|
| Emergent Responding: Recent Advances and Future Directions |
| Monday, May 25, 2020 |
| 11:00 AM–12:50 PM |
| Walter E. Washington Convention Center, Level 2, Room 202A |
| Area: AUT/VBC; Domain: Applied Research |
| Chair: Kathleen Emily Marano (Caldwell University) |
| Discussant: Caio F. Miguel (California State University, Sacramento) |
| CE Instructor: Caio F. Miguel, M.A. |
| Abstract: The current symposium provides a discussion of research studies aimed at producing generative responding, including evaluations of multiple exemplar training, instructive feedback, and matrix training. The first paper will present a study that compared the efficacy and efficiency of serial multiple exemplar training, concurrent multiple exemplar training, and instructive feedback for producing generalization of tacts of various stimuli types with individuals with autism spectrum disorder. The second paper will present a study that assessed recombinative generalization with novel combinations of abstract stimuli by programming specific training histories for undergraduate students during matrix training. The third paper will present a study that taught who, what, and where intraverbal-tacts using matrix training, and evaluated the efficacy of matrix training across two matrices. The fourth paper will present a study that evaluated the efficacy and efficiency of incorporating instructive feedback within matrix training to teach children with autism spectrum disorder to label objects and adjectives. |
| Instruction Level: Intermediate |
| Keyword(s): generative responding, instructive feedback, matrix training, multiple exemplars |
| Target Audience: The target audience is professionals and researchers in behavior analysis. |
| Learning Objectives: Attendees will be able to describe a variety of methods for producing generative responding. Attendees will be able to describe how to use matrix training to produce emergent responding. Attendees will be able to describe how to use multiple exemplar training to produce emergent responding. Attendees will be able to describe how to use instructive feedback to produce emergent responding. |
| |
| Teaching Who, What, and Where Using Matrix Training |
| MARIA CLARA CORDEIRO (Marquette University), Mary Halbur (Marquette University), Tiffany Kodak (Marquette University), Gabriella Rachal Van Den Elzen (University of Nebraska Medical Center Munroe-Meyer Institute), Jessi Reidy (Marquette University) |
| Abstract: Teaching alternating Wh- questions consists of training conditional discriminations
under multiple sources of control which may lead to response errors (Sundberg &
Sundberg, 2011). The purpose of the current study was to teach who, what, and where intraverbal-tacts in the presence of 2D stimuli. Additionally, we wanted to determine the efficacy of matrix training across two 5x5x5 matrices. Matrix training consists of teaching two or more responses in the presence of a single stimulus comprised of multiple stimulus components (Pauwels, Ahearn, & Cohen, 2015). Two children diagnosed with autism spectrum disorder (ASD) first learned to tact individual components (e.g., “towel” in the presence of a towel on a white background). We then implemented non-overlap training of diagonal 1 from Matrix 1. After training, participant one demonstrated stimulus generalization in Matrix 1 and stimulus generalization to novel stimuli in Matrix 2. The second participant is still in data collection. Results suggest that training who, what, and where in the presence of compound stimuli from one diagonal in one matrix (i.e., 15 intraverbal-tacts) may to lead to intraverbal-tacts across novel stimuli combinations (i.e., recombinative generalization) and in the presence of entirely novel stimuli (i.e., response generalization). |
| |
| Matrix Training With and Without Instructive Feedback |
| ALEXANDRA MARIE CAMPANARO (Caldwell University), Bryan Rickoski (Caldwell University), Jason C. Vladescu (Caldwell University), Sharon A. Reeve (Caldwell University), Danielle L. Gureghian (Garden Academy) |
| Abstract: The current study examined the efficacy and efficiency of incorporating the use of instructive feedback within matrix training to teach children with autism spectrum disorder (ASD) to label common objects and adjectives. The study was conducted in a private school providing educational services to students with ASD based on the principles of applied behavior analysis. We taught one set of responses using a non-overlapping matrix, a second set of responses using an overlapping matrix, and a third set of responses using a non-overlapping matrix along with secondary targets to three individuals with ASD. The results demonstrated that all teaching
methods were effective and all trained and untrained responses were acquired. Additionally, results will be discussed across different measures of efficiency, including training sessions and training time to mastery. Our findings will be discussed in light of the extant matrix training and instructive feedback literatures. Additionally, we will provide directions for future research. |
| |
| The Effects of Varying Matrix Training Arrangements on Recombinative Generalization |
| REBECCA DURHAM (University of North Texas), Samantha Bergmann (University of North Texas ), Karen A. Toussaint (University of North Texas) |
| Abstract: Recombinative generalization is a stimulus control process that involves responding to novel stimulus combinations, and it can be facilitated through an instructional approach, matrix training. A learner’s history with constituent stimuli and the arrangement of combination stimuli within the instructional matrix may affect
the likelihood of recombinative generalization. To investigate this further, the current project assessed recombinative generalization with novel combinations of abstract stimuli by programming specific training histories for undergraduate student participants. The matrix training conditions were: (a) overlap with known
(i.e., previously acquired) constituents, (b) overlap with unknown (i.e., not previously acquired) constituents, (c) nonoverlap with known constituents, and (d) nonoverlap with unknown constituents. We evaluated whether and the extent to which recombinative generalization occurred in each matrix training condition in comparison to a condition that included teaching the constituents and providing a word-order rule. Finally, we compared the total training trials to a condition in which we directly trained all constituents and combinations. The results
suggested both overlap conditions and the nonoverlap with known constituents condition produced recombinative generalization, and the nonoverlap with known constituents condition was the most efficient. These results could inform the training order and stimulus arrangements practitioners employ to program for
recombinative generalization. |
| |
Comparing the Efficacy and Efficiency of Tact Training Procedures for Generalization With Children With Autism Spectrum Disorder |
| GABRIELLA RACHAL VAN DEN ELZEN (University of Nebraska Medical Center's Munroe-Meyer Institute ), Regina A. Carroll (University of Nebraska Medical Center Munroe-Meyer Institute) |
| Abstract: Programming for generalization is a critical component of applied behavior analysis. Previous research has evaluated several procedures for achieving stimulus generalization in the context of tact training with children with autism spectrum disorder, including serial multiple exemplar training (S-MET), concurrent multiple exemplar training (C-MET), and instructive feedback (IF). Although previous research has compared some or all of these procedures, results have been mixed. In the present study, we used an adapted alternating treatments design to directly compare the efficacy and efficiency of S-MET, C-MET, and IF for producing generalization of tacts of various types of stimuli (e.g., color photographs, black and white outlines, colored drawings) with four males with autism spectrum disorder. For most participants, C-MET led to generalization in the fewest training sessions, followed by IF. These results suggest that S-MET is unlikely to lead to generalization more efficiently than other conditions, but that the ideal training arrangement may be idiosyncratic. |
|
| |