Acquisition and extinction phenomena in human trial-and-error learning under different schedules of reinforcing feedback.

Abstract

The present program of research on non-verbal selective learning (Noble, 1957a) considers reinforcement ( G ) factors as a major class of experimental variables in human learning and performance. This paper describes an investigation designed to obtain quantitative information about the role of reinforcing feedback schedules in four-choice serial trial-and-error learning by human Ss who performed on the Selective Mathometer. Two widely-accepted generalizations about the effects of partial reinforcement in simple learning situations are ( 1 ) that acquisition rates are faster the higher [he relative frequency of reinforcement and ( 2 ) that resistance to extinction is greater the lower the relative frequency of reinforcement during original learning (Jenkins & Stanley, 1950; Lewis, 1960). Since these two principles are complicated by interactions with task factors, amount of practice, and species differences (Spence, 1960, pp. 91-112), it is necessary to test them under the specific boundary conditions of this research program where human Ss are engaged in complex learning problems for incentives which are not primary. The most comprehensive review of motor-skills learning to date (E. A. Bilodeau & I. McD. Bilodeau, 1961) emphasizes the paucity of our knowledge about frequency schedules in this area.