Learning techniques have much smaller effects on deep learning and far transfer than on surface learning and near transfer¶

Claim · All claims
Evidence · 1 study · quant-synthesis r? · q3 · i1 small · n=242 studies
Currency · newest study 2021 · checked against its sources 2026-10-16 (automated, 1 of 1 entry) · not reviewed by a person

Whether the ten learning techniques Dunlosky et al. (2013) reviewed help with deep, relational learning and transfer, or mainly with surface recall.

Subclaims¶

q3 i1 Effects were d = 0.60 on surface outcomes (1,473 effects) against 0.26 on deep outcomes (109 effects), and 0.61 on near transfer against 0.39 on far transfer; most studies measured surface or factual outcomes. → Donoghue & Hattie 2021

Evidence¶

Donoghue & Hattie 2021¶

Donoghue, G. M., & Hattie, J. A. C. (2021). A Meta-Analysis of Ten Learning Techniques. Frontiers in Education, 6, 581216. doi:10.3389/feduc.2021.581216

q3 · meta-analysis (moderator analysis, 1,619 effects) · i1 · small effect on deep outcomes, d = 0.26 (0.60 surface); far transfer d = 0.39 · n=242 studies · quant-synthesis · r?

The authors coded each outcome for depth (surface, content-specific, against deep, more generalisable) and for transfer distance. Effects on surface outcomes were d = 0.60 (1,473 effects) and on deep outcomes d = 0.26 (109 effects). Near-transfer effects were d = 0.61 and far-transfer effects d = 0.39. They name the dominance of surface and factual outcomes as the review's major limitation.

"The effects were higher for Surface (d = 0.60, SE = 0.021, N = 1,473) than for Deep processing (d = 0.26, SE = 0.064, N = 109)."

Studies bearing on this claim¶

Does instruction that raises near transfer also raise far transfer to a different domain or form? This table gathers the studies behind this claim and the 5 narrower claims that bear on it: 6 distinct studies (3 causal, 2 quant-synthesis, 1 review), q2–q4, 1 of 6 with an effect size coded. One row per study, codes as each entry records them; no effect is averaged and no verdict is drawn. Built by scripts/synthesize_claim.py from eval/synthesis/; a row that looks wrong is fixed on the claim it comes from.

Discussion¶

A course aiming at understanding or transfer cannot take the headline effects of these techniques at face value. Only 109 of 1,619 effects measured deep outcomes, so the evidence for deep learning is thin rather than negative.

Related Claims¶

• Retrieval Practice Improves Transfer — the general claim this finding bears on, with the studies behind it

• Across ten learning techniques, distributed practice and practice testing have the largest effects and underlining and summarization the smallest — related

• Statistically significant differences in approaches to learning can constitute small effects of little theoretical or practical importance — related
• Test-score increases occurred only for identical and near-transfer problems; no increases were found for far-transfer problems in any of the three conditions — a narrower finding that bears on this claim
• Self-explanation effects hold on far-transfer but not near-transfer questions in a geometry classroom study — related
• The review reports that simulation games support near transfer of knowledge but evidence for long-term or far transfer is less robust — related
• In one 51-learner experiment on help formats in computer-based electrical engineering instruction, far-transfer scores were at the floor in every condition, and time spent on instruction did not differ significantly between conditions — related
• Transfer of test-enhanced learning is greatest across test formats, to application and inference questions, to medical diagnosis problems, and to mediator and related word cues — related