How generalizable is the relationship between feedback and reading literacy across different competence levels, multilingual learners, and countries? A meta-analytic approach

Contrary to theoretical assumptions, only weak relationships have emerged between feedback and reading literacy in non-experimental settings. Our study contributes to a better understanding of those weak relationships by shedding light on generalizability across specific student groups (different competence levels, multilingual learners) and contexts (countries). Using a meta-analytic approach, we analyzed the PISA 2018 dataset with data from 505,906 students from 75 countries. Our results indicate a weak overall relationship between feedback and reading literacy, an uneven distribution of feedback across competence levels, and heterogeneity across countries. In contrast, multilingual learners do not profit differently from feedback compared to their monolingual peers. Future research should focus on the mechanisms behind these differences.