Prediction of production: Elicited imitation and spontaneous speech productions of language disordered children

Abstract

Two forms of an imitation task designed to elicit eleven language behaviors were presented to 32 language impaired children (mean age 7;4, mean MLU 2.9). The tasks differed in terms of the presence or absence of contextual support. To determine how well the imitation tasks could predict spontaneous productions, a multiple regression analysis was performed using the children's spontaneous productions of each behavior as the criterion variable and responses to the same behaviors on each imitation task plus age, sex, MLU, and a measure of memory as predictor variables. The major finding was that predictions of spontaneous production varied among the eleven language behaviors. Additionally, the presence of contextual support had little effect.