Partial Reinforcement Effects as a Function of Number of Acquisition Trials and Percent Reinforcement

Abstract

Recent investigations have indicated that resistance to extinction is an inverted U-shaped function of percentage of reinforcement. Grant and Schipper (1952) suggest that two processes are involved in producing this function: discrimination and learning. The greater the percentage of reinforcement during acquisition, the more discriminable is the acquisition phase from the extinction phase, thereby, producing fewer extinction responses. On the other hand with the CR near zero strength, the greater the percent reinforcement, the greater the effect on habit strength and the greater the number of extinction responses. Lewis (1960, p. 3 ) has pointed out that, if Grant and Schipper are correct, the point of inflection of the inverted U-shaped function would need to vary with the degree of learning. The greater the degree of learning, the more the poinc of inflection should move toward the low end of rhe percentage scale. . The present study1 was an attempt to test this hypothesis in a classical aversive-conditioning paradigm measuring resistance to extinction of the GSR. The general design was to test independent groups under one of six treatments utilizing three percentages of reinforcement ( lo%, 30%, 100% ) and two levels of acquisition ( 10, 30 trials). Procedz~re.-Thirty male undergraduate students at the University of Maryland were assigned to six groups with five in each group. Recording and shock electrodes were placed on the palmar surfaces of the hands and the fingers, respectively. S was told to get comfortable and that he would receive some electric shocks during the course of the experiment. He was further told he could terminate the experiment whenever he wished; but none chose to do so. Next, the shock level was obtained by determining when S could just noticeably detect any stimulation and then increasing the number of turns on the inductoriurn by one and one half. The shock was then presented and maintained at this level for the course of the study. The CS was a 5-sec. 1000-cps tone at 40 db that terminated with a .5-sec. electric shock. The 2 )( 3 design yielded six independent groups which received either 10 or 30 presentations of the CS paired with the UCS on lo%, 30%, or 100% of the trials. The extinction series began immediately after