Writing instruction improves students’ writing skills differentially depending on focal instruction and children: A meta-analysis for primary grade students

24 studies, k = 166 effect sizes, N = 5,589 participants

Kindergarten to Grade 3 (roughly ages 5-9)

The most directly relevant meta-analysis in the archive for the youngest writers (Educational Research Review 34, 100408). Effects are decomposed by outcome dimension (quality, productivity, fluency, other), by instructional focus (transcription, self-regulation strategies including SRSD, and others), and by learner (initially weak vs typical writers). Outcomes are researcher-scored composition measures. Also tests dose as a moderator, which most syntheses in this literature do not.

Overall mean ES = 0.31 (quality 0.32, productivity 0.31, fluency 0.15, other 0.34) - moderate and positive, and notably smaller than the elementary and adolescent syntheses report for their headline treatments. SRSD had "large and consistent effect sizes across the outcomes (.59 to 1.04)". TRANSCRIPTION INSTRUCTION DID NOT YIELD STATISTICALLY SIGNIFICANT EFFECTS ON ANY DIMENSION OF COMPOSITION, owing to large variation of effects across studies - the sharpest available tension with Santangelo & Graham (2016), and the reason the transcription topic cannot be graded higher. Variation in instructional DOSAGE (total length of instruction) did NOT explain variation in effect sizes, which is a genuinely awkward finding for any dose-response story. The average effect on writing quality was LARGER for writers with weaker writing skills than for those with typical skills.

Low for direction. The larger effect for initially weak writers is a subgroup comparison within randomised or quasi-randomised trials and is at least partly regression to the mean; it should not be read as evidence that instruction closes heritable gaps.