Response to Ahrendt, Houlihan, and Buchanan

Abstract

This paper reports some puzzling results from a token economy in an inpatient behavioral treatment facility. A seemingly insignificant change from poker chip tokens to sticker tokens produced substantial increases in problem behavior-as measured by frequency of time-outs for problem behavior. The results are puzzling because it is generally assumed that qualitative aspects of the tokens-such as whether or not they are handled-should not matter.What should matter is simply the correlation between tokens and other reinforcers for which they are exchanged. Indeed, the successful use of check marks, stickers, and stars in scores of token systems over the years shows pretty clearly that stickers and other nonhandled tokens are up to the job.Why, then, did behavior deteriorate in this case? Without additional information on other aspects of the program, it is hard to know for sure. Laboratory research may provide a few clues, however. To begin with, the findings seem relevant to the literature on incentive contrast-behavioral changes due to shifts in reinforcer value. Dozens of studies with a range of species have shown that downward shifts in reinforcer value induce predictable changes in behavior, including aggression and other behavior incidental to the reinforcement contingencies (Flaherty, 1982;Pellegrini et al., 2008). What is notable about these findings is that a reinforcer capable of sustaining behavior in its own right loses its value when contrasted with a higher-valued reinforcer. In the present case, the value of the stickers may have been degraded by contrast with the higher-valued poker chip tokens.This raises another question, namely, why should poker chips work better than sticker tokens as conditioned reinforcers? Looking again to laboratory research, classic experiments by Kelleher (1958) andMalagodi (1967), with chimpanzees and rats, respectively, have shown that animals frequently engage in various forms of token-directed behavior (e.g., handling and manual contact), arising from the frequent token-food pairings. While such handling does not seem necessary for tokens to function as effective reinforcers-as other research with pigeons has shown functional control by visual nonmanipulable tokens-it is certainly possible that manual contact enhances the reinforcing efficacy of tokens. Such handling may also heighten the discriminative functions of the tokens, clearly delineating the token-earning responses (an effect sometimes called marking), as well as bridging the gap between tokens and exchange periods. Based on these findings, it is not unreasonable to think that handling the poker chip tokens, if even briefly upon earning and then again upon exchange, enhances their reinforcing efficacy.To properly evaluate this hypothesis, it would be helpful to know more about the token-production and exchange schedules, both of which have been shown to be critical determinants of behavior in token economies. In a classroom-wide token economy with fourth and fifth grade students, for example, McLaughlin and M...