Timing of feedback and retrieval practice: a laboratory study with EFL students

The study examines how the timing of feedback affects the benefits of retrieval practice for long-term retention among EFL learners. Building on a century of research showing that retrieval practice enhances long-term memory more than restudying, recent work has highlighted feedback as a key moderator of the testing effect. Prior studies report mixed results: some favor delayed feedback for long-term retention, others find no timing difference, and still others show immediate feedback superiority. No prior work had tested these issues with Saudi EFL students using prose passages and short-answer questions across multiple retention intervals. This study addresses that gap by evaluating immediate versus delayed feedback effects on retrieval practice outcomes at immediate, one-week, and one-month intervals.

The review outlines multiple purposes and types of feedback (e.g., knowledge of results, knowledge of correct response, elaborated feedback, answer-until-correct) and notes that more informative feedback often outperforms simple correctness signals. Empirical findings on feedback timing during retrieval practice are mixed. Several studies report advantages for delayed feedback (after a block or ≥1 day), while others find no timing effect or benefits for immediate feedback, possibly depending on task type (rule learning vs declarative memory), feedback duration, and assessment format. In language learning, timing effects have been frequently studied in vocabulary learning with inconsistent results, and little is known for EFL prose comprehension using short-answer tests. Prior research often combined testing with feedback without isolating feedback’s unique contribution relative to restudy and testing without feedback, especially in EFL contexts. The present study compares restudying, retesting without feedback, and retesting with immediate or delayed feedback among Saudi EFL students to clarify feedback timing effects on long-term retention.

Design: Laboratory experiment with a 4 (learning condition) × 3 (retention interval) repeated-measures framework at test. Conditions: (1) Repeated Studying; (2) Repeated Testing—no feedback; (3) Repeated Testing—immediate feedback; (4) Repeated Testing—delayed feedback. Participants: 190 male Saudi undergraduate EFL students (native Arabic speakers; ages 19–23, M=21.4, SD=1.6) from Umm Al-Qura University were recruited; 13 attrited, yielding N=177: Repeated Studying (n=46), Repeated Testing—no feedback (n=46), Immediate feedback (n=45), Delayed feedback (n=40). Materials: One 487-word prose passage (Reading Explorer 2) and a 15-item short-answer test targeting recall of key details. Four mathematical distractor tasks were interleaved to limit working-memory reliance. A tailored PowerPoint controlled stimulus and timing. Procedure: Phase 1 learning differed by group. Repeated Studying: read passage 10 min + 5-min math distractor, repeated four times. Repeated Testing—no feedback: read 10 min + 5-min distractor, then three short-answer test trials (15 items each), with a distractor task between trials; 1 min per item; no feedback. Immediate feedback group: same as testing-no feedback, but after each 1-min response period, the correct answer appeared for 10 s (per item) on all three trials. Delayed feedback group: same as immediate feedback group, but feedback was presented after each test trial as an answer review for 150 s total (10 s per answer). Phase 2 immediate test: 15 short-answer items in random order, 1 min per item + 10-s blank screen after each. Phase 3: unexpected one-week delayed test with re-randomized order; same timing. Phase 4: unexpected one-month delayed test; same timing. Scoring and Analysis: Two raters scored each response: fully correct=2, partially correct=1, incorrect/blank=0; maximum 30 points per test. Data were analyzed via ANOVA comparing four learning conditions across three retention intervals; additional one-way ANOVAs by interval and post hoc comparisons were conducted. Effect sizes (η²/E²) were reported.

• Descriptive performance patterns (means out of 30): • Immediate test: Repeated Studying M=16.78 (SD=7.54); Testing—no feedback M=16.35 (SD=4.83); Immediate feedback M=27.96 (SD=3.37); Delayed feedback M=25.80 (SD=4.36). • One-week delayed: Repeated Studying M=12.78 (SD=7.99); Testing—no feedback M=13.65 (SD=4.87); Immediate feedback M=24.62 (SD=5.42); Delayed feedback M=20.10 (SD=6.27). • One-month delayed: Repeated Studying M=9.82 (SD=6.39); Testing—no feedback M=12.60 (SD=4.49); Immediate feedback M=24.18 (SD=5.42); Delayed feedback M=19.85 (SD=6.98).
• ANOVAs by interval: • Immediate test: F(3,173)=57.794, p<0.001, η²≈0.50 (E²=0.50). Feedback groups > no-feedback and restudy; immediate vs delayed feedback not significant; restudy vs testing-no feedback not significant. • One-week delayed: F(3,173)=36.210, p<0.001, η²≈0.39 (E²=0.39). Immediate and delayed feedback groups > restudy and testing-no feedback (p<0.001). Immediate feedback > delayed feedback (p<0.001). Difference between restudy and testing-no feedback small and not statistically significant at authors’ threshold. • One-month delayed: F(3,176)=56.731, p<0.001, η²≈0.49 (E²=0.49). Testing-no feedback > restudy (M=12.60 vs 9.82, p<0.001). Both feedback groups > restudy and testing-no feedback (all p<0.001). Immediate feedback > delayed feedback (p<0.001).
• Overall: • Retrieval practice without feedback outperformed repeated studying after one month, evidencing a testing effect for long-term retention. • Providing feedback during practice testing substantially boosted performance at all intervals relative to no feedback and restudy. • Immediate feedback consistently yielded higher means than delayed feedback; differences were significant at one week and one month, but not in the immediate test.

The study demonstrates that retrieval practice benefits long-term retention beyond restudying and that adding feedback further amplifies this advantage. Both immediate and delayed feedback enhanced performance relative to no feedback, indicating feedback’s corrective and reinforcing roles during practice testing. Immediate feedback produced superior outcomes at delayed intervals, addressing the research questions by showing that feedback improves long-term retention and that its timing matters: immediate feedback was more beneficial than delayed feedback, especially at one week and one month. The findings are interpreted via the bifurcation model, whereby testing strengthens retrieved items, while feedback strengthens nonretrieved items, preventing bifurcation and broadening learning gains. The cognitive processing window account also supports the immediate feedback advantage, positing that learners can better compare their responses with corrective information shortly after responding, leading to deeper processing and consolidation. Methodological differences from prior mixed findings (e.g., short-answer vs multiple-choice, 10-s per-item feedback duration, provision of both immediate and delayed feedback) may explain why immediate feedback prevailed here. The results underscore the educational value of integrating timely, item-level feedback with retrieval practice to foster robust, durable learning in EFL contexts.

This laboratory study with Saudi EFL undergraduates showed that feedback during retrieval practice markedly improves retention of prose passage content compared to no feedback and restudying. Immediate feedback was most effective, producing superior performance at one-week and one-month delays relative to delayed feedback, highlighting the importance of timely corrective information for error correction and consolidation. Practical implications include providing immediate, sufficient-duration feedback after each item and enabling learners to engage with both correct and incorrect responses to maximize learning. Future work should extend these findings across learner populations, contexts, and assessment formats to refine guidelines for feedback timing in EFL instruction.

• Sample comprised only male Saudi undergraduates due to campus segregation; generalizability to female learners is unknown.
• Laboratory setting may limit ecological validity; effects should be examined in varied educational contexts.
• Outcomes were based on short-answer tests administered at set intervals; inclusion of qualitative methods (e.g., interviews, focus groups) could deepen understanding of learner experiences with feedback.
• The study focused on short-answer items; other item types (e.g., true/false, multiple-choice) warrant investigation for timing effects.