as the
ratio of maximal Fv (in kN) to ∆CoM [30]. However, for AS, leg stiff-
ness was calculated as the ratio of concentric force (when vv is
equal to zero) to ∆CoM, as proposed by Dalleau et al. [16]. ∆CoM
was calculated by integrating vv during the grounding phase from
its minimal position (i.e. tvmin afterpeak) to its zero position (v0).
In order to check the linearity of the lower limb movements and
its accordance with theoretical linear spring behaviour, the linear-
ity of the curve of Fv in function of ∆CoM was veried (Figure 2).
An r²>.80 was chosen as a threshold to consider the bouncing
behaviour as a linear spring oscillation. All the retained jumps met
this criterion.
Reactivity index corresponds to the ratio of FT to contact time(CT).
CT corresponds to the time of presence of a ground reaction force
signal over a jump (oscillation period) for PF measurement, whereas
it corresponds to the time that elapses from the position of the
maximal velocity (
maxv
t
) to
afterpeakv
t
min
(see Equation 2).
Statistical analysis
All descriptive statistics were used to verify whether the basic as-
sumption of normality of all studied variables was met. Shapiro-Wilk
tests revealed no abnormal data pattern. The statistical tests were
processed via SPSS® (version 16.0, Chicago, IL). In addition, statis-
tical power and effect sizes were calculated using G*Power 3. Statis-
tical power was 1 for all jump modalities with a sample size inferior
to 20 subjects and large effect sizes.
The test-retest reliability of the accelerometric system was assessed
with the intraclass correlation coefcient (ICC) (2, 1) (relative reli-
ability) [8] in order to describe how strongly individual scores in the
same session and throughout test and retest sessions resembled each
other. An ICC of r=0.8 represents good agreement, and a value r>0.9
is considered to indicate excellent agreement [18]. Coefcients of
variation (CV %) were also calculated to measure the dispersion of
the scores of the test and retest. A coefcient of variation CV ≤ 10%
was interpreted as an insignicant difference between test and retest
sessions [4]. Besides, the method of Bland and Altman (absolute
reliability) [7] allowed determination of test-retest systematic bias ±
random error as well as lower and upper limits of agreement (LoA).
According to Atkinson and Nevill, systematic bias refers to the gen-
eral trend for the measurements to be different in a particular direction
(either positive: upper LoA or negative: lower LoA) whereas the random
error refers to the degree to which the repeated measurements vary
for the individuals [4]. Paired Student T-tests were used to detect any
signicant systematic bias between the scores of the two sessions
(test and retest).
FIG.
2.
TYPICAL SHAPE OF EXPERIMENTAL VERTICAL FORCE TO CENTRE
OF MASS DISPLACEMENT CURVE, REPRESENTING A TYPICAL LOWER
LIMB FLEXION-EXTENSION DURING THE HOPPING IN PLACE TEST.
Note: The dotted line represents the leg stiffness.

58
Choukou M.-A. et al.
The concurrent validity was assessed using ICCs (2, 1) [8] in order
to describe how strongly individual scores obtained by the two
methods resembled each other. The Bland-Altman method allowed
determination of systematic bias between the accelerometric
system and the force platform (± random error) and the lower
and upper LoA [7]. Besides, coefcients of correlation (R2)
of the between-device differences were plotted. The level of
heteroscedasticity was set at R2 = 0.1; thus, a coefcient of
correlation less than 0.1 (R2 <0.1) means that the variables are
homoscedastic [4]. Additionally, independent-samples Student
T-tests were used in order to detect any signicant systematic bias
between AS and PF data at p<0.05
.
RESULTS
The results are shown in Table 1 and Table 2.
Test-retest reliability
No signicant differences between the test and retest were
reported for all studied variables (p>0.05) (Table 1). All CVs were
lower than 10% for all studied variables except for Vcmj and Pcmj,
which were 11.09% and 13.36%, respectively. Besides, the ICC
values were between 0.74 and 0.89 for jumping heights, and
0.86 and 0.96 for reactivity index and leg stiffness, which was
not the case for force and power during the countermovement
jump (0.29 < ICCs < 0.79).
CV% ICC (95% CI) Systematic Bias Random Error Student T test
Jump Height
5H-H (cm) 6.42 0.74 - 0.85 1.1 4.4 NS
SJ-H (cm) 4.25 0.82 - 0.84 -1 6.2 NS
CMJ-H (cm) 4.31 0.80 - 0.89 - 0.2 4.8 NS
Force, Velocity & Power
Fsj (N · kg-1) 3.30 0.85 - 0.92 - 0.5 1.7 NS
Vsj (cm · s-1) 6.41 0.85 - 0.92 - 7.8 29.7 NS
Psj (W · kg-1) 6.03 0.74 - 0.83 - 1.8 6.4 NS
Fcmj (N · kg-1) 4.24 0.66 - 0.79 - 0.6 2.3 NS
Vcmj (cm · s-1) 11.09 0.66 - 0.42 3.8 22.1 NS
Pcmj (W · kg-1) 13.36 0.29 - 0.45 2.8 6.9 NS
Contact Time, Leg Stiffness & Reactivity Index
CT (ms) 5.70 0.88 - 0.93 + 3.9 19.4 NS
IR 7.86 0.94 - 0.96 - 0.2 0.3 NS
kleg (kN · m-1) 6.03 0.86 - 0.92 + 2.8 8 NS
ICC
(95 % CI)
Systematic Bias
(cm)
Random Error
(cm)
Lower LoA
(cm)
Upper LoA
(cm)
Jump Height
5H-H (cm) 0.9 - 0.94 + 1.8 ± 15.3 -13.4 17.1
SJ-H (cm) 0.71 - 0.79 + 5.6 * ± 11.7 -6.1 17.4
CMJ-H (cm) 0.79 - 0.86 + 3.6 * ± 13.1 -10.7 17.4
Force, Velocity & Power
Fsj (N · kg-1) 0.63 - 0.78 - 1.4 ± 2.4 - 6.6 3.8
Vsj (cm · s-1) 0.32 - 0.35 + 11.1 * ± 4.4 - 12.9 35.1
Psj (W · kg-1) 0.18 - 0.31 + 11.7 * ± 16.9 - 22.4 46
Fcmj (N · kg-1) 0.68 - 0.79 + 0.1 ± 3 - 6.3 6.6
Vcmj (cm · s-1) 0.37 - 0.47 + 15.8 * ± 14.4 - 19.7 51.4
Pcmj (W · kg-1) 0.19 - 0.46 + 16.7 * ± 21.6 - 38 71.6
Contact Time, Leg Stiffness & Reactivity Index
CT (ms) 0.73 - 0.91 - 69 * ± 21 7 131
IR 0.74 - 0.80 + 0.4 ± 0.9 -1.5 2.4
kleg (kN · m-1) 0.76 - 0.87 + 7.8 ± 12.7 -23.6 39.3
TABLE 2.
CONCURRENT VALIDITY OF ACCELEROMETRIC SYSTEM VS FORCE PLATFORM
TABLE 1.
TEST-RETEST RELIABILITY OF ACCELEROMETRIC SYSTEM
Note: NS: no signicant difference between test and retest mean values (p < .05)
Note: the signes (+) and (-) respectively refer to a higher and a lower values of AS compared to the reference value obtained by PF. * Statistically
signicant systematic bias between both systems at p < .05

Biology of Sport, Vol. 31 No1, 2014
59
Validation of an accelerometric device
Concurrent validity
Regardless of signicance level, the mean values of AS were
higher than those of PF for all studied variables except force
during SJ and CT during hopping in place, as shown in Table 2.
The Student T-test showed signicant differences between AS
and PF for jump height during SJ (SJ-H) and CMJ (CMJ-H), and
for vertical velocity and power during SJ and CMJ. The difference
between both devices was also signicant for CT assessment with
lower values when using AS (Table 2).
DISCUSSION
The aim of this validation study was to investigate the reliability of
an autonomous and transportable accelerometric system, and its
validity compared to the force platform for estimating (a) vertical
jump height, (b) vertical force and power, and (c) leg stiffness and
reactivity index during vertical jump tasks.
AS showed high test-retest reliability (Table 1) for assessing (a),
(b) and (c). In addition, the results showed good CVs (< 10%) for
all studied variables, except for velocity and power during the coun-
termovement jump. The ICCs showed moderate to high values for
(a) [from 0.74 to 0.89], (c) [from 0.86 to 0.96] and force, velocity
and power during SJ [from 0.74 to 0.92], by following the criterion
of the literature regarding the magnitude of the group-levelcorrelation
[18]. Our results are in accordance with the literature regarding the
jumping height recorded during hopping in place (ICC: 0.86-0.96,
CV: 5.1%), SJ (ICC: 0.86-0.96, CV: 4.93%) and CMJ (ICC: 0.93–
0.98, CV: 3.62%) [10].
The results showed that AS is able to reproduce the same mea-
surement precisions at different moments for the above-mentioned
variables. Considering validity, PF and AS showed good accuracy as
demonstrated by good ICC (>0.73) and low bias (<1%) for 5H
height, and leg stiffness and reactivity index, moderate ICC (>0.63)
for force during SJ and CMJ, and insignicant T-test, which shows
a strong association with the reference method. What are the pos-
sible explanations of the lack of validity for the other parameters,
i.e. (a) SJ and CMJ height, (b) velocity and power during both SJ
and CMJ, and (c) CT during 5H?
AS validity for vertical jump height assessment
As regards jumping heights measurement, the systematic biases of
SJ height (5.63 cm) and CMJ height (3.66 cm) were signicant
(p<0.05), with weak to moderate ICC values (0.71<ICC<0.86).
These biases seem to be related to FT estimation, which was
different according to each assessment device. In the ight
time method (Equation 1) [5, 20], it is assumed that the CoM
position at takeoff is the same as the CoM position at landing. So,
the vertical jump height corresponds to the CoM elevation between
the instant of landing and the instant of takeoff—namely, the ight
time.
When using the force platform, FT is measured as the difference
between the two instants of “actual” take off and “actual” land-
ing(Figure 3); that is, when force is equal to zero [29]. This is not
the case for AS, which considers FT as the lapse of time between
the maximum value of positive velocity and the minimum value of
negative velocity, which are both accessible from the velocity-time
curve (Figure 3), thus estimating the “effective” takeoff and landing,
respectively. This method could induce bias of ight time measure-
ment between AS and PF. According to our data, a maximal veloc-
ity could be achieved at the end of the concentric phase shortly
before the actual takeoff (Figure 3). This could be considered as the
beginning of ight time, which induces a slight underestimation at
the start of the takeoff. That was also the case of the effective land-
ing, which occurs shortly after the actual landing, inducing a slight
overestimation at the start of the touchdown (Figure 3). This has
also been reported by Casartelli et al., who compared the jumping
heights obtained by AS to those obtained by photoelectric cells [10].
Both of these approximations involve an FT overestimation which
reects the difference of measurements between AS and PF.
Differences of AS validity level between the three jumping mo-
dalities are mainly dependent on the prior jumping concentric phase
(propulsive phase of the jump), which is specic to each type of
jump. As shown in the method section, the jumping height was
measured using the ight time data (Equation 2).
The sources of error could be the detection of the minimum value
of negative velocity (vminafterpeak), which is considered as the “effective
touchdown” while calculating FT, and the detection of the maximum
FIG.
3.
A COUNTERMOVEMENT JUMP AS RECORDED BY THE AS AND PF.
Note: The upper curve represents the force (Fz) and its corresponding instants
of takeoff and touchdown. The lower curve shows the velocity (which results
from the double integration of force) and its corresponding takeoff and
touchdown. The ight time is slightly overestimated when using velocity data.

60
Choukou M.-A. et al.
value of positive velocity (vmax), which is considered as the “effective
takeoff”. A maximal velocity could be achieved at the end of the
concentric phase shortly before the actual takeoff. This could be
considered as the beginning of the ight time, which induces a slight
underestimation of the instant of the takeoff. That was also the case
of the effective landing, which is considered to occur shortly after
the actual landing, inducing a slight overestimation of the instant of
touchdown, as previously reported by Casartelli et al., who compared
AS jumping scores to those obtained by photoelectric cells. Both of
these approximations involve an overestimation of ight time, which
could explain the countermovement jump height difference (3.6 cm).
It is important to mention that the mean value of CMJ height mea-
sured in this study was lower than the values obtained in university
students (45-46 cm) [31] and was close to the jumping height of
male rhythmic gymnasts (36 cm) [17], and 14-year-old boys (36.9
cm) using the Ergojump Bosco System [34], showing that our par-
ticipants achieved moderate CMJ heights. Based on our results, we
suggest that a similar overestimation would be observed in partici-
pants within the range of values from 22.6 to 51.1 cm, close to our
study. Additionally to its high reproducibility for assessing the coun-
termovement jump height, the validity of AS is deemed acceptable
by taking into account the amount of systematic bias (3-4 cm) re-
corded in this study, the variability of the jumping behaviours and
the practical purposes of this jumping test.
Therefore, the ight time is more likely to be the major source of
bias since it is dependent on the jumping modalities. Hopping in
place particularly required a very short contact time (about 90 ms in
this study), which is why AS encountered low probability to make
errors while detecting vmax. That is why the difference of the hopping
in place heights between the two devices was very low (1.8 cm),
showing that “effective takeoff” was close to “actual takeoff”. There-
fore, the hopping in place height could be estimated using an ac-
celerometer system with the insurance that the ight time is as near
as possible from the lapse of time between actual takeoff and touch-
down.
That was not the case of squat and countermovement jump heights
estimation, which showed high and signicant systematic biases
(Table 2). The protocol of squat and countermovement jumps could
be the reason for ight time overestimations. Indeed, knee angle has
to be 90° prior to jumping during SJ and was freely chosen during
CMJ. Mechanical variations due to these modalities seem to increase
the probability of error while detecting vmax, probably because of lon-
ger contact time due to knee exion compared to hopping in place. In
spite of the stabilization moment at the end of lowering during a squat
jump trial, the accelerometric system showed error while estimating
FT. The static nature of the squat jump does not seem to reduce
mechanical variability compared to the countermovement jump as
one might imagine. Both modalities affected the moment of vmax.
AS validity for force and power assessment
The difference of velocity and power estimations between AS and
PF was signicant (p < 0.05) with higher values coming from
the AS during the SJ task (11.1 cm and 11.7 cm, respectively)
as well as CMJ (15.8 and 16.7 cm, respectively), and weak
ICCs (from 0.18 to 0.47). The main reason seems to be related
to the heteroscedasticity of the data and the specicity of the
task. The data of velocity were homoscedastic (R2 = 0.01) for Vsj
and slightly heteroscedastic (R2 = 0.11) for Vcmj. The difference
between AS and PF was signicantly high (Table 2), showing
poor validity of AS for assessing velocity during squat and
countermovement jumps (p < 0.05).
The heteroscedasticity of velocity scores during the countermove-
ment jump could be explained by the specicity of the countermove-
ment jump technique. Moreover, this could be caused by the con-
straint of the task, which consisted in jumping with hands over the
waist over a force platform. The previous literature showed that,
without arm motion, the eccentric phase is used in order to maintain
the balance of the system rather than shortening this phase. Conse-
quently, it is difcult for AS to nd the real beginning of the concen-
tric phase, which affects the initial value of velocity [2, 25]. The poor
validity of AS for assessing power during squat and countermovement
jumps seems to be the direct consequence of biases in velocity as-
sessment since it corresponds to the product of force, which is esti-
mated at its just value, times the velocity which is overestimated
when assessed by AS.
AS validity for leg stiffness and reactivity assessment
AS was deemed valid for assessing kleg, as shown in Table 2.
However, its validity for assessing CT remains critical because
it systematically underestimated the “actual” values of CT by
69 ms. This was due to the CT calculation protocol applied
by AS. In this method, CT was considered as the lapse of time
between the minimal position of velocity after touchdown and its
maximal value during the successive takeoff, i.e., when force is
equal to body weight (F = m × g). This was not the case for PF,
FIG.
4.
COMPARISON BETWEEN THE ACTUAL AND EFFECTIVE CONTACT
TIME AS MEASURED BY THE FORCE PLATE AND THE ACCELEROMETRIC
SYSTEM, RESPECTIVELY.
Note: The dashed line represents the criteria of determination of touch-down
and take-off according to the “m × g” line as assumed by AS. The gure
shows that the accelerometric system underestimates the contact time
(effective) compared to the force platform (actual contact time).

Biology of Sport, Vol. 31 No1, 2014
61
Validation of an accelerometric device
1. Arampatzis A., Schade F., Walsh M.,
Bruggemann G.P. Inuence of leg
stiffness and its effect on myodynamic
jumping performance. J. Electromyogr.
Kinesiol. 2001;11:355-364.
2. Ashby B.M., Heegaard J.H. Role of arm
motion in the standing long jump. J.
Biomech. 2002;35:1631-1637.
3. Atkinson G., Reilly T. Circadian variation
in sports performance. Sports Med.
1996;21:292-312.
4. Atkinson G., Nevill A.M. Statistical
methods for assessing measurement
error (reliability) in variables relevant to
sports medicine. Sports Med.
1998;26:217-238.
5. Bosco C., Luhtanen P., Komi P.V.
A simple method for measurement of
mechanical power in jumping. Eur. J.
Appl. Physiol. Occupat. Physiol. 1983;
50:273-282.
6. Bosquet L., Berryman N., Dupuy O. A
comparison of 2 optical timing systems
designed to measure ight time and
contact time during jumping and
hopping. J. Strength. Cond. Res. 2009;
23:2660-2665.
7. Bland J.M., Altman D.G. Statistical
methods for assessing agreement
between two methods of clinical
measurement. Lancet 1986;1:307-
410.
8. Bland J.M., Altman D.G. A note on the
use of the intraclass correlation-coefcient
in the evaluation of agreement between
2 methods of measurement. Comput.
Biol. Med. 1990; 20:337-340.
9. Buśko K., Madej A., Mastalerz A. Effect
of the cycloergometer exercises on power
and jumping ability measured during
jumps performed on a dynamometric
platform. Biol. Sport 2010;27:35-40.
10. Casartelli N., Müller R., Mauletti N.
Validity and reliability of the Myotest
accelerometric system for the assessment
of vertical jump height. J. Strength Cond.
Res. 2010;24:3186-3193.
11. Cavagna G.A. Force plateforms as
ergometers. J. Appl. Physiol. 1975;
39:174-9.
12. Cavagna G.A. Symmetry and asymmetry
in bouncing gaits. Symmetry 20.12; 2:
1270-1321.
13.
Cavagna G.A., Franzetti P., Heglund N.C.,
Willems P. The determinants of the step
frequency in running, trotting and
hopping in man and other vertebrates. J.
Physiol. (London) 1988;399:81-92.
14. Chelly M.S., Denis C. Leg power and
hopping stiffness: relationship with sprint
running performance. Med. Sci. Sports.
Exerc. 2001;33:326-333.
15. Comstock B.A., Solomon-Hill G,
Flanagan S.D., Earp J.E., Luk H.Y.,
Dobbins K.A., et al. Validity of the
Myotest® in measuring force and power
production in the squat and bench press.
J. Strength. Cond. Res. 2011;25:2293-
2297.
16. Dalleau G., Belli A., Viale F., Lacour J.R.,
Bourdin M. A simple method for eld
measurements of leg stiffness in hopping.
Int. J. Sports. Med. 2004;25:170-176.
17. Di Cagno A., Baldari C., Battaglia C.,
Monteiro M.D., Pappalardo A., Piazza M.
Factors inuencing performance of
competitive and amateur rhythmic
gymnastics--gender differences. J.
Science. Med. Sport 2009;12:411-416.
18. Donner A., Eliasziw M. Sample size
requirements for reliability studies. Stat.
Meth 1987;6 :441-448.
19. Farley C.T., Morgenroth D.C. Leg stiffness
primarily depends on ankle stiffness
during human hopping. J. Biomech.
1999;32:267-273.
20. Frick U. Comparison of biomechanical
measuring procedures for the
determination of height achieved in
vertical jumps. Leistungsport
1991;21:448-453.
21. Glatthorn J.F., Gouge S., Nussbaumer S.,
Stauffacher S., Impellizzeri F.M.,
Mafuletti N.A. Validity and reliability of
Optojump photoelectric cells for
estimating vertical jump height. J.
Strength Cond. Res. 2011;25:556-560.
22. Hobara H., Inoue K., Muraoka T.,
Omuro K., Sakamoto M., Kanosue K. Leg
stiffness adjustment for a range of
hopping frequencies in humans. J.
Biomech. 2010;43:506-511.
23. Hobara H., Muraoka T., Omuro K.,
Gomi K., Sakamoto M., Inoue K.,
Kanosue K. Knee stiffness is a major
determinant of leg stiffness during
maximal hopping. J. Biomech.
2009;42:1768-1771.
24. Laffaye G., Taiar R., Bardy B.G. The
effect of instruction on leg stiffness
regulation in drop jump. Sci. Sports
2005;20:136-143.
25. Laffaye G., Bardy B., Taiar R.
Upper-limb motion and drop jump: effect
of expertise. J. Sports Med. Phy. Fit.
2006;46:238-247.
26. Laffaye G., Choukou M.A. Gender bias in
the effect of dropping height on jumping
performance in volleyball players. J.
Strength Cond. Res 2010;24:2143-
2148.
27. Lees A., Vanrenterghem J., de Clercq D.
Understanding how an arm swing
enhances performance in the vertical
jump. J. Biomech 2004;37:1929-
1940.
REFERENCES
which determines the “actual” CT, i.e. the time between actual
touch-down and take-off from the GRF-time curve, irrespective of
the magnitude of reaction force exerted on the ground [12,13].
In contrast, AS measures the “effective” CT, which is considered
as the period of time during which vertical force is equal to or
greater than body weight (F ≥ m × g) (Figure 4). Obviously, these
biases have a knock-on effect on the reactivity index since its
calculation uses the contact time data. To conclude, AS could
be used for assessing contact time by taking into account this
systematic underestimation.
CONCLUSIONS
The aim of this study was to evaluate the reliability of an
accelerometric system and its validity compared to a force
platform for assessing vertical jump performance. The results
showed a high level of reliability for assessing jumping height,
leg stiffness, reactivity index, velocity and power during squat
jump and force measurements using the accelerometric system.
However, force and power measurements were weakly reliable.
The accelerometric system was deemed valid for assessing
hopping in place height, force during squat and countermovement
jumps as well as leg stiffness and reactivity index. However, the
evaluation of jumping height, velocity and power during both
squat and countermovement jump was not valid. The main
causes of non-validity for velocity and power as well as contact
time assessment are due to biases occurring while detecting
the takeoff and touchdown moments. That being said, the
accelerometric system remains highly reliable for assessing the
studied variables. Thus, it could be useful notably to follow up
rehabilitation programmes or for long-term athletic monitoring.
Funding: This work has not received any funding resources
Conicts of interest: none

62
Choukou M.-A. et al.
28. Llyod R.S, Oliver J.L., Hughes M.G.,
Williams C.A. Reliability and validity of
eld-based measures of leg stiffness and
reactive strength index in youths. J.
Sports Sci 2009;27:1565-73.
29. Linthorne N.P. Analysis of standing
vertical jumps using a force platform.
Am. J. Phys 2001; 69: 1198-1204.
30. McMahon T.A., Valiant G., Frederick E.C.
Groucho running. J. Appl. Physiol. 1987;
62:2326–2337.
31.
Nuzzo J.L., Anning J.H., Scharfenberg J.M.
The reliability of three devices used for
measuring vertical jump height. J.
Strength Cond. Res. 2011;25:2580-
2590.