Classroom ICT Use and Reading Achievement: Evidence from Spanish Primary Schools

This paper explores the relationship between students’ ICT use and reading achievement using data on Spanish pupils participating in PIRLS 2021. Specifically, we examine whether the intensive use of digital tools for certain classroom activities contributes to improvements in reading comprehension. To address this question, we apply a causal machine learning approach based on conditional inference forests (CIF), which allows us to approximate counterfactual outcomes while conditioning on a rich set of student, family, and school covariates. This method enables the identification of both average and heterogeneous treatment effects, supporting a causal interpretation of the estimated relationships based on observables. Statistical uncertainty is quantified through resampling-based confidence intervals, which reinforce the robustness of the results. Overall, our findings indicate that the effects associated with more intensive classroom use of digital tools are statistically significant but very small in magnitude, suggesting a limited educational relevance. Nevertheless, clear socioeconomic differences emerge, with information-search and content-creation activities showing the strongest positive effects among students from more advantaged backgrounds, while structured digital writing tasks yield comparatively larger benefits for those from less advantaged families.