Four adult, chronic psychotic inpatients were observed before, during, and after the contingent application of time-out. Time-out was a part of their individual treatment programs, and consisted of placement in a furnitureless room for 15 minutes. Observations were made of 16 behaviors grouped into seven categories; these categories included both consequated and nonconse-quated behaviors. Observation sessions lasted for 15 minutes with interval recording of each behavior for the first 10 seconds of every 30-second block.Results for each subject were analyzed by means of three-way factorial ANOVAs, and indicated the following: a) three subjects exhibited significant differences in response frequency before, during, and after time-out that were consistent across all instances of their being placed in time-out; b) the particular behaviors that reliably varied with the observation conditions for one subject were not the same as those which reliably varied for another subject; c) two subjects significantly increased the frequency of their consequated responses during time-out in spite of the fact that time-out was effective as measured by a significantly lower emission rate after time-out compared to before time-out.