Button pressing by 44 college students intermittently produced points and the words "GOOD" or "POOR" on a computer screen. The events were arranged according to a paced random-interval 10-s schedule in which the target interresponse-time categories were 1 to 3, 3 to 5, or 6 to 12 s. The degree to which instructions specified certain aspects of the contingency (e.g., whether response spacing was critical) was also varied, and in some conditions the experimenter prompted specifically paced responses during the first 2 min of the session. The procedures shaped the local patterning of behavior of some subjects in less than 30 min of exposure to the contingencies. Most subjects who, in a postexperimental questionnaire, accurately identified the schedule contingencies also responded more accurately than those whose verbal descriptions were inaccurate or imprecise.
