Resistance to extinction in the rat as a function of percentage and distribution of reinforcement.

Abstract

5 groups of rats were trained in discrete trials to press a retractable lever. 1 group was consistently reinforced; for the other 4 (partially reinforced) groups, percentage and distribution of reinforcement were varied factorially. 2 effects of partial reinforcement appeared in the data of extinction: 1 which distinguished the consistent from all partial groups early in extinction, and a second which distinguished among the partial groups much later in extinction. For the partial groups, resistance to extinction increased with number of successive nonreinforcements in training, but it was not significantly affected by percentage of reinforcement alone, or by total number of nonreinforcements in training.

Increased resistance to extinction following partial reinforcement: A function of N-length or percentage of reinforcement?

“…However, that situation did not occur in the present study. A possible reason for the difference between the present study and the findings of Capaldi (1964) and Gonzalez & Bitterman (1964) could be, as suggested by Kotesky & Stetner (1968 ), differences in the length of the intertrial intervals (ITIs) used. This appears to be a consistent difference between the studies that report sequential effects and those that do not.…”

“…This appears to be a consistent difference between the studies that report sequential effects and those that do not. For example, the present study and the studies by Kotesky employed ITIs of 15 min or more; the studies by Gonzalez & Bitterman (1964) and those by Capaldi employed ITIs of 30 sec or less. Although Capaldi & Spivey (1964) found sequential effects tQ be operative following a 24-h ITI, Surridge & Amsel (1966) offered evidence suggesting that Rn may be regulated by sequential effects only under conditions of massed trials.…”

“…Specifically, it was not found that the groups with the larger values of N-length were more resistant to extinction. Gonzalez & Bitterman (1964) reported Rn to be a function of N•length but not significantly affected by percentage of reinforcement. They suggested that the inverse relation between Rn and percentage of reinforcement found previously may have been due to the usual inverse relation between N-Iength and percentage of reinforcement.…”

Comparative analysis of learning in honeybees

“…In rats, for example, the extent of single-alternation patterning varies markedly with effortfulness ofresponse and amount ofreward (Gonzalez, Bainbridge, & Bitterman, 1966). The Ishida experiment also failed to show increased resistance to extinction with longer runs ofN trials, a vertebrate phenomenon presumably generated by cumulative aftereffects ofN trials (Capaldi, 1964;Gonzalez & Bitterman, 1964). Foragers repeatedly fed to repletion on a distinctive target have, however, been found to show greater resistance to extinction when on half the training trials the feeding was immediately preceded by several minutes ofexposure to the target without food (Couvillon & Bitterman, 1980).…”

Children’s resistance to extinction: Two tests of the discrimination hypothesis

“…One process undoubtedly important is the ease with which the child can discriminate between acquisition and extinction conditions. According to the discrimination hypothesis (Gonzalez & Bitterman, 1964), the rate of extinction is inversely related to the degree of similarity between acquisition and extinction conditions. Thus, the PREE occurs because it is harder for an organism trained with partial reinforcement to discriminate between acquisition and extinction than it is for an organism trained with continuous reinforcement.…”