CoChat
Sign Up
Cited by 137Memory Processes and InfluencesMemory and Neural MechanismsSecond Language Acquisition and Learning

Is expanding retrieval a superior method for learning text materials?

Open access
Karpicke, J. D., Roediger, H. L.·2010·Memory & Cognition
DOI 10.3758/mc.38.1.116
Open this paper in CoChat — free accountOpen PDF

Abstract

Expanding retrieval practice refers to the idea that gradually increasing the spacing interval between repeated tests ought to promote optimal long-term retention. Belief in the superiority of this technique is widespread, but empirical support is scarce. In addition, virtually all research on expanding retrieval has examined the learning of word pairs in paired-associate tasks. We report two experiments in which we examined the learning of text materials with expanding and equally spaced retrieval practice schedules. Subjects studied brief texts and recalled them in an initial learning phase. We manipulated the spacing of the repeated recall tests and examined final recall 1 week later. Overall we found that (1) repeated testing enhanced retention more than did taking a single test, (2) testing with feedback (restudying the passages) produced better retention than testing without feedback, but most importantly (3) there were no differences between expanding and equally spaced schedules of retrieval practice. Repeated retrieval enhanced long-term retention, but how the repeated tests were spaced did not matter.

AI snapshot

Written from the paper’s full text.

TL;DR

In two experiments with expository texts, repeated retrieval improved 1-week retention, and feedback after retrieval further boosted retention, but expanding and equally spaced retrieval schedules did not differ. The spacing schedule of repeated tests did not matter under these conditions.

Key findings

  • Expanding versus equally spaced retrieval produced no meaningful differences on final recall 1 week later. In Experiment 1, final recall was .66 for immediate-expanding, .65 for immediate-equal, .53 for delayed-expanding, and .57 for delayed-equal; in Experiment 2, no-feedback final recall was .55/.53 (immediate expanding/equal) and .45/.52 (delayed expanding/equal), and feedback final recall was .80/.78 and .84/.83.
  • Repeated testing enhanced long-term retention more than a single test. In Experiment 2 no-feedback conditions, single-test final recall was .37 (immediate) and .32 (delayed), while repeated-test conditions ranged from .45 to .55; in feedback conditions, single-test final recall was .34 and .46, while repeated-test conditions ranged from .78 to .84.
  • Feedback after retrieval improved retention. In Experiment 2, repeated-test data showed a main effect of feedback [F(1,60) = 59.65], and repeated retrieval with feedback produced a greater than 400% improvement in long-term retention relative to studying the texts once.
  • Subjects recalled texts in forward serial order rather than recalling the last ideas first: average recall-output/input-order correlations were r = .89 in Experiment 1 and r = .81 in Experiment 2, and no subject recalled the last passage ideas first.

Methods

  • Experiment 1: 80 Washington University undergraduates studied two brief passages (Bamboo and Vanuatu; 162 and 155 words; 20 idea units each) and took three free recall tests. The design crossed first-test position (immediate vs. delayed) with repeated-test spacing (expanding vs. equal): 0-2-4, 0-2-2, 2-2-4, and 2-2-2 schedules; final recall occurred 1 week later.
  • Experiment 2: 64 Washington University undergraduates studied four critical passages (Bamboo, Diving, Eugene O’Neill, Vanuatu; 20 idea units each) plus filler passages. Learning condition was within subjects (study only, single test, expanding retrieval, equally spaced retrieval), while first-test position (immediate vs. delayed) and feedback (feedback vs. no feedback) were between subjects; final recall occurred 1 week later.
  • Free recall was scored by idea units; 40 tests were double-scored in each experiment, with interrater correlations of r = .98 (Experiment 1) and r = .97 (Experiment 2). Analyses used ANOVAs comparing first-test position, spacing schedule, test number, repeated testing, and feedback.

Limitations

  • Stated: the authors note that differences between expanding and equally spaced schedules might emerge with different paradigm features, such as conditions producing rapid intertest forgetting or much longer delays between tests (perhaps days apart), but say these are empirical questions not yet sufficiently examined.
  • Stated: in Experiment 2, the authors could “only speculate” about why the immediate and delayed first-test groups did not differ on the initial test in the feedback conditions.
  • Evident: both experiments used brief expository texts, undergraduate samples, and within-session spacing intervals of minutes (total lags of 6–8 min), so the null expanding-versus-equal result may not generalize to longer intervals, other materials, or other populations.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.3758/mc.38.1.116

Cites

41 of 41 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 40 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    There are no differences between expanding and equally spaced schedules of retrieval practice for learning text materials.

    “there were no differences between expanding and equally spaced schedules of retrieval practice”
    • Also stated in: § Abstract · plain
      “how the repeated tests were spaced did not matter”
    Quote in the paperSubstantial claim
  • Findingown data

    In Experiment 1, groups with an immediate first recall test recalled more idea units during the initial learning phase than groups with a delayed first test.

    “In general the two groups with an immediate first test recalled more idea units than did the two groups with a delayed first test.”
    • Also stated in: § Initial recall. · hedged
      “There was a main effect of initial test position [F(1,76) 20.27, p 2 .21],indicating that delaying the first test produced forgetting.”
    Quote in the paperNot a substantial claim
  • Findingown data

    Testing with feedback (restudying the passages) produces better retention than testing without feedback.

    “testing with feedback (restudying the passages) produced better retention than testing without feedback”
    Quote in the paperSubstantial claim
  • Findingown data

    In Experiment 1, recall increased modestly across the three initial tests (hypermnesia).

    “There was also a main effect of test number [F(2,152) 4.43, p 2 .06], indicating that a modest amount of hypermnesia occurred across the three tests.”
    Quote in the paperNot a substantial claim
  • Findingown data

    Repeated retrieval practice enhances long-term retention more than taking a single test.

    “Overall we found that (1) repeated testing enhanced retention more than did taking a single test”
    Quote in the paperSubstantial claim
  • The quote only partly states this claim. 5
  • May combine two claims. 12
  • Classified as probably not a substantial claim. 12

  • The quote only partly states this claim.

    Findingown data

    In Experiment 1, subjects tended to recall ideas from the texts in serial order rather than recalling the last few ideas first.

    “The figure clearly shows that the subjects tended to recall the ideas from the texts in serial order.”
    • Also stated in: § EXPERIMENT 2 · boosted
      “In fact there was not a single subject who recalled the last ideas in the passages first.”
    Quote in the paperNot a substantial claim

  • The quote only partly states this claim.

    Findingown data

    In Experiment 1 final recall, expanding versus equal spacing did not affect recall.

    “but no effect of spacing schedule and no interaction (Fs 1).”
    Quote in the paperNot a substantial claim

  • The quote only partly states this claim.

    Findingown data

    In Experiment 1 final recall 1 week later, subjects who took an immediate initial recall test outperformed those who took a delayed initial test.

    “The subjects who took an immediate initial recall test outperformed those who took a delayed initial test, but the particular spacing of repeated tests did not appear to matter.”
    • Also stated in: § EXPERIMENT 2 · plain
      “There was a main effect of initial test position [F(1,76) 10.07, p 2 .12]but no effect of spacing schedule and no interaction (Fs 1).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    In Experiment 2 no-feedback initial recall, there was no main effect of learning condition and no interaction.

    “A 2 (position of the first test) 3 (learning condition) ANOVA showed that there was a main effect of the position of the first test [F(1,30) 4.69, p 2 .14]but no main effect of learning condition and no interaction (Fs 1).”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    In Experiment 2 feedback conditions, the pattern of results was somewhat different from the no-feedback conditions.

    “The pattern of results was somewhat different from the pattern in the no-feedback conditions.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    In Experiment 2 no-feedback initial recall, there was a marginally significant small degree of hypermnesia across tests.

    “There was a marginally significant main effect of test number [F(3,93) 2.32, p 2 .07,p .08], reflecting a small degree of hypermnesia across tests.”
    Quote in the paperNot a substantial claim

  • The quote only partly states this claim.

    Findingown data

    In Experiment 1, the average correlation between recall output order and input order was r = .89.

    “The overall average correlation was r .89.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    In Experiment 2, the recall order results wholly replicated the results of Experiment 1.

    “The recall order results wholly replicated the results of Experiment 1.”
    Quote in the paperNot a substantial claim

  • The quote only partly states this claim.

    Findingown data

    In Experiment 2, taking a single free recall test produced better long-term retention than did reading the text and not recalling it, even without feedback, in both immediate and delayed first test conditions.

    “First, within the no-feedback groups, both the immediate and delayed first test conditions showed significant testing effects.”
    • Also stated in: § Final recall. · plain
      “Taking a single test produced better long-term retention than did only studying the text-even when the subjects did not reread the text after recalling it (cf.Roediger & Karpicke, 2006b).”
    • Also stated in: § GENERAL DISCUSSION · plain
      “First, there was a testing effect: Taking a single free recall test produced better long-term retention than did reading the text and not recalling it—even when there was no feedback after the test.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Roediger & Karpicke, 2006b

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    In the present experiments, recall remained relatively constant or even exhibited a modest degree of hypermnesia across tests.

    “but the present data show that recall remained relatively constant or even exhibited a modest degree of hypermnesia (see too Logan & Balota, 2008).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Logan & Balota, 2008

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductnone given

    In Experiment 2, the four critical passages were presented in the same order for every subject, but the order of the four learning conditions was counterbalanced to create four orders.

    “The four critical passages were presented in the same order for every subject (Bamboo, Diving, Eugene O'Neill, Vanuatu), but the order of the four learning conditions was counterbalanced to create four orders.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    In Experiment 1, the order in which subjects studied the two passages was counterbalanced to create two counterbalancing orders within each condition.

    “The order in which the subjects studied the two passages was counterbalanced to create two counterbalancing orders within each condition.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductown data

    In Experiment 1, 40 recall tests were scored by two raters, and the Pearson correlation between their scores was r = .98; remaining tests were scored by one rater.

    “Initially, 40 recall tests were scored by two raters, and the Pearson product-moment correlation between their scores was r .98.Given the high interrater reliability, the remaining recall tests were scored by one rater.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductown data

    In Experiment 2, 40 recall tests were scored by two raters, and the Pearson correlation between their scores was r = .97; remaining tests were scored by one rater.

    “As in Experiment 1, 40 recall tests from Experiment 2 were scored by two raters.The Pearson product-moment correlation between their scores was r .97.Given the high interrater reliability, the remaining recall tests were scored by one rater.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    In Experiment 1, preliminary pilot testing showed that 4 min was enough time for subjects to reach asymptote in amount recalled.

    “Preliminary pilot testing showed that 4 min was enough time for subjects to reach asymptote in amount recalled (cf.Roediger & Thorpe, 1978).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Roediger & Thorpe, 1978

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    The authors' paired-associate learning experiments led them to conclude that expanding retrieval was not superior to equally spaced retrieval and, in fact, the opposite may be true for long-term retention.

    “In short, our paired-associate learning experiments led us to conclude that expanding retrieval was not superior to equally spaced retrieval and, in fact, the opposite may be true for long-term retention.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Karpicke & Roediger, 2007a

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Experiments that used longer delays (e.g., 24 h or more) often found no effect of expanding retrieval, or sometimes an advantage of equally spaced practice.

    “In addition, experiments that used longer delays (e.g., 24 h or more) often found no effect of expanding retrieval, or sometimes an advantage of equally spaced practice (e.g., Cull, 2000;Logan & Balota, 2008).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Cull, 2000

      Not checked in the free run.

    • cites Logan & Balota, 2008

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationreasoning

    The critical factor in paired-associate learning experiments was the position of the first retrieval attempt, not the schedule of repeated tests.

    “Our theory was that the critical factor in paired-associate learning experiments was the position of the first retrieval attempt, not the schedule of repeated tests.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Recall reflects immediate/primary memory when a word pair is recalled immediately after it was studied.

  • May combine two claims.

    Premisecitation

    Some experiments in which subjects were given feedback after retrieval attempts also found no advantage of expanding retrieval or an advantage of equally spaced practice.

    “And finally, some experiments in which subjects were given feedback after retrieval attempts also found no advantage of expanding retrieval or an advantage of equally spaced practice (for a review, see Balota, Duchek, & Logan, 2007).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Balota, Duchek, & Logan, 2007

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation

    If students practice retrieval on their own, they would likely restudy after attempting recall.

    “Of course, examining feedback is practically relevant, because if students practice retrieval on their own, they would likely restudy after attempting recall (Karpicke, Butler, & Roediger, 2009).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Karpicke, Butler, & Roediger, 2009

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Expanding and equally spaced retrieval schedules have been examined with paired-associate materials but not with texts, and the results with paired-associate tasks have been equivocal.

    “Expanding and equally spaced retrieval schedules have been examined with pairedassociate materials but not with texts-and moreover, the results with paired-associate tasks have been equivocal (see Balota et al., 2007).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Balota et al., 2007

      Not checked in the free run.

  • May combine two claims.

    Interpretationcitation

    The conclusion that spacing schedule does not matter is not consistent with Landauer and Bjork's (1978) original idea but is supported by the bulk of recent research.

    “This conclusion is not consistent with Landauer and Bjork's (1978) original idea but it is supported by the bulk of recent research.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Landauer and Bjork's (1978)

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Only a few studies have examined feedback in the form of rereading a text after students attempted recall.

    “but we know of only a few studies that examined feedback in the form of rereading a text after students attempted recall (e.g., McDaniel, Howard, & Einstein, 2009).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites McDaniel, Howard, & Einstein, 2009

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    Even a single, nominally massed test can enhance later retention, as long as recall on the test does not represent retrieval of items residing in immediate memory.

    “Even a single, nominally massed test can enhance later retention, as long as recall on the test does not represent retrieval of items residing in immediate memory.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    Increasing the difficulty of retrieval attempts may promote long-term retention.

    “Therefore it may be true that increasing the difficulty of retrieval attempts does in fact promote long-term retention (Pyc & Rawson, 2009)”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Pyc & Rawson, 2009

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Testing an individual item immediately after it was studied in a paired-associate learning task would prompt subjects to recall the item when it still resided in their immediate awareness, akin to rehearsing an item in massed fashion, a technique known to have minimal effects on later recall.

    “Testing an individual item immediately after it was studied in a pairedassociate learning task would prompt subjects to recall the item when it still resided in their immediate awareness, akin to rehearsing an item in massed fashion, a technique known to have minimal effects on later recall (e.g., Craik & Watkins, 1973).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Craik & Watkins, 1973

      Not checked in the free run.

26 of 26

  • The authors state this as an assumption.

    Retrieval difficulty produces a positive effect on long-term retention.

    “Retrieval difficulty itself is assumed to produce a positive effect on long-term retention (see Gardiner, Craik, & Bleasdale, 1973;Pyc & Rawson, 2009).”

  • The paper relies on this without support.

    Recall reflects immediate/primary memory when a word pair is recalled immediately after it was studied.

    “It is clear that recall reflects immediate/ primary memory when a word pair is recalled immediately after it was studied.”

  • The authors state this as an assumption.

    Recall latency is a viable index of retrieval difficulty.

    “Assuming that recall latency is a viable index of retrieval difficulty (Benjamin, Bjork, & Schwartz, 1998;Koriat & Ma'ayan, 2005)”
3 of 3

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • For almost all learning paradigms, the training intervals are fixed, although in one type of spaced training paradigm, the intervals between sessions progressively lengthen.

    “For almost all learning paradigms, the training intervals are fixed, although in one type of spaced training paradigm, the intervals between sessions progressively lengthen2,84.”

    From The right time to learn: Mechanisms and optimization of spaced learning (2016) · cites this paper as [84]

  • A meta-analysis and a text learning study found no substantial evidence for the superiority of progressively lengthening intervals in promoting long-term memory formation.

    “However, a meta-analysis2 and a text learning study84 found no substantial evidence for the superiority of this approach in terms of promoting long-term memory formation.”

    From The right time to learn: Mechanisms and optimization of spaced learning (2016) · cites this paper as [84]

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Experiment 1 spacing interval is assigned to the wrong experiment

    Minor

    The text assigns a six-minute lag to Experiment 1, but Experiment 1 used four-minute periods. Two such periods give an eight-minute lag; the six-minute lag corresponds to Experiment 2's three-minute periods.

  • Inconsistent denominator degrees of freedom in the same ANOVA

    Minor

    The reported feedback main effect and feedback-by-test-position interaction come from the same two-factor ANOVA on the single-test data, but their denominator degrees of freedom are given as 30 and 60, respectively. For this between-subjects analysis, they should use the same denominator degrees of freedom.

2 of 2

Cites

Flagged first, then by citations

3 of the 44 references couldn’t be matched to a known paper.

Research works printed but not linked: 2

  1. Landauer, T. K., & Bjork, R. A. (1978). Optimum rehearsal patterns and name learning. In M. Gruneberg, P. E. Morris, & R. N. Sykes (Eds.), Practical aspects of memory (pp. 625-632). London: Aca- demic Press.
  2. Tulving, E. (1968). Theoretical issues in free recall. In T. Dixon & D. Horton (Eds.), Verbal behavior and general behavior theory (pp. 2-36). Englewood Cliffs, NJ: Prentice Hall.

Cited by

Sorted by citations

Related

Authors
Karpicke, J. D., & Roediger, H. L.
Affiliations
Purdue University West Lafayette (US)
Corresponding author
Karpicke, J. D.
ORCID
Karpicke, J. D.
Venue
Memory & Cognition
Publisher
Springer Science and Business Media LLC
Published
2009-12-04
Volume
38
Issue
1
Pages
116-124
ISSN
  • 0090-502X (ISSN-L)
  • 1532-5946
DOI
10.3758/mc.38.1.116
OpenAlex
W2116385289
Semantic Scholar
29cd968699b66b38a019b6f1058038b62ff21fa1
Cited by
137
References
44
Citations by year
2026: 2 · 2025: 1 · 2024: 5 · 2023: 7 · 2022: 4 · 2021: 8 · 2020: 14 · 2019: 6 · 2018: 16 · 2017: 8 · 2016: 9 · 2015: 12 · 2014: 10 · 2013: 12 · 2012: 10
FWCI
3.71
Type
journal article
Language
en
Open access
Licence
www.springer.com/tdm (tdm)
Locations
  • Memory & Cognition (publishedVersion)
  • PubMed (publishedVersion)
Funders
Purdue University
Topics
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Memory Processes and Influences
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Memory and Neural Mechanisms
  • Social Sciences › Psychology › Developmental and Educational Psychology › Second Language Acquisition and Learning
Keywords
retrieval practice, spacing effect, long-term retention, text recall, retesting, restudy, feedback
MeSH terms
Humans, Practice, Psychological, Reading, Mental Recall, Retention, Psychology, Verbal Learning, Recognition, Psychology, Feedback, Psychological

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-10.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-10.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review