Syntes: Racsmány, Szőllősi & Marián (2020) — reversing TE by feedback depends on practice success

Källa: Racsmány, M., Szőllősi, Á., & Marián, M. (2020). Reversing the testing effect by feedback is a matter of performance criterion at practice. Memory & Cognition, 48, 1161–1170. https://doi.org/10.3758/s13421-020-01041-5

Metadata

• ID: LEARN-015
• Tier: RCT (three experiments)
• N / design: Swahili–Hungarian pairs; prior-study load 1× / 3× / 6× then six no-feedback practice cycles (restudy vs test vs baseline); 7-day delay; six final tests with feedback
• Population: vuxen-livslangt (Exp 1 n=29; Exp 2 n=30; Exp 3 n=24)
• Subject: whether feedback-induced reversal of TE after delay depends on practice-phase retrieval success

Finding

1. Exp 1 (1× study; practice success 25.3–28.7%; no practice feedback): first final test > restudy, t(28)=2.29, p=.03, d=0.43 (Fig. 3: retested ~23% vs restudied ~16%); TE reversed after subsequent final-test feedback cycles.
2. Exp 2 (3×; practice 66.1–75.0%): first final TE d=1.33; no reversal across final feedback cycles.
3. Exp 3 (6×; practice 76.1–80.9%): first final TE d=1.71 (Fig. 3: ~61% vs ~28%); TE persisted through several final feedback cycles.
4. Authors conclude practice-success level, not feedback-during-practice conflation, was the key factor in earlier reversal findings (Storm et al. 2014 Exp 1).

Mechanism

Supports bifurcation-style accounts: low no-feedback practice success leaves most tested items unrehearsed; final-test feedback then helps restudied items more. High practice success preserves TE despite later feedback.

Limits

Word-pair lab paradigm; young adults; first-final means partly from figure caption; reversal after final-test feedback is a different design force from C0’s single delayed criterion. Between-experiment comparison of TE magnitude (not a single within-experiment Training×Repetition interaction like Gupta 2022).

This claim dies if …

(For C1 residual no-feedback equality prong:) equal delayed TE vs restudy when practice success is experimentally low and high without practice feedback — Racsmány shows TE present at low success but not equal (d=0.43 vs 1.71).

Swedish implication

Without feedback, keep practice success reasonably high if the goal is a durable retrieval-over-restudy advantage; low unaided success shrinks (does not erase) that advantage on a first delayed test.

Links

• DOI: https://doi.org/10.3758/s13421-020-01041-5
• Related: gupta-pan-rickard-2022-prior-episodic-learning, rowland-2014-testing-effect-meta, chan-etal-2024-testing-effect-independent, storm-friedman-murayama-bjork-2014 (cited; not separately shelved)