Content uploaded by Iker J. Bautista
Author content
All content in this area was uploaded by Iker J. Bautista on Jul 30, 2019
Content may be subject to copyright.
Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=rspb20
Sports Biomechanics
ISSN: 1476-3141 (Print) 1752-6116 (Online) Journal homepage: https://www.tandfonline.com/loi/rspb20
Letter to editor
Iker J. Bautista & Fernando Martín
To cite this article: Iker J. Bautista & Fernando Martín (2019): Letter to editor, Sports
Biomechanics, DOI: 10.1080/14763141.2019.1640280
To link to this article: https://doi.org/10.1080/14763141.2019.1640280
Published online: 30 Jul 2019.
Submit your article to this journal
View Crossmark data
Letter to editor
We have read with great interest and care the study carried out by García-Orea,
Belando-Pedreño, Merino-Barrero, and Heredia-Elvar (2019) entitled ‘Validation of
an opto-electronic instrument for the measurement of execution velocity in squat
exercise’DOI: 10.1080/14763141.2019.1597156. We applaud the authors for their
thoughtful approach to the study. The general idea of the study is good and has an
important practical utility, unfortunately there are some aspects, regarding both statis-
tics and results that, in our modest opinion, need to be addressed.
The use of devices to control velocity execution while training with external resis-
tances is an interesting issue (García-Ramos et al., 2017; Spitz, Gonzalez, Ghigiarelli,
Sell, & Mangine, 2019). The cost of these devices ranges from the most economical (i.e.,
video camera and app) to the most expensive (T-Force system) (Courel-Ibañez et al.,
2019). However, regardless of the price, it is crucial that all devices that evaluate
execution velocity do so as accurately as possible. Systematic and/or random errors in
the measurement of the velocities at which the bar moves could generate large errors in
the determination of the intensity of the exercise according to its maximum repetition
(1RM) or when comparing the performance between different athletes.
Here is a list of what are, in our opinion, errors that we have detected in the study
mentioned above:
Page 3, ‘experimental design’section. The authors express that they have used
a unifactorial repeated subject design. That is not true. Two independent variables are
available in this study (i.e., load [6 levels] and devices [2 levels]). Therefore, the most
appropriate design for the data collection they have carried out would be a factorial design
(since there is more than 1 independent variable) of repeated measurements. At the
statistical level, this could be solved with an analysis of the variance of repeated measure-
ments (ANOVA RM [2, devices x 6, loads]) where we would find the following effects:
(i) Main effect of ‘device’. Regardless of the load factor, the ANOVA compares the
mean of all velocities across both devices (i.e., Velowin vs. T-Force). This is an
interesting effect.
(ii) Main effect of ‘load’. Regardless of the device factor, the ANOVA compares the
mean of all velocities across all loads (i.e., 20 vs. 30 vs. 40 . . .). This is not an
interesting effect.
(iii) Interaction ‘device x load’. The ANOVA compares the mean of all velocities
across load and intensity factors. This is an interesting effect.
It would be interesting for the reader to be able to interpret the results obtained in
terms of main effect of device ‘factor’and interaction effect ‘device x load’of the
aforementioned ANOVA.
SPORTS BIOMECHANICS
https://doi.org/10.1080/14763141.2019.1640280
© 2019 Informa UK Limited, trading as Taylor & Francis Group
Page 6, statistical analysis section. The authors express that they have used the
intraclass correlation coefficient (ICC). However, there are more than 6 different
types of ICC depending on the treatment that receives the term error, i.e., whether to
include the systematic and/or random error, if a 1-way or 2-way ANOVA is performed,
if a fixed or random model is used, and if the data come from one single measurement
or from an average of measurements (Weir, 2005). On the other hand, measurement
reliability is defined as stability against multiple repetitions at different times. As we
have been able to read in the study, no measurements have been made on different days
and only the best repetition of each load has been analysed. Therefore, we do not
understand how the authors can speak about reliability.
Page 6, statistical analysis section. The authors express that only the fastest repetition
at each load is used for the reliability analysis (page 7). They declare to have carried out,
among others, the ICC test, it is therefore impossible to evaluate the reliability of the
device, since to carry this out, at least two measurements should have been taken into
account for each load, and preferably in different experimental sessions. We understand
that the ICC has been made to study the concordance between both devices, but the
latter is not a reliability test, only a validation. Moreover, this is not the best statistical
procedure to evaluate that. Bland-Altman plot would be the best procedure to evaluate
the concordance between devices and also assess systematic and random error. Finally,
we understand that the best repetition on device 1 (e.g., T-Force) does not have to
match with the best repetition on device 2 (e.g., Velowin).
Page 7, the authors state that Pearson’s Correlation Coefficient showed values from
r= 0.70 to r= 0.96. This is not consistent with Lin’s CCC results, since Lin’s CCC value
can never be higher than Pearson’s Correlation Coefficient value, a fact that the same
authors take into account on page 6 when they explain the statistical procedures they
have performed.
Page 7, normal test results section. The authors state that the variable Mean Velocity
for Velowin did not have a normal distribution, but they express a p= 0.172, a value
that evidences the normality of the variable. This is probably a writing error.
Page 7, Table 2. The authors state ‘Table 2. Coefficient of Variation (CV), Intra-class
Correlation Coefficient (ICC) and Lin’s Concordance Coefficient (CCC) for each variable
with both devices in Squat exercise’. However, the variable ‘device’is not observed
anywhere.
Page 7, Table 2. The authors present the SEM values for each of the loads and as
a function of each of the variables analysed (i.e., mean velocity, mean propulsive
velocity and peak velocity). It is striking that the SEM average of all loads is equal to
0.296 m/s; that the ICC is, on average, 0.94; and that Pearson’s correlation coefficient
has fluctuated between 0.70 and 0.96 (the authors have not expressed in which loads
they have obtained such coefficients). The ICC, depending on which formula it is used,
may or may not include systematic error. However, the SEM is affected by the
systematic and random error. In this case, Pearson’s correlation coefficient cannot
detect systematic errors, but it can detect random errors (Weir, 2005). For this reason,
a type of statistical analysis such as the Bland-Altman graph is missing. Firstly, the error
could have been expressed in terms of systematic and random components. Secondly,
the presence of heterocedasticity could have been analysed.
2LETTER
Page 8, ‘ANOVA’section. As I have previously argued, the experimental design of
the cited study was an ANOVA RM (2 x 6). The authors should have expressed, for
each dependent variable (i.e., VM, VMP and VP), the statistics of Snedecor’s F, the
degrees of freedom and the effect size.
We firmly believe that the main problem lies in the experimental design of the manuscript
itself. On one hand, one thing is to evaluate the reliability of a device and/or evaluation
procedure and, on the other hand, to evaluate the concordance between different devices
(T-Force vs. Velowin). Due to all the arguments we have presented above, we consider that
the assertion made by the authors in the conclusions of the study ‘The main finding of this
study was the high reliability and concurrent validation of the Velowin opto-electronic system
for measuring the execution velocity’is not supported by the results presented in the manu-
script. We are not saying that such a device (i.e., Velowin) is not valid or reliable, however,
with the results obtained in the study, making such a claim is risky. In fact, a paper recently
published by Courel-Ibañez et al. (2019) reported a correlation coefficient between T-Force
and Velowin of r= 0.991 in full squat exercise and the limits of agreement’svalueswere
−0.08 m/s and 0.05 m/s for systematic and random error, respectively.
Disclosure statement
No potential conflict of interest was reported by the authors.
References
Courel-Ibañez, J., Martínez-Cava, A., Morán-Navarro, R., Escribano-Peñas, P., Chavarren-
Cabrero, J., González-Badillo, J. J., & Pallarés, J. G. (2019). Reproducibility and repeatability
of five different technologies for bar velocity measurement in resistance training. Annals of
Biomedical Engineering,47, 1523–1538. doi:10.1007/s10439-019-02265-6
García-Orea, G. P., Belando-Pedreño, N., Merino-Barrero, J. A., & Heredia-Elvar, J. R. (2019)
Validation of an opto-electronic instrument for the measurement of execution velocity in
squat exercise. Sports Biomechanics. doi:10.1080/14763141.2019.1597156
García-Ramos, A., Torrejón, A., Feriche., B., Morales-Artacho, A. J., Pérez-Castilla, A., Padial, P.,
&Haff,G.G.(2017). Prediction of the máximum number of repetitions and repetitions in
reserve from barbell velocity. International Journal of Sports Physiology and Performance,13,
353–359. doi:10.1123/ijspp.2017-0302
Spitz, E. W., Gonzalez, A. M., Ghigiarelli, J. J., Sell, K. M., & Mangine, G. T. (2019). Load-velocity
relationships of the back vs. Journal of Strength and Conditioning Research,32, 301–306.
doi:10.1519/JSC.0000000000002962
Weir, J. P. (2005). Quantifying tet-retest reliability using intraclass correlation coefficient and the
SEM. Journal of Strength and Conditioning Research,19, 231–240. doi:10.1519/15184.1
Iker J. Bautista
Physical Education and Sport Science, University of Granada, Granada, Spain
ikerugr@gmail.com http://orcid.org/0000-0002-7409-6290
Fernando Martín
Physical Education and Sports, University of Valencia, Valencia, Spain
http://orcid.org/0000-0003-1996-8276
Received 13 June 2019; Accepted 1 July 2019
SPORTS BIOMECHANICS 3
