Trials to criterion: a within‐program prediction of post‐program generalization

Abstract

Data‐based instruction is an important and effective procedure when teaching handicapped learners; at present, however, no validated decision rules are available to predict post‐program generalization. This study investigated the extent to which learning set analysis of a severely handicapped subject's performance was capable of providing decision rules when teaching for a generalized outcome. Generalized compliance was trained and training data analyzed to determine how data trends across training examples compared with post‐program generalization. Results show that the greatest generalization effect occurrred when training data trends were steepest downwards. Discussion of the findings includes suggestions concerning future research to further define the parameters of decision rules necessary to teach for generalization.

Additional information

Notes on contributors

Paul M. Gannon

Paul M. Gannon, Ph.D., is Lecturer in Special Education, School of Education, James Cook University, Townsville, Queensland 4811.