Abstract
Research generally supports the use of error-correction procedures that require active student responses (ASRs). However, some recent research suggests that requiring ASRs after errors is not always advantageous for individuals with autism spectrum disorder (ASD). We examined acquisition in a receptive identification task for two learners with ASD under three error-correction conditions: required ASRs, model, and modified model that minimized spontaneous ASR. In the model condition, both individuals emitted spontaneous responses to the model even though none was required, extending previous research to the receptive domain. For both learners, the model error-correction was effective; however, minimizing the spontaneous ASR in the modified model condition impacted its efficacy compared to the required ASR condition for one learner. For the other learner, error-correction procedures that required ASRs were not effective. Results are discussed in terms of empirical evaluation of error-correction procedures and suggest that imitative behavior (e.g., emitting an ASR when none is required) might function as a prerequisite skill that predicts the efficacy of model-based error-correction procedures for learners with ASD.