A Synthesis of Research on Language of Reading Instruction for English Language Learners

17 studies meeting inclusion standards

English language learners, predominantly Spanish-dominant elementary students in the United States

Best-evidence synthesis - a systematic search with explicit inclusion standards and effect-size quantification, rather than an all-comers meta-analysis. The authors themselves close by saying the number of high-quality studies is small and that longitudinal randomised designs are needed, which is the correct reading of a +0.45 median from 13 studies.

Among 13 studies of elementary reading for Spanish-dominant students, 9 favoured bilingual approaches on ENGLISH reading measures and 4 found no differences; median effect size +0.45, sample-weighted +0.33 (p < .05). One of two heritage-language studies and both secondary studies favoured bilingual approaches. The synthesis favours paired bilingual strategies that teach reading in the native language and in English at different times of day.

Medium-to-high. Most included studies are quasi-experimental comparisons of programmes that families and districts chose; the meta-analysis inherits that confounding.

Superseded in part by the same authors' 2012 update, which reports a smaller pooled effect (+0.21) and, crucially, reports that the largest and longest evaluations - including the only multiyear randomised evaluation of transitional bilingual education - found NO difference by the end of elementary school. That is the standard deflation pattern this archive expects, appearing inside one research programme.