View article

The current study identifies relations among students’ natural language input, individual differences in reading commitment, and learning gains in an intelligent tutoring system. Students (n= 84) interacted with iSTART across eight training sessions. Linguistic features of students’ generated self-explanations (SEs) were analyzed using Coh-Metrix. Results indicated that linguistic properties of students’ training SEs were predictive of learning gains, and that the strength and nature of these relations differed for students of low and high commitment to reading.