Association for Behavior Analysis International

Matrix training approaches have been an area of clinical research for decades, yet many variables related to the effectiveness of the procedures have yet to be empirically studied. This symposium examines several of these variables with a wide range of participants. The first study examined the use of over-lapping or non-overlapping designs when teaching language skills to typically developing toddlers. The second study examined whether teaching pre-requisite component skills prior to combinated skills is necessary by comparing three different teaching arrangements. The third study examined the role of testing sequence in promoting recombinative generalization across matrices with children diagnosed with autism spectrum disorder (ASD). The fourth study examined applications of matrix training with language skills with 2-year olds diagnosed with ASD. Findings from these studies have implications for clinical programming and future directions for research in the area of matrix training. Results and common themes will be discussed by Dr. Linda LeBlanc.

Recombinative generalization involves the production of novel responses to untrained stimuli when the component skills are in an individuals behavioral repertoire. Matrix training, which has been shown to result in recombinative generalization, involves arranging targets so that some combinations are exposed to direct teaching while others may emerge without instruction (e.g., Axe & Sainato, 2010). The current study examined the efficacy and efficiency of three matrix-training procedures employing a non-overlapping approach and arranged in an adapted alternating treatments design to teach textual responses and noun-verb tacts with three participants. The component skills necessary to emit the combination skill were not in the participants repertoire prior to the study. The matrix-training procedures included: teaching the component skills with the combination skill, teaching only the combination skill, and teaching the component skills to mastery prior to teaching the combination skill. Thus far, results show that the procedure which includes teaching the component skills along with the combination skill has met the criterion for recombinative generalization in the fewest sessions and least instructional time. Procedural arrangements to promote efficient acquisition and recombinative generalization with matrix training will be discussed.

Research in matrix training has shown that when the diagonal targets are trained the individual may demonstrate correct responses to the non-diagonal targets within the same matrix and novel targets (Frampton, Wymer, Hansen, & Shillingsburg, 2016). The purpose of this study was to replicate prior research with an altered order of post-tests to isolate whether effects are due to multiple exemplar training (MET) or unique to matrix training. Three males and 1 female diagnosed with autism were exposed to matrix training with mastered tacts of nouns (e.g., cat) and verbs (e.g., jumping). Following baseline of a Generalization Matrix and Matrix 1, the diagonal targets within Matrix 1 were trained (e.g., The cat is jumping). Post-tests were conducted for the Generalization Matrix prior to post-tests for Matrix 1. Three participants showed immediate recombinative generalization within the Generalization Matrix, results similar to those obtained through MET. One participant showed recombinative generalization with the Generalization Matrix only after exposure to the Matrix 1 post-test. This result suggests that the opportunity to respond to the trained targets and their recombinations had a facilitative effect; a component not included in MET. These findings have bearing on language programming strategies and conceptualization of matrix training approaches.

Communication deficits are common in children diagnosed with Autism Spectrum Disorder (ASD). Using generative instruction for communication skills, such as matrix training, instructors teach a subset of skills and new skills emerge without direct teaching. Such an approach can result in a faster acquisition of functional language in children with communication delays. Matrix training has been used to teach preschool-aged children with ASD spelling, writing, receptive identification, and pretend play skills. This approach also has been effectively used to teach tacting and listener skills to children with intellectual disabilities. We employed matrix training with much younger participants:2-year-old boys, diagnosed with ASD with a limited communication repertoire. We taught noun-verb tacting and listener responding skills. During baseline, participants demonstrated limited correct noun-verb tacting. Correct responses for targets directly taught increased progressively across trials. Further, we observed spontaneous generalization to novel noun-verb tacts in the natural environment. These findings suggest matrix training can be a useful instructional strategy for expanding the communication repertoire of young children diagnosed with ASD.