Objectives: The aim of the present study was to promote self-initiated question asking in school-aged children with ASD by conducting a robot-intervention and a human trainer-intervention. The objective was to investigate the effectiveness of both interventions.
Methods: Data were collected using a combined crossover multiple baseline design across participants. Six children were randomly assigned to two experimental groups. During the sessions a statement-question-action scenario was used to elicit self-initiated questions. During the intervention, using a least-to-most prompting hierarchy, the children were prompted either by the robot or the human trainer to initiate a question related to a statement. Data-analysis involved visual inspection and the calculation of Taunovlap that examines data nonoverlap between phases. The overall Taunovlapwas also calculated.
Results: The results revealed that the number of self-initiated questions significantly increased between baseline and the first intervention for both experimental groups. The values of Taunovlap showed a significant increase for all children. The number of self-initiated questions remained high during the subsequent phases of the study.
Conclusions: Both the robot-intervention and the human trainer-intervention effectively promoted self-initiated question asking in children with ASD. The high number of self-initiated questions during follow-up indicates that both experimental groups maintained this skill. Practical implications and directions for future research will be discussed.