Strict phonics instruction outperforms mixed phonics: Evidence from a large-scale study

Abstract

Background Learning to read is a foundational milestone in formal education, and a large body of research supports the effectiveness of systematic phonics instruction in early literacy development. However, much less is known about whether different implementations of phonics instruction influence children’s reading outcomes. This study investigates whether variation within phonics-based instruction—specifically, strict versus mixed phonics approaches—is associated with differences in early reading development. Methods We combine survey data from more than 9000 first-grade teachers with longitudinal assessment data for over 139,000 pupils in France. Teachers were classified according to the instructional approaches they reported using and the characteristics of the reading materials employed in their classrooms. Pupils’ reading fluency and reading comprehension were measured using nationally administered standardized assessments conducted during Grade 1 and at the beginning of Grade 2. Results Multilevel analyses showed that pupils taught using strict phonics approaches achieved higher reading fluency and reading comprehension than those taught with mixed phonics approaches, with differences corresponding to approximately 0.11–0.16 standard deviations. These effects were substantially larger among pupils with weaker pre-reading skills and those attending socioeconomically disadvantaged schools. The findings remained robust across multiple complementary analyses designed to address potential selection effects and omitted-variable bias. Conclusion The findings demonstrate that meaningful differences in reading outcomes may arise not only from whether systematic phonics instruction is provided, but also from how it is implemented. They further suggest that optimizing the coherence of early phonics instruction may be particularly beneficial for pupils most at risk of reading difficulties.

The authors' abstract, as published at the source. Learning and Instruction, 2026 · DOI ↗