Human performance differences following the learning of behavior sequences by chaining procedures.

Abstract

Chaining procedures (forward chaining, backward chaining, and whole task craining) are used to teach sequences of behaviors to diverse populations. Recent research comparing the three procedures has revealed differences between the procedures in terms of acquisition variables e.g., number of errors made during learning, time required to the learn a sequence). Much less research has examined whether the procedures produce differences in performance following acquisition. A series of six experiments investigating performance differences following exposure to four chaining procedures are presented. The four chaining procedures included the three traditional chaining procedures--forward, backward, and whole task--as well as a random chaining technique. Overall, results indicated that forward chaining led to the most accurate and fastest performance of responses in the absence of the external antecedent stimuli used during training. Forward chaining also led to more accurate ordering of these stimuli following training. Random chaining led to the most accurate and the fastest performance of responses following presentation of the antecedent external stimuli out of sequence order. The theoretical and practical implications of these findings are discussed.