Recombinative generalization and matrix training : methods for acquisition of untaught components

Document Type

Thesis

College

College of Arts and Sciences

Department

Psychology

Degree

M.S. Applied Behavior Analysis

Date Completed

2012

First Committee Member

Ahearn, William

Second Committee Member

Dickson, Chata

Third Committee Member

Karsten, Amanda

Abstract

"Instructional efficiency is an important factor in selecting teaching procedures for individuals with autism as many demonstrate slow rates of acquisition. Key factors in instructional efficiency are the rate of acquisition and generalization of responding. Foss (1968ab) investigated two methods for structuring instruction (“overlap” and “non-overlap”) to promote recombinative generalization using a matrix training procedure with undergraduate students. The present study used a similar procedure to teach tacting of prepositions and kitchen items to three girls, ages 13-20, diagnosed with an autism spectrum disorder. Kitchen items and prepositions were trained in combination (e.g., “the strainer is to the right of the box”) down the diagonal of the matrix according to a “Non-overlap” instructional sequence. Untaught combinations were probed then more combinations were taught in a stepwise pattern or “Overlap” sequence followed by probe sessions. Additional instructional phases, including a second “Non-overlap” condition, were also conducted followed by probe sessions. Probes tested expressive and receptive generalized responding. Across all sessions, expressive identification of combinations of kitchen items and prepositions was explicitly taught using faded verbal prompting and discrete trial training procedures. Results varied across participants in number of trials to mastery and from minimal to nearly complete generalization of all untaught combinations following the first “non-overlap” condition. Across all participants, generalized responding increased with further training sequences. Generalized identification of components occurred receptively across all three participants and expressively for two participants. Additionally, generalized responding occurred more frequently in probes of receptive responses than in probes of expressive responses for two participants."

This document is available upon request to Western New England University faculty, students, and staff. Please contact D'Amour Library at for access.

Share

COinS