Association for Behavior Analysis International

In this symposium, we will review the effects of reinforcement history on the persistence of behavior. The first presenter will share data on the response patterns of children with autism under lean and rich token schedules and yoked tandem schedules at varying response requirements. Our second presentation will explore the effects of aggregate reward history on overjustification and behavioral persistence, and to extend to which these may be a function of stimulus-reinforcer relations. Our final presenter will share data comparing the persistence strengthening effects of a DRA procedure in which reinforcement histories are developed concurrently vs. a DRA procedure in which reinforcement histories for two responses are developed separately.

Researchers interested in the pattern of responding and persistence of behavior under different reinforcement schedules and parameters.

Token economies are commonly used in educational and clinical settings as tools for reinforcing appropriate behavior as well as the absence of problem behavior. However, little applied research has been conducted to investigate the behavioral mechanisms through which token economies exert their effects. What’s more, research with non-human animals has shown that tokens may serve discriminative functions that may actually suppress responding under specific conditions. The purpose of this study was to evaluate the effects of various schedules of token delivery and exchange as well as yoked tandem schedules on response patterns in children with autism spectrum disorders. Four children ranging in age from 4 to 11 participated in the current study. We compared responding under lean and dense token delivery schedules with responding on tandem schedules at varying response requirements. Visual inspection of session to session graphs revealed variable responding within and across participants. However, aggregated response patterns revealed effects consistent with basic research on token and tandem schedules. Notably, token training was associated with increased response rates in most participants. Tokens were also found to suppress responding under some conditions (e.g., high response requirements). These findings provide further insight into the generality of basic research findings on token schedules to clinical settings.

The overjustification effect remains controversial across psychological perspectives. Cognitive researchers frequently find evidence of the phenomenon (Deci, Koestner, & Ryan, 2001), whereas behavioral researchers rarely observe the effect (Cameron & Pierce, 2002). Importantly, researchers have generally employed methods of investigation common to their own field. This has occasioned consistent differences across perspectives in the rate and aggregate history of reinforcement prior to tests of overjustification. This is notable considering behavioral momentum literature indicates rate of reinforcement has a central importance in governing response persistence during disruption (such as extinction, as applied in the overjustification effect). Aggregate reward history may have similar effects, to the extent that heightened stimulus-reward associations may obtain over longer durations of stimulus-reinforcer pairings. This study examined the effects of aggregate reward history on overjustification and behavioral persistence, and the extent to which overjustification effects, like persistence, may be a function of stimulus-reinforcer relations. Results indicated that in 3 of 4 cases, longer histories of reinforcement were associated with stronger persistence values and less frequent overjustification effects, suggesting aggregate reward history may strengthen responding as it relates to these phenomena.

Behavioral momentum theory research has demonstrated that DRA based interventions, while decreasing the rate of a target response (i.e., problem behavior), may in fact increase the persistence of the target response when reinforcement for the alternative response is disrupted. Previous studies have demonstrated that by training the alternative response in a context in which a target response has no history of reinforcement and then combining this context with one in which the target response has a history of reinforcement, the persistence increasing effects of DRA can be minimized when disruption occurs. The present study replicated previous findings with individuals with developmental disabilities by showing that combining stimuli associated with a rich FI schedule for an alternative response with a comparatively leaner FI schedule for a target response reduced target respond persistence compared with concurrent training of the alternative and target responses. Participants engaged in comparatively similar novel responses across three colored conditions, in which target and alternative responding was reinforced separately and concurrently. Extinction tests were used to test response persistence across the concurrent and combined stimulus conditions. The current study extends previous research by replicating this finding within participants following resumption of reinforcement phases and subsequent disruption phases.