|
| Science Board Translational Series: The Science and Practice of Discrete-Trial Training |
| Sunday, May 24, 2009 |
| 10:30 AM–11:50 AM |
| North 120 BC |
| Area: AUT/EAB; Domain: Applied Behavior Analysis |
| Chair: Kathryn Saunders (University of Kansas) |
| Discussant: Kathryn Saunders (University of Kansas) |
| Abstract: Discrete-trial training is widely used to teach early language skills to children with autism and other developmental disabilities. These procedures were developed within longstanding translational-research programs targeted at understanding basic principles of stimulus control. The three talks will underscore the scientific basis for instructional programming, and illustrate contemporary lines of research that will lead to refined teaching procedures. Kathryn Saunders will present a component analysis of arbitrary matching to sample, a procedure that often is used in receptive-language instruction. She will analyze common teaching procedures (e.g., delayed prompting), with the goal of illustrating why some procedures are more effective than others. Next, Carol Pilgrim will describe two promising lines of research from her laboratory. One procedure facilitates the acquisition of arbitrary matching by first ensuring accurate thematic matching. The other line of work, directly from the animal laboratory, involves differential-outcomes procedures. To wrap things up with a new way of conceptualizing the problem of seemingly identical procedures that vary in their effectiveness, William McIlvane will present the concept of “joint abstraction.” He will describe two laboratory research programs that apply this concept to studies of stimulus control in individuals with severe intellectual disabilities and in nonhuman primates. |
| |
| Why Some Procedures for Teaching Arbitrary-Matching-to-Sample Are More Effective Than Others |
| KATHRYN SAUNDERS (University of Kansas) |
| Abstract: The arbitrary-matching-to-sample-procedure is often used in teaching receptive-language and other skills to students with autism and other developmental disabilities. For example, a child might be taught to select, from a choice pool of 2 or more pictures, the picture that goes with a spoken word. As another example, matching procedures are used to teach relations between photos and objects. In this talk, I will present a component analysis of arbitrary-matching teaching procedures, trace the roots of the analysis to basic laboratory research, and show how this analysis helps predict which teaching procedures will be effective, and why. The component analysis will be applied to commonly used teaching procedures such as delayed prompting. Two procedures that establish arbitrary matching by ensuring acquisition of its components will be described. |
| |
| Strategies for generating arbitrary match-to-sample performances |
| CAROL PILGRIM (University of North Carolina - Wilmington) |
| Abstract: Although critical to a number of functional skill-training programs, arbitrary match-to-sample (MTS) performances have proven a challenge for many participant populations. This talk will include data from a series of studies involving young typically-developing children as participants. In each case, the target performance was a three-choice arbitrary match-to-sample task involving unfamiliar nonrepresentational line drawings. Study 1 shows the facilitative effects of a programmed instruction series involving identity MTS training, thematic MTS training (matching nonidentical stimuli from a common category or functional grouping), and arbitrary MTS training. Study 2 shows the facilitative impact of adding class-specific reinforcers to the MTS training procedures (choosing B1 given A1 produced Reinforcer 1, while choosing B2 given A2 produced Reinforcer 2, etc). Study 3 shows the emergence of arbitrary MTS performances following simple discrimination training with class-specific reinforcers (choosing A1, B1, or C1 from a group of distractor stimuli produced Reinforcer 1, while choosing A2, B2, or C2 produced Reinforcer 2, etc.). Collectively, these results may provide some directions to explore in applied settings when discrete-trial procedures fail to produce the targeted discriminations.? |
| |
| Modeling “Joint Abstraction” in Conditional Discrimination Procedures |
| WILLIAM J. MCILVANE (University of Massachusetts Medical School) |
| Abstract: This paper will introduce the concept of “joint abstraction” as an adjunct to the concept of “joint attention” that is important in the development of communicative repertoires (and deficit in neurodevelopmental disorders such as autism). Joint abstraction refers to the presence and/or degree of joint control by stimulus features or dimensions in dyads such as those involving an experimenter and a research participant, a therapist and a client, or a teacher and student. In each case, one member of the dyad (e.g., the experimenter) endeavors to arrange a programmed contingency (implicitly or explicitly) designed to influence the behavior of the other member (e.g. the research participant). A high degree of joint abstraction means that the dyad member who arranges the programmed contingency and the other member who responds to the contingency both attend to the same stimulus features/dimensions (or features/dimensions positively correlated with them). In the first part of the presentation, I will review the theory behind the concept of “joint abstraction.” Thereafter, I will discuss two programs of translational research, one with humans with severe neurodevelopmental disabilities and the other with nonhuman primates, each of which has been designed to increase the level of joint abstraction between members of dyads in modeling naturally occurring processes in the development of joint attention. |
|
| |