Parametric comparisons of quantified EMG activity are not valid across subjects because of individual differences in skin resistance, input impedance, amount of subcutaneous tissue, and muscle anatomy. Intrasubject parametric comparisons of integrated soleus muscle EMG data, therefore, were performed to obtain information on individual responses to the stretches. A separate ANOVA was performed on the integrated soleus muscle EMG data of each subject. Stretching conditions (4 levels), trials (3 levels), and periods (the time intervals between H-reflexes, 4 levels) were treated as between factors.Because the levels of the random factor in an ANOVA are assumed to be independent observations,29 the one-second time segment values were tested for independence and determined to be the appropriate random variable.