The present research was undertaken to investigate the effects of two independent variables considered potentially important to the reduction of fear through systematic desensitization. The first independent variable investigated was the importance of making covert motor responses when instructions were given to imagine motor behavior. Electromyographic measures were obtained on subjects' covert muscular activity as they imagined themselves raising their arms. The subjects were then classified, on the basis of their average electromyographic responsiveness, as high-responders (those showing relatively high levels when imaging movement) and low-responders. A 2 X 2 analysis of covariance showed a significant difference in the posttreatment scores of the high- and low responders on performance measures, the high-responders performing better. The difference between the two instruction groups was not statistically significant. Additional analyses indicate the importance of the instruction variable is mitigated by the extent to which subjects actually follow the particular instructions given. These data imply electromyography may be used in clinical practice as an assessment tool to determine which subjects are likely to respond to systematic desensitization. They also suggest the possibility of using response measures to train self-monitoring of imagery. Concerning the imagery construct, the present study offers an empirical alternative to the traditional conceptualization. The data obtained support the utility of the analysis of imagery advanced in this paper.