Temporal comparator rules and responding in multiple schedules

Abstract

Three response rules for explaining the role of temporal factors in the control of responding were examined. These were the cycle-to-trial comparator rule from scalar expectancy theory (SET; Gibbon & Balsam, 1981), the "deletion" comparator rule proposed by Cooper, Aronson, Balsam, and Gibbon (1990), and a "bad/good" comparator rule-a type of ITI-to-trial comparator. Two of these rules were designed to explain the acquisition of responding in simple associative learning paradigms (i.e., autoshaping). Here, their generality as predictors of response levels in response-dependent multiple schedules was examined. SET's overall cycle-to-trial comparator rule was the best predictor of the pattern of responding. Contrary to previous findings regarding contrast in multiple schedules, which show greater contrast with shorter component durations, there was no effect of absolute component duration. As predicted by SET, relative, not absolute, component durations controlled response levels. © 1993 Psychonomic Society, Inc.