Association for Behavior Analysis International

Although punishing consequences have been shown to reduce harmful or unwanted behavior when other methods fail, the use of punishment in behavioral interventions has largely been abandoned due to ethical concerns. There is a need for new approaches to punishment that do not produce unwanted side effects. Such methods must be assessed in terms of ethical viability and overall punishment efficacy. That is, the extent to which the punishment approach successfully reduces the future probability of targeted behavior. This symposium presents new approaches to the investigation of punishment and the development of punishment procedures. Presentations will show that a functionally negative stimulus may alone act as a punishing consequence with animals, and considers the extent to which this method may be used to punish sub-optimal choice behavior in animals and humans. Presentations also consider the use of delay to reinforcement as a punishing consequence, and the extent to which environmental variables may moderate the efficacy of punishment of positively reinforced behavior, thereby modelling a component process of some psychopathological phenomena.

The use of punishment in treatments designed to reduce harmful behavior has largely been abandoned for ethical reasons. However, the use of positive reinforcement in behavioural treatments may fail to reduce harmful behavior to safe levels. Use of a functionally "negative" stimulus as an operant punishing consequence offers a new approach to punishment. The present study uses an animal model to investigate whether presentation of a negative stimulus will punish the behavior it follows. Six pigeons are used. The target behavior is key pecks for positive reinforcement. One stimulus (S+) predicts response-contingent food deliveries on a variable interval schedule. Simultaneously, a negative stimulus previously associated with the absence of food (S-) is presented on a variable rate schedule. Food deliveries are not withheld when the S- stimulus is presented. Results show that the overall rate of key pecking for food by pigeons is suppressed when key-pecking also produces the negative stimulus. Therefore, a negative stimulus alone has the potential to be a punishing consequence. Ongoing research investigates the extent to which different environmental variables impact the functional reliability of a negative stimulus as a punishing consequence, and the extent to which a negative stimulus may punish choice. This research provides a foundation for the continued investigation of new approaches to punishment that reduce unwanted behavior without raising ethical concerns.

Behavioral principles have significant relevance to the prevention and minimisation of harm in a number of health-related areas. They are, however, under-utilised. Recent research in New Zealand shows that compulsory interruptions of gambling have marginal benefits. Basic research using animal models suggests that, given appropriate pre-training, punishment of sub-optimal choice strategies may be demonstrated using an ethically sound approach, and that these effects might be mirrored in humans. Our ongoing research investigates the extent to which this approach may be integrated into a simulation of an Electronic Gaming Machine (EGM, poker/pokie machine, slot machine). In most jurisdictions, the majority of gambling-related harm is attributable to gambling excessively on EGMs which are underpinned by sophisticated variations of variable ratio schedules. The present research investigates how, as part of the ongoing gaming experience, specific game symbols may acquire aversive properties, be embedded within the game, and made contingent on excessive or harmful gambling. The implications will be discussed in terms of harm minimization in gambling and related contexts with humans.

When responding delays the delivery of a reinforcer and the reinforcer is obtained by not responding in a specific way, such procedure (named resetting delay / omission training) is an effective method to reduce specific target behaviors. The efficacy of such procedure depends on the duration of the delay, on delays being signaled or not, on application of delays from the outset of training or after training is well established, and on the specific contingency between behavior and consequence, among other factors. Variations in the response-dependent resetting delay procedure involve the use of non-resetting delays (delays being not reset during the delays) or the use of protective response-reinforcer delays. These other methods have advantages in order to investigate theoretical issues, but in general response-contingent delays result in a decrement in response rate that represent effective alternative techniques to suppress behavior in comparison to traditional punishment. This will be illustrated here by their use on modulating schedule-induced drinking in rats, a laboratory model that has been regarded useful for understanding behavioral excess.

Reinforced acts are often subjected to punishment, as when the incautious consumption of prey is terminated in its theft by a rival, or when an enjoyable conversation is cut short by an unconsidered comment. The termination of threats of such punishment generally requires a temporary cessation of ongoing reinforced acts, a cessation which may become less probable as the terminal reinforcer for ongoing behavior approaches. In a test of this notion, pigeons' pecks to a green key were reinforced on ratio schedules; after a random number of pecks during each ratio run, a second key was concurrently and briefly lit red, accompanied by a tone. A single keypeck to red terminated the warning signal and averted a forthcoming blackout period and a zeroing of the green keypeck counter. The probability of warning signal termination was observed to be an inverse function of the number of responses remaining for terminal reinforcement, despite a greater loss of effort when punishment occurred later in a ratio run. The preemptive self-exposure to warnings observed in many psychological disorders, introduced early to maximum effect in evoking a course of behavior incompatible with the continuation of enticing censured acts, may constitute an instance of this process.