Behavioral Clinical Psychology (BCP) - Ventura County Office of Education (VCOE) Partnership

Project Summary: CSUN and VCOE will engage in a partnership lo evaluate the effects of extinction and error correction on biased responding during match-to-sample training among students with intellectual disabilities. Background Several behavioral assessments (e.g., preference assessments, discrimination training) for individuals with disabilities involve the client's selection of a stimulus from an array (e.g., Fisher et al., 1992; Saunders & Spradlin, 1989) and are used to inform behavioral interventions or to evaluate academic skills. However, outcomes may be hindered by selections that come under the control of an irrelevant stimulus feature,such as stimulus position (Bourret, Iwata, Harper ,& North, 2012). That is, if the client always selects the stimulus to the right of a rotating preference array, regardless of stimulus quality, a preference hierarchy cannot be determined . Likewise,if a client always selects the right-most stimulus during a conditional discrimination procedure,the appropriate discrimination will not be acquired. Although this pattern of responding­ referred to as a position bias- has been noted to occur in both animals and humans, to date only three studies have evaluated relevant treatments (Bourret et al., 2012; Galloway , 1967; Kangas & Branch, 2008). Although previous research has demonstrated effective treatments for biased responding, therapeutic effects may have been delayed in previous studies by the presence of intermittent reinforcement for biased responding. That is, biased responding sometimes resulted in selection of the correctly matched stimulus in the match-to-sample (MTS) procedures used by Galloway (1967) and Kangas et al. (2008), and biased responding intermittently resulted in HP or large magnitude reinforcers in the preference assessment procedures reported by Bourret et al. (2012). Extinction of biased responding (i.e., elimination of reinforcement for any biased response) may result in faster elimination of biases; however, no systematic investigation of the effects of extinction on biases exists to date. Further, no study has attempted to identify the extent to which the treatment of bias generalizes lo novel situations. Aim # I: To evaluate the effects of extinction and error correction on biased responding during match-to-sample training in individuals with intellectual disabilities. Aim #/2: To extend current research by examining the extent to which treatment effects generalize to responding in a non-treated context. Method During MTS training, two sample stimuli each will contain a different feature (color, shape, symbol,form ,etc.) and two comparison stimuli each will be identical to one sample stimulus (identity matching) or different than both sample stimuli (non-identity matching). For non-identity matching, each sample stimulus will be arbitrarily "matched" lo one comparison stimulus by the experimenter prior to the onset of the study . Two MTS tasks will be used in the study -an experimental task and a generalization task. In the experimental task, sample and comparison stimuli will be images displayed on a touch-screen monitor. In the generalization task, sample and comparison stimuli will be 3-dimensional objects; samples will be small trinkets and comparisons will be containers (e.g., cups) in which those trinkets can be placed, or sorted. Correct responding will result in access to a high-preference (HP) edible item (Fisher et al., 1992) and incorrect responding will be followed by a verbal indication that the selection was incorrect ,unless otherwise specified by the condition (below). The baseline condition will involve the presentation of sample stimuli rotated evenly across all positions in the array. Extinction will involve presenting the "correct" stimulus in the non-biased position on every MTS trial. Error correction will involve presenting repetitions of the same trial contingent on each "incorrect" response while rotating "correct" stimuli evenly across positions. Generalization of treatment effects will be examined by testing responding on a novel MTS task. Participants will include students (ages 3-12) enrolled in a special education classroom in one of two Ventura County Office of Education (VCOE) schools with whom the lab has established a contract. CSUN students will be trained to record data with iTouches using ABC DataPro software. The experiment will be conducted using a non-concurrent multiple baseline design (Watson & Workman , 1981). Anticipated Results & Implications Selected participants will demonstrate >60% biased responding during baseline. The addition of an extinction component may be sufficient to eliminate bias; however, the extinction condition also may inadvertently reverse bias. The addition of error correction should eliminate bias in any subject for whom extinction alone is ineffective. It is unclear whether the effects of extinction or error correction will result in generalization of treatment effects to an untreated context, or whether treatment in the novel context will be necessary. Results of this study can help identify the most efficient and effective method for eliminating position -biased responding in individuals with intellectual disabilities. This outcome has clinical implications for the programming of effective behavioral intervention and for the training of conditional discrimination, a requisite skill for many academic programs.