Two criteria reported in recent literature were compared with the criterion of 95% probability (using analysis of variance) for their abilities to detect significant differences in M-waves due to repositioning of the ankle joint during human soleus H-reflex testing. Data from five normal adult males were analyzed using all three rejection criteria, and results from the three analyses were compared. Variability of M-wave amplitudes within a condition was found to be quite small. Commonly used M-wave rejection criteria which do not account for within-condition variability of M-waves were shown to do a relatively poor job of detecting alterations of M-wave amplitude due to a change in ankle position.