|
| Further Analysis of Variables that Influence Mand Training |
| Monday, May 25, 2009 |
| 1:30 PM–2:50 PM |
| North 128 |
| Area: DDA/AUT; Domain: Applied Behavior Analysis |
| Chair: Lisa C. Winborn-Kemmerer (West Virginia University) |
| Discussant: Joel Eric Ringdahl (University of Iowa) |
| CE Instructor: Bryan Davey, Ph.D. |
| Abstract: Individuals with developmental disabilities and autism often present a variety of communication deficits. This can include the use of vocal sounds or babbling and the absence of vocal speech. Depending on the individual’s existing communication skills, different topographies of mands (e.g., vocal sounds, manual signs, communication cards, microswitch devices) may be targeted for training and alternative strategies may be needed to effectively increase communication. For example, training and reinforcing more than one mand may increase the individual’s functional communicative repertoire. In addition, training mands across multiple settings and functional contexts may affect the errors made with mands, the individual’s ability to generalize mands, and the amount of problem behavior displayed during training. In this symposium, the presenters will discuss the role of several variables (e.g., training multiple mands, extinction of mands, mand errors, stimulus generalization, and use of lag schedules of reinforcement) that may influence mand training and the use of functional communication skills. |
| |
| Stimulus Generalization and Extinction of Mands During Functional Communication Training |
| LISA C. WINBORN-KEMMERER (West Virginia University), Jennifer Wolfe (University of Louisville), Allison Cheek (University of Louisville) |
| Abstract: The purpose of this study was to evaluate the stimulus generalization of two mand topographies across settings and to further evaluate problem behavior and mands when one of the mands was placed on extinction. Two children with developmental disabilities and autism participated in this study. A functional analysis was conducted across one setting to identify the reinforcers for problem behavior. Next, problem behavior was assessed for the escape condition of the functional analysis in an alternative setting. Concurrent FCT programs were then implemented across both settings for problem behavior maintained by negative reinforcement. A microswitch was trained in one setting and a communication card was trained in the other. Following FCT, stimulus generalization (novel setting) was assessed for each mand. Finally, both mands were available for reinforcement in each setting, however, extinction was provided for one of the mands. All phases of this study were conducted within a multielement and reversal designs. Results showed that both mands generalized to novel settings and that the children used the alternative mand, when one of the mands was placed on extinction without increased problem behavior. Interobserver agreement was obtained across 30% of sessions and averaged above 80%. |
| |
| An Evaluation of Mand Errors Across Functional Contexts During Functional Communication Training. |
| TERRY FALCOMATA (University of Iowa), David P. Wacker (University of Iowa), Joel Eric Ringdahl (University of Iowa), Kelly M. Vinquist (University of Iowa), Anuradha Salil Kumar Dutt (University of Iowa) |
| Abstract: The purpose of this study was to evaluate stimulus generalization errors during and following FCT and to examine the conditions under which errors were most likely to occur. Specifically, we evaluated the occurrence of stimulus generalization errors in manding across functional contexts in which one or more functions of problem behavior were identified. Errors were evaluated across three functional contexts (i.e., restricted tangible, attention, and demand) to allow for a direct evaluation of stimulus generalization within and across reinforcement classes (i.e., positive reinforcement, negative reinforcement) and functional contexts. We first conducted functional analyses to identify the function(s) of problem behavior with three children diagnosed with developmental disabilities. Next, we implemented FCT in which three novel manual signs across three respective functional contexts (i.e., tangible, attention, demand) were trained and the occurrence of stimulus generalization errors was evaluated across all three functional contexts. The results suggested that variables relating to reinforcement class affected patterns of stimulus generalization regardless of the presence or absence of functions of problem behavior within respective functional contexts. Interobserver agreement was obtained during at least 30% of sessions and averaged above 90%. |
| |
| The Use of Lag Schedules of Reinforcement to Increase
the Variability of Vocal Production in Children with Developmental Disabilities |
| ALLISON TETREAULT (West Virginia University), Claire St. Peter Pipkin (West Virginia University), Brittany Glass (West Virginia University) |
| Abstract: Children diagnosed with developmental disabilities present a wide range of communication deficits. These deficits may range from complete mutism and nonuse of words to mild grammar deviations. While there is abundant literature on teaching strategies for children who use words, less attention has been given to strategies for subjects who engage in only the production of sounds. Intuitively, shaping procedures that reinforce successive approximations to the target behavior (here, a word) would seem to be a useful strategy. However, some children do not imitate vocal sounds and do not produce enough sounds to allow for the selection of a shapeable vocal unit (e.g., “eat” or “candy” cannot be shaped from the sounds /b/ or /p/). We investigated a lag reinforcement schedule to increase the vocal variability of young pre-verbal children with autism. By increasing variable vocal production, a wider array of sounds developed in the children’s vocal repertoire, which can be selected from and shaped into functional words. Future applications of this technology are suggested. Interobserver agreement was obtained for 30% of sessions and averaged above 80%. |
|
| |