Association for Behavior Analysis International

Individuals with autism often engage in rigid or stereotyped patterns of behavior. A multiple baseline across participants design was used to examine the effects of an across session (second-order) lag schedule on the choice of leisure activities of three elementary students diagnosed with autism. During baseline, two participants showed low levels of response variability when choosing a leisure activity and one participant displayed a pattern of frequent within session switching. Variable responding across sessions increased for all participants following the implementation of the across session lag schedule of reinforcement. During a non-treatment follow-up, one participant continued to respond according to the Lag 2 schedule and one participant engaged in reparative responding. This study supports the use of lag schedules of reinforcement to increase response variability in students with autism. Additionally, it addresses the limitation of frequent within-session switching by utilizing an across session schedule.

Prior research has demonstrated that when positive reinforcement is presented contingent upon response variability, organisms will increase the degree to which they vary their behavior (e.g., Page & Neuringer, 1985). However, relatively little research has been conducted on the relationship between negative reinforcement and variability. Although positive and negative reinforcement operate similarly (Michael, 1975; Perone, 2003), some have suggested that negative reinforcement may restrict an organisms behavioral repertoire (Sidman, 2001). We evaluated the effects of a schedule in which negative reinforcement was contingent upon response variability. Several college students enrolled in an introductory course in special education served as participants. They played a computer game in which a high-pitched tone was played through earphones, and clicking on one of nine buttons on the screen could terminate the tone. In successive conditions, termination of the tone depended on the current response differing from the previous 3 responses (lag 3) or repeating one of the previous 3 responses (rep 3). Distinctive background colors were associated with the lag 3 and rep schedules. A subsequent series of manipulations assessed the extent to which these background colors exerted stimulus control over responding. All participants responded more variably under the lag 3 condition. Half of the participants also demonstrated stimulus control.

Basic behavioral literature suggests that behavioral variability is a dimension of behavior that can be increased with operant reinforcement and controlled by discriminative stimuli (Page and Neuringer, 1985). In the current study, a multiple schedule was first used to establish stimulus control over repeated and variable responding in the presence of two different colored cards in 2 college students. Reinforcement was available on an FR schedule of reinforcement during the REPEAT component and on a Lag schedule of reinforcement during the VARY component. Once discriminated responding was observed, a novel task was introduced and repeated and variable responses during the novel task were measured during an extinction component in the presence of the previously acquired same discriminative stimuli. The presence of the discriminative stimuli evoked variable responding both participants, even when reinforcement was removed for all components.