Association for Behavior Analysis International

Establishing conditioned reinforcers is often a critical step in service delivery for individuals with autism spectrum disorder (ASD) and/or intellectual and developmental disabilities (IDDs). Therefore, identifying effective procedures for establishing conditioned reinforcers is critical. In this review, we systematically searched the literature for published and unpublished single-case design studies that evaluated stimulus-stimulus (S-S) pairing, response-stimulus (R-S) pairing, and/or operant discrimination training (ODT) for establishing conditioned reinforcers for individuals with ASD and/or IDDs. We searched two electronic databases for studies that included pre- and post-pairing measurements of responding that resulted in access to the stimulus targeted for conditioning. Thirty-one studies, including twelve theses and dissertations, met inclusion criteria. Eight studies evaluated S-S pairing, 20 evaluated R-S pairing, and nine evaluated ODT. Combined, S-S pairing, R-S pairing and ODT were effective at establishing conditioned reinforcers only about half the time. However, analyses of the effectiveness of each procedure indicate that R-S pairing is the most effective of the three procedures.

We recently surveyed certified clinicians about their commonly used practices when training and implementing token economies with individuals with autism and other neurodevelopmental disorders. Overwhelmingly, clinicians reported selecting stimuli to be used as tokens based on the learner’s existing interests (e.g., characters, animals, etc.). Additionally, clinicians reported that learners are allowed to manipulate tokens during both token production and exchange. Tokens based on a learner’s interest (i.e., interest-based tokens) have been demonstrated to produce greater increases in the levels of on-task behavior compared to novel tokens (Carnett et al., 2014) Interest-based tokens have also been shown to increase correct responding, relative to pre-existing tokens, within the context of skill acquisition (Charlop-Christy & Haymes, 1998). However, several limitations in both Carnett et al. (2014) and Charlop-Christy and Haymes (1998) limit the conclusions that can be drawn to inform clinical practice. The present study evaluated that how novel and interest-based tokens affect skill acquisition with three children diagnosed with autism spectrum disorder. Results suggest that the type of stimuli used as a token does not have a significant effect on skill acquisition but can produce longer handling times which can result in fewer learning opportunities across time. Suggestions for clinical practice and future research will be discussed.

We examined the effects of token production delays on the rate of skill acquisition during discrete trial instruction (DTI) for three children with an intellectual or developmental disability (IDD). We used a multielement design to compare the rate of skill acquisition in a DTI with accumulated reinforcers context across three conditions: a) immediate reinforcement; B) 5-s token-production delay; and c) 10-s s token-production delay. All participants acquired the skills more quickly in the immediate reinforcement condition. Delays of both 5 s and 10 s decreased the efficiency and efficacy of skill acquisition for all participants.

Basic research has shown that nonhuman animals generally display a preference for variable ratio (VR) rather than fixed ratio (FR) schedules of reinforcement, particularly when low individual ratios are included (Field et al., 1996). Minimal applied research has investigated preference for these schedules among children with autism spectrum disorder (ASD) despite a clinical phenotype that suggests a general preference for sameness (which may theoretically extend to a preference for fixed schedules). In a preliminary investigation of second-order schedule effects within a token economy, Argueta et al., (2019) found a similar preference for VR rather than FR schedules for their participant with ASD. This study extends Argueta et al. by evaluating preference for FR and VR exchange schedules across an escalating range of exchange ratios within a token economy. We used a concurrent chains assessment to evaluate preference for FR or VR exchange-production schedules of reinforcement at equal ratios of 5 and 10. Preliminary results did not indicate a strong preference for either schedule at a ratio of 5, however an increase to a ratio of 10 resulted in an emergence of preference.