|Optimizing Assessment and Treatment through Methodological and Translational Research
|Sunday, May 25, 2014
|9:00 AM–10:50 AM
|W186 (McCormick Place Convention Center)
|Area: DDA; Domain: Applied Research
|Chair: Tara A. Fahmie (California State University, Northridge)
|Discussant: Per Holth (Oslo and Akershus University College of Applied Sciences
|CE Instructor: Tara A. Fahmie, Ph.D.
This symposium features methodological and translational research evaluating behavioral processes involved in preference and reinforcer assessment, conditioned reinforcement, and differential reinforcement. Lisa Hunter will present a study comparing the effects of stimulus-stimulus pairing and discriminative control in the establishment of conditioned reinforcers. Janine Urbano will present a study evaluating a new approach to the analysis of preference hierarchies obtained through pairwise preference assessment. The traditional percentage method was compared to the Thurstone comparative law to test whether the latter analytical strategy may provide better predictions of reinforcing effects. Lorraine Becerra will present a review and analysis of the relation between assessment consistency and validity of multiple stimulus without replacement preference assessments. Finally, Michael Kelley will present a translational study allowing for a close evaluation of the reinforcement processes underlying behavior change during differential reinforcement of alternative behavior. Dr. Per Holth, with a background in both experimental and applied behavior analysis, will close the symposium with remarks on the contributions of this research.
|Keyword(s): conditioned reinforcement, differential reinforcement, preference assessment, translational research
Pairing vs. Discriminative Training for Establishing Conditioned Reinforcement Effects
|LISA HUNTER (St. Amant Research Centre), Alison Cox (University of Manitoba), Gabriel Schnerch (University of Manitoba), Javier Virues Ortega (University of Manitoba, St. Amant Research Centre, University of Auckland)
Establishing new reinforcers is an endeavor of paramount importance for the implementation of reinforcement-based approaches to treatment among individuals with developmental and intellectual disabilities. Identifying effective reinforcers for low functioning clients may be particularly challenging. Two methods have been proposed to induce reinforcing effects to initially neutral items: stimulus-stimulus pairing and discriminative training. First, stimulus-stimulus pairing (SSP) consists of the concurrent presentation of a neutral item with an already established reinforcer. Second, the discriminative control procedure (DCP) features a neutral item as a discriminative stimulus signaling the availability of an already established reinforcer contingent upon an arbitrary response. The goal of the present study was to evaluate which of these methods induces greater conditioned reinforcement effects among individuals with intellectual disabilities. We conducted a series of preference assessments to identify established reinforcers, neutral leisure items, and arbitrary responses with no (automatic) reinforcing effects. We evaluated the effects of the SSP and DCP methods in a multi-element manipulation combined with a multiple baseline design across subjects. The results showed that for most participants both interventions induced some conditioned reinforcing effects. While participants engaged more often in the arbitrary response during contingent reinforcement probes following training with either method, responding was highly variable. Moreover, a clear superiority of one approach over the other was not demonstrated in any of the participants.
Reinforcing Effects of Items Ranked According to the Thurstone Comparative Law
|JANINE URBANO (University of Manitoba), Flavia Julio (University of Manitoba), Javier Virues Ortega (University of Manitoba, St. Amant Research Centre, University of Auckland)
Preference may be defined as the relative strength of behaviors among two or more choice options and it is often measured as a pattern of choosing. Assessing the preferences of persons with intellectual and developmental disabilities (ID/DD) is important for several reasons. Preferred items often function as reinforcers and they can be used in intervention programs for establishing new skills and reducing problem behaviors for people with ID/DD. Pairwise preference assessment is often used to evaluate potential reinforcers in this population. The outcome of a typical pairwise preference assessment is a hierarchy of items ranked according to the percentage of trials in which each item was chosen out of the times the item was presented. This hierarchy is an ordinal scale that hardly accounts for variability of choice over time. By contrast, the Thurstone paired comparative method generates interval-level scales over multiple assessments. Therefore, the latter analytical strategy may account better for time-dependent changes in preference. The purpose of the present study was to determine whether the Thurstone approach to data analysis would more accurately predict reinforcing effects, thereby enhancing the validity of pairwise preference assessments. We conducted a series of pairwise preference assessments analyzed through the traditional percentage method and the Thurstone method. The reinforcing effects of items with diverging ranks according to either method were subsequently evaluated in a concurrent schedule reinforcer assessment embedded in an ABAB design. Overall, the results indicated that scale values resulting from the Thurstone analysis provided better predictions of reinforcing effects.
|A Review and Analysis of the Consistency of MSWO Assessments
|LORRAINE BECERRA (California State University, Northridge), Tara A. Fahmie (California State University, Northridge)
|Abstract: The consistency of stimulus rankings across repeated multiple stimulus without replacement (MSWO; DeLeon & Iwata,1996) preference assessments may influence the predictive validity of its outcomes. For instance, inconsistent stimulus rankings might be a function of behavioral biases (e.g. side biases), rule governed behavior (e.g., "save the best for last"), or changes in preference over time. However, MSWO consistency rarely has been reported in published research. We first reviewed the consistency and validity of published MSWO data. Next, we conducted an analysis of MSWO data from 11 individuals diagnosed with an intellectual disability between the ages of 5 and 22 years old, each of whom participated in five assessments of three different arrays containing eight stimuli each. Spearman rank correlation coefficients across assessments were moderate to weak (range, rs = 0.04 to 0.96) for more than half of the participants. Methodological and practical implications of these data, as well as potential areas for future research, will be discussed.
An Animal Model of Differential Reinforcement of Alternative Behavior
|MICHAEL E. KELLEY (The Scott Center for Autism Treatment, Florida Institute of Technology), Christopher A. Podlesnik (The University of Auckland)
Translational research often consists of replicating and extending the results of basic findings. Replications might consist of obtaining evidence of generality (e.g., across species) or application of basic findings to socially important behaviors (e.g., enhancing treatment of socially important problems). In this collaboration, we extended previous translational research by exposing non-human animals to an experimental preparation more consistent with typical application with humans--differential reinforcement of alternative behavior. This approach is in contrast with typical preparations in which humans are exposed to more typical non-human, basic arrangements. Preliminary findings reveal consistent resurgence of our analogue of problem behavior upon discontinuing reinforcement for alternative behavior. These findings provide a platform to assess thoroughly and efficiently factors influencing long-term treatment maintenance of behavioral treatments. For example, we can assess the extent to which multiple contingency reversals, which are common in applied differential reinforcement arrangements to establish experimental control, might influence the occurrence and magnitude of resurgence. This collaboration offers the opportunity to understand the behavioral processes underlying behavior during treatment while developing avenues to improve treatment effectiveness.