Reliability was assessed by a three-layered approach recommended
by Weir [22]. A repeated-measures ANOVA was
performed in order to detect possible systematic error. Relative
and absolute intra-session reliability of uncorrected acceleration-,
corrected acceleration- and motion capture-derived parameters
were evaluated by calculating the intraclass correlation coefficient
(ICC (2,1)) and the standard error of measurement
(SEM = SDH(1 ICC)), respectively. In order to explore to what
minimum amount of trials the testing procedure can be reduced
while still yielding excellent repeatability, relative intra-session
reliability was calculated for ten, five, and three trials. ICC values
were interpreted according to recommendations of Cicchetti [23]
(ICC 0.75 rated as excellent).
For the evaluation of the concurrent validity, average values of
the data obtained by the two different measurement devices
(smartphone and motion capture system) were compared using
Pearson correlation coefficients (r) and Bland–Altman 95% limits of
agreement [24]. Correlation coefficients were interpreted according
to Zou et al. [25] (rating r = 0.5 as moderate and r = 0.8 as strong).
Bland–Altman plots were made in order to discuss possible
systematic deviations of results obtained by the smartphone.
Statistical analysis was performed using SPSS 21 (IBM Corp.,
Armonk, NY, USA) and level of statistical significance was set to
p = 0.05.