The pooled effect of writing instruction for US ELL students is robust to sensitivity analyses but funnel plot asymmetry suggests possible missing positive effect sizes¶

Claim · All claims
Evidence · 1 study (2 entries) · quant-synthesis r? · q3
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q3 i? Pooled estimates remained at g = 0.33 across varied within-cluster correlations and after excluding studies with attrition concerns, but the funnel plot appeared asymmetrical while Egger's test detected no significant asymmetry. → Aitken 2026

Evidence¶

Aitken 2026¶

Aitken, A. A., Graham, S., Cao, Y., & Nusrat, A. (2026). The Impact of Writing Instruction for English Language Learners in the United States: A Meta-Analysis. Educational Psychology Review. https://doi.org/10.1007/s10648-026-10197-6

q3 · i? · quant-synthesis · r?

Sensitivity analysis varying the assumed within-cluster correlation ρ across 0.20, 0.40, 0.60, and 0.80 in the RVE framework. The article reports "The pooled ES estimates remained at g = 0.33", supporting robustness to dependence assumptions.

"The pooled ES estimates remained at g = 0.33, indi- cating that the results were not sensitive to the choice of ρ."

Aitken 2026 (2)¶

Aitken, A. A., Graham, S., Cao, Y., & Nusrat, A. (2026). The Impact of Writing Instruction for English Language Learners in the United States: A Meta-Analysis. Educational Psychology Review. https://doi.org/10.1007/s10648-026-10197-6

q3 · i? · quant-synthesis · r?

Publication bias assessment using funnel plots, trim-and-fill, and Egger's regression test on study-level aggregated effect sizes. The article reports the funnel plot "appeared asymmetrical", while Egger's test (intercept = 0.49, p =.363) found no statistical evidence of publication bias.

"The funnel plot appeared asymmetrical, suggesting the potential absence of larger positive effect sizes"

Discussion¶

Related Claims¶

• Writing intervention effects for US ELL students do not vary significantly by study quality indicators, participant characteristics, or intervention characteristics — related
• Publication bias is present in the mathematics–writing literature, though the pooled estimate is robust in sensitivity analyses — related
• Publication-bias checks indicate the pooled effect is unlikely to result from publication bias — related
• Evidence of publication bias exists, but the pooled video effects are robust to it — related
• Writing instruction produces a statistically significant positive overall effect on the writing performance of K-12 English language learners in the United States (g = 0.33) — possibly the same claim (merge candidate)
• Writing interventions improve writing quality, the most frequently assessed outcome, with an effect of 0.33 for US ELL students — related
• Funnel-plot asymmetry ran contrary to small-study effects because larger-sample studies reported larger effect sizes, likely reflecting longer training programs — related
• Publication bias is present in the embodied vocabulary learning literature, suggesting pooled effects may overestimate the true benefit — related
• Pooled active-break effect estimates are robust to sensitivity analyses varying within-study correlation and excluding individual studies — related
• Publication bias is unlikely to explain the consistency of SRL intervention effect sizes — related
• The text generation research field shows a small-study effect suggesting publication bias — related
• No evidence of publication bias was found in the meta-analysis — related
• Publication-bias assessments substantially reduce the intercept prediction, often indicating no positive transfer when the identified moderators are absent — related
• Process writing yields a meaningful but not statistically significant effect (g = 0.51) on US ELL students' writing — a narrower finding that bears on this claim
• Strategy instruction significantly improves the writing of US ELL students (g = 0.36) — related
• A cluster-robust Egger-type test found no statistically detectable small-study effects, interpreted as absence of evidence rather than evidence of no publication bias — related
• Funnel-plot analysis finds no evidence of missing small-effect studies, though asymmetry from few negative estimates suggests possible small bias against null-negative results — related
• The pooled self-efficacy–partnership correlation appears robust to the file-drawer problem — related
• Publication bias is indicated for the performance-approach and performance-avoidance anxiety effects but not for the depression effects — related
• Publication bias evidence is mixed: marginal asymmetry for tutees but no missing studies, and no bias for tutors — related
• Publication bias analyses suggest possible small-sample bias for SCED studies, with trim-and-fill reducing the adjusted mean effect to 1.01 — related
• No evidence of dissemination bias was detected in the meta-analysis — related
• Publication-bias checks indicate limited risk of publication bias in the extensive reading literature synthesized — related
• Publication bias is minimal in the meta-analytic estimate of the inhibitory control–mathematical ability correlation — related
• No evidence of publication bias was detected in the pooled L2 gender-difference literature — related
• Formal tests find no substantial evidence of publication bias or small-study effects in the meta-analysis — related
• Funnel-plot asymmetry suggests possible selection or publication bias, but p-curve analyses indicate evidential value for the pooled relationships — related
• Publication bias could not be ruled out in the pooled teacher-student SEC estimate, though the association persisted across sensitivity analyses — related