In this meta-analysis, we investigate whether collaborative discussion about texts have generative value for students by enabling them to demonstrate transfer of their comprehension abilities to new texts and novel tasks. We examine evidence of the effects of text-based discussion from quasi-experimental and experimental studies where researchers have assessed the effects of discussion on comprehension-related measures that are independent of the texts discussed. There is evidence that classroom discussions can enable students to acquire the habits of mind to transfer their comprehension capabilities to new texts and tasks. Results show wide variability in effects with some discussion approaches demonstrating stronger effects than others. Findings also suggest that lower-achieving and younger students may benefit more from these discussions, potentially mitigating disparities in reading development. The duration of interventions may also play a role, with shorter, focused discussions yielding better results than extended interventions.