trending toward significancep = 0.096
Post hoc Conover–Iman tests for pairwise comparisons with multiplicity adjustment using false-discovery rate correction revealed that the classifier was also able to significantly differentiate between adjacent pairs of clinical scores for resting tremor constancy and resting tremor amplitude ( p ≤ 0.05), with the exception of scores 2 and 3 ( p = 0.067) and 3 and 4 ( p = 0.054) for resting tremor constancy (which were trending toward significance), and scores 0 and 1 ( p = 0.096) and 2 and 3 ( p = 0.1) for resting tremor amplitude.