Feedback in behavior modification: an experimental analysis in two phobic cases.

Two illustrations of single-case research are described in which an isolated therapeutic variable was sequentially introduced, withdrawn, and reintroduced while changes in a clinically relevant behavior were measured. A claustrophobic patient and a knife-phobic patient received graduated practice in facing their phobic stimuli; length of time the claustrophobic patient stayed in a small dark room per trial, and length of time the knife-phobic patient kept knife exposed per trial were measured. In both experiments, when feedback of these time scores was withdrawn, ongoing progress was retarded. Reinstatement of feedback led to renewed improvement. In Experiment 2, adding and removing contingent verbal praise against a constant background of precise feedback did not significantly alter rate of progress.