This paper deals with problems of measuring change in motor behavior. Conventional measurement procedures and statistical analyses involving change are presented in the first section. This section discusses difference scores as the criterion measure and the use of all scores as the dependent variables. This latter category involves using either univariate or multivariate analysis of variance (ANOVA, MANOVA). The author gives suggestions for generally conservative researchers who want to use the methods discussed, but the second section describes alternative methods for analyzing change. The statistical techniques described are gradually being adopted by associated disciplines and are probably more appropriate for describing and predicting performance over time. This second section includes stochastic processes, time-series, factor analytic models of change, and curve-fitting as a change indicator. An appendix provides empirical comparisons among the numerous statistical methods described. (PB)