|
A Focus on Generalization of Skills for Young Children with Autism and Developmental Delays |
Saturday, May 23, 2009 |
1:00 PM–2:20 PM |
North 125 |
Area: AUT/EDC; Domain: Applied Behavior Analysis |
Chair: Ilene S. Schwartz (University of Washington) |
Discussant: Carol Ann Davis (University of Washington) |
Abstract: Slow acquisition and generalization of skills continues to be a struggle for children with disabilities. Children with autism and developmental disabilities often exhibit low rates of engagement, deficits in language and symbolic play with the lack of generalization or transfer of skills. Each of these deficits provides significant challenges for children with autism both in terms of their learning and their long-term outcomes. Given the abundant number of skills young children with autism and developmental delays with must learn, behavior analysts must evaluate and apply generative approaches to instruction so that some skills are taught and others emerge without direct teaching. Three papers will offer research findings on studies of generalization. Axe and Sainato will present their study of matrix training for young children with autism,” “Matrix Training of Instruction Following of Pre-Academic Skills with Preschoolers with Autism”. Duckett and Schwartz will offer a paper examining the use of matrix training for receptive language skills. Garfinkle will present a study of the untrained generalization of engagement in very young children with a focus on play materials. Carol Davis will serve as a discussant in her review of these papers offering implications for practice and recommendations for future directions for research. |
|
“Matrix Training of Instruction Following of Pre-Academic Skills with Preschoolers with Autism” |
JUDAH AXE (Simmons College), Diane M. Sainato (The Ohio State University) |
Abstract: Given the abundant number of skills children with autism must learn, behavior analysts must evaluate and apply generative approaches to instruction so that some skills are taught and others emerge without direct teaching. Matrix training is a generative approach and was evaluated with four preschoolers with autism learning to follow trained and untrained action-picture instructions (e.g., circle the pepper, underline the deer). Six actions and six pictures were each arranged on axes of matrices. Errorless teaching was used to train the instructions along the diagonal of the matrices. A multiple probe across behaviors design was employed. Mean interobserver agreement of responses across participants, phases, and tiers was above 93%. Two participants readily followed untrained instructions with the minimal number of trained instructions. Two participants required training on more than the minimal number of instructions to demonstrate generalization to untrained instructions. Three of the four participants followed instructions to perform the actions with previously known pictures, letters, and numbers. This study extended previous research by using matrix training to teach picture identification and writing skills to preschoolers with autism. Matrix training is an efficient approach to instruction and can be used to teach academic skills to children with autism. |
|
The Matrix Revisited: Recombinative Generalization and Young Children
with Autism |
LAURELIN DUCKETT (The University of Washington), Ilene S. Schwartz (University of Washington) |
Abstract: Slow acquisition and generalization continue to be a struggle for
children on the autism spectrum. This study looked at the use of recombinative
generalization and the effectiveness of using matrix-training procedures to
teach receptive language to young children with autism. An AB design with
multiple probes was used and then replicated across two children and three
matrices, resulting in a total of 6 AB designs. Two preschool children with
autism participated in the study, a five-year-old girl and a three-year-old boy.
The three matrices involved teaching receptive language combinations of
adjective/noun, agent/action, and emotion/preposition. Both children were able
to acquire new receptive language using discrete trial training and then
generalized to more then 3 times the number of stimuli originally taught. The
results of the study show that matrix-training may be an effective and highly
efficient way to enhance an intervention with young children with autism by
decreasing the number of stimuli being taught and increasing the number of
stimuli acquired. |
|
The Untrained Generalization |
ANN N. GARFINKLE (University of Montana) |
Abstract: : Children with autism often have low rates of engagement, deficits in symbolic play, and with the generalization or transfer of skills. Each of these deficits
provides significant challenges for children with autism both in terms
of their learning and their long-term outcomes. The current study
examines toy materials and their attributes that may have facilitated
the untrained generalization of engagement across play materials. Three
male children under the age of three, diagnosed with autism, were
observed over a three-month period. In addition three age-match
typically developing children were also measured. In this time period a
momentary time sample recorded children's engagement and play material.
The data show, with a high degree of confidence, that both the typically
developing children and those diagnosed with autism increased the number
of different play materials with which they interacted. This increase
took place in the absence of training during the observation period. By
examining the features of the toy materials it may possible to determine
which attributes facilitate the generalization of toy play in young
children with autism. Both practical and theoretical implications will
be discussed. |
|
|