*2.5. Statistical Analysis*

We recorded the within-day trial-to-trial intraclass correlation coefficients (ICC) and standard errors of measurement (SEM) [47] values for the participants wearing the different heel height (T3, T6, T9) vs. unshot in each plane of motion for PPH bone and IMTT bone in static and dynamic conditions. Landis and Koch [47] proposed that coefficients below 0.20 indicate slight agreement, and coefficients from 0.20 to 0.40 indicate fair reliability. Coefficients from 0.41 to 0.60 indicate moderate reliability, coefficients from 0.61 to 0.80 indicate substantial reliability, and coefficients from 0.81 to 1.00 indicate almost perfect reliability. We considered coefficients of 0.90 or larger to reflect a sufficient magnitude of reliability because they increase the likelihood that a measure is also reasonably valid. To check the concordance correlation between high heels' variables, Concordance Correlation Coefficient (CCC) [48] was done, where statistically significant *p*-values (<0.05) would mean a perfect correlation [48]. SEM was utilized to determine the minimal detectable change (MDC) for all evaluations. This was also considered the Reliable Change Index (RCI). We utilized the RCI as a statistical method to determine the clinical significance according to Jacobson and Truax [49].

An initial Kolmogorov-Smirnov test showed that the data was not normally distributed (*p* < 0.05). The *p*-values for multiple comparisons were corrected with a non-parametric paired Friedman test to prove that all high-heels variables were different. Bivariate correlations with a Wilcoxon test were carried out to determine whether there were significant differences between "unshod" vs. "with high heels of T3, T6 and T9" in static and dynamic conditions in FPH and IMTT bones; an alpha level of 0.01 was established for all tests of significance. All data were studied to establish the effect of heel height in 3 axes of movement with the medial line of the body taken as reference movement. In addition, Spearman's Rho (ρ) rank correlation coefficient both in static and dynamic conditions was done to check the possible relation between the height of the heels and the different bones' deviations. The results showed descriptive summaries as the mean ± SD. Analyses of total 96 variables led to *p*-values <0.05 (within a 95% confidence interval) that considered statistically significant. We conducted data analysis with SPSS software version 19.0 (SPSS Science, Chicago, IL, USA).
