Inter-response time shaping by percentile schedule with discrete trial procedure

Abstract

The percentile schedule has been used as a convenient procedure for the automatic shaping of operant responses. Experiments 1, 2, and 4 of this study examined whether two variables related to the percentile schedule—“update the reinforcement criterion for each response” and “retreat of the reinforcement criterion”—are critical factors of the percentile schedule on shaping. Experiment 3 examined the effect of combining one of the shaping rule “reinforce movement, not position” with the percentile schedule. Eight pigeons were exposed to a discrete trial procedure. In each trial, pigeons were required to peck one key and then to the other, and the interval between pecks was defined as the inter-response time (IRT). A percentile schedule was applied to this IRT aiming to lengthen it. The results showed that (1) IRTs were lengthened by the percentile schedule; (2) "update the reinforcement criterion for each response," "retreat of the reinforcement criterion," and "reinforce movement, not position" had no significant effect on lengthening IRT; and (3) the frequency of IRT occurrence decreased when the "retreat of the reinforcement criterion" was removed. These results indicate that the aforementioned percentile schedule variables are not critical for the percentile schedule on IRT shaping. Because the discrete trial procedure can dissociate the effects of variables on the length and frequency of IRT, this procedure seems effective for future shaping studies.