Positive Contrast following a Shift from Partial Punishment to Continuous Reinforcement

Abstract

The primary emphasis in the area of incentive contrast has centered around the effects of shifts in reward magnitude. However, with increasing evidence of the broad nature of contrast phenomena, investigators have explored the effect under other experimental contingencies. One procedure of potential theoretical interest deals with shifts in percentage or frequency of reward. Lehr (1974) and Shanab, Birnbaum, and Cavallaro (1974) have recently found positive contrast following shifts from partial reinforcement to continuous reinforcement schedules. The purpose of the present study was to expand the scope of research on percentage shift to include an examination of preshift incentive level by introducing punishment as a partial determinant of incentive. Twenty-four albino rats, 100 days old, were initially given 40 trials of continuous reinforcement (10 .045-gm. pellets on each trial) in a straight, 1.2-m., wooden alley (4 trials/day). All subjects received a constant 20-sec. delay of reward throughout the experiment to hold down running speed and allow for the emergence of positive contrast. The alley was 10 cm wide, 13.75 cm high, and covered by a hinged Plexiglas top. During a subsequent 40-trial preshift-phase 3 conditions (8 subjects/group) experienced different reward schedules. As a control, continuous reinforced subjects received 10 .045-gm. Noyes pellets per trial throughout the experiment (4 trials/day). Subjects partially reinforced received 2 reinforced and 2 nonreinforced trials each day and subjects partially punished received 2 rewarded and 2 punished trials each day (punishment was superimposed on nonreinforcement). During this phase, the alley was modified such that bIack and white vertical stripes were placed over the grey interior and a grid floor was installed in the delay and goal sections of the alley. A Lehigh Valley, Model 1531 shock unit was employed to administer a brief (.5-sec. duration) electric shock at 2.5 mA. Following preshift training, in order to avoid the possible contaminating influence of conditioned fear, the alley was reconverted to grey interior and wooden floor and all subjects shifted to delayed continuous reinforcement (10 pellets on each trial). This postshift phase lasted 32 trials (4 trials/day). A 3 groups X 8 blocks of 4 trials repeated-measures analysis of variance was performed on total speeds during postshift. The finding of greatest interest was a significant interaction of groups X blocks (K,,I,, = 1.82, p < .05). Newman-Keuls post hoc procedure indicated that a significant positive contrast effect emerged on Blodc 4; the condition of partial punishment was superior to those of partial and continuous reinforcement, which did not differ (p < .05). Since parcial punishment would have been expected to have had a lower preshift-incentive level than partial reinforcement, the present experiment supporcs the position that in studies of percentage shift, preshift-incentive level determines the magnitude of contrast in postshift.