Joint effects of pattern of reinforcement, intertrial interval, and amount of reinforcement in the rat.

Abstract

The joint effects of patterns of reinforcement (composed of different values of n length, number of N-R transitions, and total number of nonreinforcements), large and small amounts of reinforcement, and massed and spaced trials on acquisition and both initial and terminal extinction were observed. Greater total numbers of nonreinforcements produced greater resistance to extinction, particularly early in extinction. Massed trials produced greater resistance to extinction than spaced trials, but only later in extinction. Large amounts of reward resulted in more resistance to extinction than small amounts for partially reinforced animals, but the opposite for continuously reinforced animals. A three-way interaction between pattern, amount, and intertrial interval was not found. The main effects and interactions are best explained by a frustration mechanism.

Successive nonreinforcements (N-length) and resistance to extinction at spaced trials.

“…The results of this investigation, along with other recent data (Capaldi & Lynch, 1968;Seybert et al, 1973), lend great evidence to the operation of sequential effects at spaced trials. In addition, the results of this investigation would seem to place in doubt other data that until now seemed to support other theories (e.g., Gonzalez & Bitterman, 1969;Jernstedt, 1971), since these results would appear to be related at least in part by the type of runway used. At the present time, no other theory (e.g., frustration theory, Amsel, 1958Amsel, , 1962 seems adequate to account for the differential results of the present experiments.…”

“…Studies where the runway was all one color have supported a sequential analysis of spaced-trials data (e.g., Capaldi & Lynch, 1968;Capaldi & 6S2 Minkoff, 1969;Seybert et al, 1973). Studies where the goal box differed in brightness from the rest of the apparatus have failed to support sequential theory at spaced trials (e.g., Gonzalez & Bitterman, 1969;Jernstedt, 1971). In order to test the above hypothesis, an apparatus with two different goal boxes was used in the present spaced-trials experiments.…”

“…Although it is generally accepted that sequential mechanisms account for most of the extinction results obtained at massed trials (e.g., Robbins, 1971), the results obtained at spaced trials are somewhat inconsistent with this conclusion. Some experiments (e.g., Capaldi, 1971;Seybert, Mellgren, & Jobe, 1973) have supported the notion that the memory of reward and nonreward may be functional at intertrial intervals (ITIs) of at least 24 hr., but others have failed to support the notion (e.g., Gonzalez & Bitterman, 1969;Jernstedt, 1971).…”

Resistance to extinction at spaced trials using the within-subject procedure.

“…It has been suggested that the partial reinforcement effect (PRE) may be governed by one kind of mechanism when a short intertrial interval (ITI) is used, but a different kind of mechanism when a more extended ITI is used (Amsel, 1967;Gonzalez & Bitterman, 1969;Jernstedt, 1971). The present experiments were designed to test this suggestion using a within-5 procedure, where a partial reinforcement (PRF) schedule is associated with one alley, and a continuous reinforcement (CRF) schedule is associated with a second alley.…”

“…ITI, which is a relatively short ITI and therefore might be considered close to an optimal ITI for demonstration of the operation of sequential variables (Amsel, 1967). EXPERIMENT I Amsel (1967) and others (e.g., Gonzalez & Bitterman, 1969;Jernstedt, 1971) have 1 Requests for reprints should be sent to Roger L. Mellgren, Department of Psychology, University of Oklahoma, Norman, Oklahoma 73069. stated that Capaldi's (1967) sequential hypothesis is adequate for explaining PRE at short ITIs (massed trials), but is not adequate for explaining long-ITI (spacedtrials) data. In order to determine the importance of sequential variables at long ITIs, the present experiment used 24-hr.…”