Syntes: Guo et al. (2024) — homework versus no homework g = 0.45 in mostly old, primary-school experiments at high risk of bias

Källa: Guo, X., Li, J., Xu, Y., Hu, X., Liu, Y., Xing, X., Li, X., White, H. & Yang, K. (2024). The relationship between homework time and academic performance among K-12: A systematic review. Campbell Systematic Reviews 20(3), e1431. https://doi.org/10.1002/cl2.1431

Verified content

• [full text] 11 controlled studies (2 RCTs, both high risk of bias; 9 non-randomised: 4 moderate, 2 serious, 3 critical risk of bias); most published 1939–1995; one with a follow-up.
• [full text] Homework versus no homework g = 0.45 (95% CI 0.24–0.66; 14 comparisons; I² = 71.3%); arithmetic computation g = 0.46 (n = 5), problem solving g = 0.17, concepts not significant.
• [full text] The pooled homework-versus-none estimate comes from 10 of the 11 studies (published 1939–1995): by the review's Table 2, six in grades 3–6, one in grade 7, one in grade 8 and two in senior high-school English classes; the grade-2 randomised trial with the only follow-up (Dolean & Lervåg 2021) compared more with less homework and is not in the pooled estimate. The review's text lists only eight studies with a grade and no grade-level subgroup analysis was possible. (Corrected 2026-11-04 against the full text; the earlier line said 8 studies, 7 in grades 2–6 and 1 in grade 8.)

Limits

• [assessment] Weak, old evidence; cannot itself test a grade contrast. Used 2026-11-03 to weaken C27: the K–6 "small" holds only for correlations.

Why it is in the library

The only recent systematic review that pools controlled homework-versus-none experiments, mostly in primary school.

Added 2026-11-03 by the project (nightly attacks; read: full text). [abstract] = from the published abstract; [full text] = read in the full text; [assessment] = the project's own assessment.