A source of possible error in the design of many motor short-term memory studies- the effect of repeated testing of the same criterion positions- was evaluated together with a secondary factor, motivation. Using an arm-positioning task, three groups of five subjects, representing different motivational levels, were tested repeatedly on the same three test positions. Constant and variable error measures indicated significant effects of the repetition factor. Subjects became more accurate with repeated testing of the same positions as reflected by a positive shift in constant error in the direction of the criterion. There was an initial decrease and subjects' motivational level. The apparent assumption in prior studies, that presentations of intervening test items obviates a repetition effect, was clearly refuted. Thus, in many motor short-term memory studies, treatment effects have been inextricably mixed with the effects of repetition.