The effect of conditioned reinforcement rate on choice: a review

Abstract

We review the nature of conditioned reinforcement, including evidence that conditioned reinforcers maintain choice behavior in concurrent schedules and that they elevate responding in the terminal links of concurrent-chains schedules. A question has resurfaced recently: Do theories of choice in concurrent-chains schedules need to include a term reflecting greater preference for higher rates of conditioned reinforcement? The review of several studies addressing this point suggests that such a term is inappropriate. Elevated rates of conditioned reinforcement (and responding) in the terminal links of concurrent-chains schedules do not lead to greater preference in the initial link leading to the higher rate of conditioned reinforcement. If anything, the opposite preference is likely to occur. This result is not surprising, since the additional putative conditioned reinforcers in the terminal link are not correlated with a reduction in time to primary reinforcement nor with an increase in value.

Fig 1. Mean relative response rates during the terminal link of the concurrent-chains schedule with the superimposed paired stimulus schedule (top), and median relative choices for this schedule (bottom), when the superimposed stimulus was paired with primary reinforcement.

Means are averages of the medians. Data from Conditioned Reinforcement (p. 213, Figure 8.7), by D. P. Hendry, 1969, Homewood, IL: Dorsey Press. Copyright 1969 by Dorsey Press. Adapted with permission.

Fig 2. Mean relative response rate on the key associated with the paired schedule of brief-stimulus presentation.

The open bars represent the initial links and the closed bars represent the terminal links. Two asterisks above a bar indicate that the result was significant at the 0.01 level; one asterisk indicates a significant result at the 0.05 level (i.e., relative responding significantly different from .50). Data from “Preference for conjoint schedules of primary reinforcement and brief-stimulus presentation” (p. 37, Figure 4) by N. Squires, unpublished doctoral dissertation, 1972, University of California, San Diego. Copyright 1972 by the University of California, San Diego. Adapted with permission.

Fig 3. The concurrent-tandem and concurrent-chains procedures used in Fantino et al. (1991).

The left panel illustrates the sequence of events in the concurrent-tandem schedule, in which the same stimulus is present on each key throughout. The right panel shows the sequence of events in the concurrent-chains schedule, in which entry into each terminal link is signaled by a unique stimulus. During both the choice and outcome phases of one schedule, the other schedule remained operative in both the chain and tandem procedures. From “Choice and conditioned reinforcement,” by E. Fantino, D. Freed, R. A. Preston, & W. A. Williams, 1991, Journal of the Experimental Analysis of Behavior, 55, p. 181. Copyright 1991 by Journal of the Experimental Analysis of Behavior. Adapted with permission.