Sample size does not significantly moderate the gloss effect: large studies (ES = 0.88) versus small studies (ES = 0.72) do not differ significantly¶

Claim · All claims
Evidence · 1 study · quant-synthesis r3 · q3
Currency · newest study 2014 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q3 i? Multiple glosses had a large, positive effect in large studies (ES = 0.88) and a medium effect in small studies (ES = 0.72), but the difference between large and small studies was statistically non-significant (Q = 0.35, p > 0.05). → Vahideh Sadat Vahedi 2014

Evidence¶

Vahideh Sadat Vahedi 2014¶

Vahideh Sadat Vahedi, Behzad Ghonsooly, and Reza Pishghadam. (2014). Vocabulary Glossing: A Meta-Analysis of the Relative Effectiveness of Different Gloss Types on L2 Vocabulary Acquisition. Teaching English with Technology, 16(1), 3-25. http://www.tewtjournal.org

q3 · i? · quant-synthesis · r3

Moderator analysis grouping studies with sample size greater than 60 as large (n=21) and below 60 as small (N=13). The article reports the difference between the subgroups' effect sizes was statistically non-significant. No effect size for the moderator contrast is printed, so impact is null.

"However, the difference between the large and small studies effects sizes proved to be statistically non-significant (Q= 0.35, p> 0.05)."

Discussion¶

Related Claims¶

• Learning context (ESL vs EFL) does not significantly moderate the gloss effect, though EFL contexts showed a nominally larger effect — related
• Research design does not significantly moderate the gloss effect: randomized (ES = 1.16) versus matched (ES = 0.71) studies do not differ significantly — related
• Program intensity is a significant moderator: high-duration gloss interventions yield larger effects (ES = 1.48) than low-duration ones (ES = 0.57) — related
• Sampling size is not an effective moderator of AR effects, with large samples (50+) at 0.647 and small samples (1-49) at 0.262 — related