A Longitudinal Randomized Trial of a Sustained Content Literacy Intervention from First to Second Grade: Transfer Effects on Students' Reading Comprehension

30 elementary schools, 2,952 students, 144 teachers, randomised

US grade 1-2 students in a high-poverty urban district

The mechanism paper for MORE, and the reason to record it separately from the 2024 long-term report: it decomposes transfer by how much of the taught vocabulary appears in the test passage. Schools randomised to treatment or control; treatment received spring grade-1 thematic content literacy lessons in science AND social studies, then summer wide reading of thematically related informational texts, then grade-2 thematic lessons in science. Outcome is a science content reading comprehension measure whose passages were deliberately built at three transfer distances: near (many directly taught words), mid, and far (no taught-word cues). Read at abstract depth - the effect-size decomposition by passage type is reported qualitatively in the abstract, and the per-band numbers have not been extracted from the article.

Overall transfer to science content reading comprehension was ES = 0.18 after 12 months of sustained instruction, plus a smaller summer loss on a domain-general reading measure for treated students. The finding that matters for the knowledge-versus-skills argument is the GRADIENT: transfer effects were strongest on passages containing networks of directly taught vocabulary (near and mid transfer) and weakest on far-transfer passages with no such cues. That is the knowledge hypothesis confirmed in its narrow form and refuted in its broad one - knowledge helps you read about what you learned about, and the help decays as the text moves away from the taught content. It is the cleanest within-study demonstration in this cluster of why aligned and independent comprehension measures diverge by roughly 2x.

Low. School-level randomisation; passive gene-environment correlation cannot produce the treatment-control contrast or the passage-type gradient.

Unreplicated by an independent team; the 2024 Developmental Psychology paper extends the same trial to grade 3-4 outcomes rather than replicating it. The passage-distance gradient has not been reproduced elsewhere, which is unfortunate because it is the most theoretically informative result in the content-literacy literature.