Doctoral Dissertations

Document Type

Dissertation

College

College of Arts and Sciences

Dissertation Defense Date

20-7-2024

Advisor(s)

Dickson, Chata

Committee Member(s)

Dickson, Chata; Thompson, Rachel; Pinkston, Jonathan; Ahearn, William

Abstract

Previous research has shown that Multiple Exemplar Training (MET) produces greater generalization than Single Exemplar Training (SET). The current study directly compares the effects of SET and MET using a parallel-treatments design. In the MET condition, we used a simplified approach to general case instruction called an instructional matrix. The instructional matrix borrowed elements from matrix training, such as organizing the relevant stimuli into a table to determine combinations for teaching and subsequent probing for generalization; and, it borrowed from general case instruction, the programming of common stimuli. The training and generalization settings were set up in the same way as multiple exemplar components are arranged in matrix training. In each axis of the matrix, there were variants of a single dimension of the context (e.g., a location or instruction). For each of 6 participants different functional skills were assigned to each of the two methods (SET and MET). Both methods, SET and MET, were effective in teaching the skills, but generalization results were superior using MET in 8 out of the 9 comparisons. Stakeholders were satisfied with outcomes, and they approved the procedures used.

Share

COinS