A digital safe space for learning: how computer-assisted instruction supports students with lower academic self-concept

Computer-assisted instruction (CAI) is a valuable approach for managing classroom heterogeneity by providing feedback tailored to students’ individual needs. While previous research has primarily focused on the cognitive mechanisms underlying CAI’s effectiveness, it has often overlooked the social-cognitive processes that may contribute to its success by offering students a private learning space away from the judgement of their peers. The present study compared CAI to conventional classroom instruction over a two-week period with 389 tenth-grade history-geography students. First, we expected that CAI would result in better academic learning compared to conventional classroom instruction. Second, we tested alternative hypotheses, that learning gains in CAI would vary depending on students’ levels of academic self-concept (social-cognitive hypothesis) and prior knowledge (cognitive scaffolding hypothesis). Consistent with the social-cognitive hypothesis, multilevel modeling revealed that students with initially lower levels of ASC experienced higher learning gains in CAI compared to conventional classroom instruction, while no benefits were observed for students with lower prior knowledge. Additionally, our results indicate a disruption of the predictive relationship between ASC and subsequent academic performance in the CAI group, compared to the presence of this association in the conventional classroom instruction group. These findings suggest that CAI benefits low-confidence students regardless of their initial ability level and that this method offers a protective effect against the impact of perceived incompetence on subsequent performance.