Statistical analysis

To maximize power for the small subject numbers in a specific group, we pre-processed the data as follows. Elevation of the left and right hindlimb were combined for straight line trotting by multiplying the difference in Pmin and Pmax between before and after hindlimb length elevation by -1. This results in positive changes in the lameness measure to indicate lameness in the elevated limb, and negative changes in the lameness measure to indicate lameness in the non-elevated limb. Lunging trials were combined into inside limb and outside limb elevation groups by multiplying all left lunge trials by -1. This results in positive changes in the lameness measure to indicate lameness in the inside limb and negative changes in the lameness measure to indicate lameness in the outside hindlimb. Changes in Pmin and Pmax with induced hindlimb length difference were assessed using a two-way repeated-measures ANOVA with trial (straight, lunge with inside limb elevation, lunge with outside limb elevation) and surface (hard, soft) as within-subject factors. Diagnostic checks on the residuals were carried out to check normality and constant variance assumptions. Multiple comparisons were carried out by using the t-test statistic with Bonferroni adjusted p-values. Significance level was set at 0.05.