Since the parameters derived from the uncorrected acceleration
data showed in general higher ICCs than the ones derived from the
corrected acceleration data, only those parameters were further
explored.
The reliability over the three, five, and ten trials did not
substantially differ among each other and therefore, subsequent
results were all based on the minimum amount of three trials. The
third to the fifth trials were used for statistical analysis because
these trials revealed slightly more reliable and valid measurements
than the first three trials.
Reliability values of smartphone and motion capture
measurements are presented in Table 1. Displacement parameters
of smartphone measurements showed good to excellent
reliability with ICCs between 0.71 and 0.80 while ICCs of time
variables ranged from 0.49 to 0.86 (fair to excellent). SEMs of
smartphone-derived CoM displacement parameters were substantially
larger than respective motion capture-derived SEMs,
while SEMs of time variables derived by both devices were
similar. The ANOVA revealed only a statistical significant
result for the smartphone-based variable Mima_disp_L, indicating that there was no significant systematic error for most
parameters.