Association for Behavior Analysis International

Token economies are among the most commonly used and successful applied technologies (e.g., Soares, et al., 2016). Tokens offer advantages over providing the directly consumable reinforcer, as they are delivered without interrupting desirable behavior. Further, tokens may allow for longer durations of reinforcer access, which can enhance the value of the directly consumable reinforcer. However, a number of unanswered questions remain about tokens economies (Hackenberg, 2018). The current symposium addresses some of the limitations of the prior literature on token economies through a large-scale survey of current practices to establish and use tokens (Study 1), as well as directly assessing the value of tokens (reinforcing efficacy) across procedures that establish tokens as conditioned reinforcers (Study 2) and assessing the value of tokens (preference) exposed to contingencies (gain, loss, free) related to their practical use (Study 3). Findings from these studies are discussed with regards to practical use of token economies.

Undergraduates, practitioners, researchers

Tokens are among the most common consequences delivered by behavior analysts who work with individuals with developmental disabilities (Graff & Karsten, 2012). However, recommendations for establishing tokens as conditioned reinforcers vary and many questions remain about best practices. In this study, children with intellectual and developmental disabilities completed preference and reinforcer assessments, from which we identified two to three backup reinforcers. We then evaluated four procedures for establishing tokens as conditioned reinforcers, usually followed by extinction tests to determine if the token had assumed any independent value. We began with stimulus-stimulus (SS) pairing of tokens with the backup reinforcers. If SS pairing did not establish tokens as conditioned reinforcers, we evaluated response-stimulus (RS) pairing and/or noncontingent token-exchange training, in which participants exchanged noncontingently delivered tokens for backup reinforcers. If neither of these procedures established tokens as conditioned reinforcers, we assessed response-contingent token-exchange training. Results suggest that (1) exchange plays a critical role in supporting reinforcer effectiveness, and (2) the conditions under which we evaluate the effects of token training might influence our results and conclusions.

Conditioned reinforcers (e.g., tokens) assume the reinforcing value of the primary reinforcer with which they were paired. Behavioral economic concepts and some applied evidence suggest that the use of conditioned reinforcers with different reinforcement and punishment contingencies could alter their perceived value (adding value or devaluing that token). In particular, working to gain tokens or losing tokens may increase their value; whereas, getting tokens for free may decrease their value. The current study examined one participant’s preference for two different tokens immediately following the use of these tokens in the context of reinforcement and punishment contingencies (contingent reinforcement, noncongtingent reinforcement, and loss). Results of the study are consistent with behavioral economic research and suggest that contingent reinforcement and loss contingencies may have value adding effects; whereas, noncontingent reinforcement contingencies may have devaluing effects. Results are discussed within the context of a behavioral economic framework and practical applied advice is provided.