CoChat
Sign Up
Cited by 76Cognitive Abilities and TestingCognitive Functions and MemoryNeural and Behavioral Psychology Studies

The early effects of external and internal strategies on working memory updating training

Open access
Laine, Matti, Fellman, Daniel, Waris, Otto, Nyman, Thomas J.·2018·Scientific Reports
DOI 10.1038/s41598-018-22396-5PMC PMC5840432
Open this paper in CoChat — free account

Abstract

The mechanisms underlying working memory training remain unclear, but one possibility is that the typically limited transfer effects of this training reflect adoption of successful task-specific strategies. Our pre-registered randomized controlled trial (N = 116) studied the early effects of externally given vs. internally generated strategies in an updating task (n-back) over a 5-day period with a single 30-minute training session. Three groups were employed: n-back training with strategy instruction (n = 40), n-back training without strategy instruction (n = 37), and passive controls (n = 39). We found that both external and internal strategy use was associated with significantly higher posttest performance on the trained n-back task, and that training with n-back strategy instruction yielded positive transfer on untrained n-back tasks, resembling the transfer pattern typically seen after the ordinary uninstructed 4–6-week working memory training. In the uninstructed participants, the level of detail and type of internally generated n-back strategies at posttest was significantly related to their posttest n-back performance. Our results support the view that adoption of task-specific strategies plays an important role in working memory training outcomes, and that strategy-based effects are apparent right at the start of training.

AI snapshot

Written from the paper’s full text.

TL;DR

In a pre-registered RCT (N = 116), a single 30-minute n-back training session with an externally provided visuospatial strategy improved trained n-back performance and transferred to harder untrained n-back tasks, whereas uninstructed training did not differ from controls. In controls, self-reported internally generated n-back strategies at posttest were strongly related to n-back performance, supporting strategy mediation over capacity expansion in early working memory training.

Key findings

  • The strategy training group reached a higher n-back level already on the fourth training block than the active control group, t(75) = 3.98, p < 0.001.
  • On the trained digit n-back task, the strategy group outperformed both controls: maximum level ANCOVA group effect F(2, 112) = 17.09, p < 0.001; strategy vs active controls t(75) = −4.18, p < 0.001, d = −0.95; strategy vs passive controls t(77) = −5.60, p < 0.001, d = −1.26. Average level showed the same pattern, F(2, 112) = 12.58, p < 0.001.
  • Transfer appeared on the more demanding untrained 3-back tasks but not on 2-back or task-general near-transfer tasks. For 3-back letters, group effect F(2, 111) = 12.83, p < 0.001; strategy vs active controls t(75) = −4.52, p < 0.001, d = −1.03; strategy vs passive controls t(76) = −4.20, p < 0.001, d = −0.95. For 3-back colors, group effect F(2, 110) = 9.85, p < 0.001; strategy vs active controls t(73) = −3.31, p = 0.005, d = −0.76; strategy vs passive controls t(76) = −4.20, p < 0.001, d = −0.95.
  • In controls, the level of detail of self-generated n-back strategies at posttest explained 45.6% of the variance in the n-back composite score, R² = 0.456, β = 0.675, t(75) = 7.83, p < 0.001. Strategy type also mattered: on the trained digit n-back, Updating, Grouping, Grouping and comparison, and Other strategies outperformed No strategy, while Rehearsal did not.

Methods

  • Pre-registered randomized controlled trial with 116 Finnish-speaking monolingual adults aged 18–40, assigned to n-back training with strategy instruction (n = 40), n-back training without strategy instruction (n = 37), or passive controls (n = 39).
  • Study conducted within five weekdays: pretest Monday, one approximately 30-minute adaptive digit n-back training session on Tuesday–Thursday, posttest Friday. The strategy group received a novel visuospatial n-back strategy before training.
  • Main analyses used ANCOVAs with posttest performance as dependent variable, pretest performance as covariate, and group as between-subjects factor, followed by planned orthogonal contrasts and Benjamini-Hochberg correction. Strategy-report analyses used regression, ANOVAs, and inter-rater reliability scoring.

Limitations

  • The authors state that due to the brevity of training, they cannot draw conclusions about strategy effects in longer-term WM training; the lack of pre-post difference between active and passive controls is most likely due to the single 30-minute session.
  • The authors note that strategy could be cause or effect: they discuss the reverse Strategy-As-Effect hypothesis and suggest both hypotheses may apply to different subgroups, so causal direction is not fully resolved.
  • Evident from the text: strategy use was assessed retrospectively by self-report, and the sample was restricted to Finnish-speaking monolingual adults aged 18–40, limiting generalization to other populations and training durations.

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

Integrity signals

Registry

Crossref · 10.1038/s41598-018-22396-5

Cites

30 of 30 references verified0 citation errors · 0 not checkable

Statistics

38 tests recomputed · 0 with the reported and recomputed p on opposite sides of .05

Claims the paper makes

0 of 39 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    In the untrained 3-back task with colors, there was a significant main effect of group on d-prime.

    “ANCOVA on d-prime did not show a significant main effect of group for the 2-back, F , but the group difference on the more demanding 3-back task was highly significant, F(2, 110) = 9.85, p < 0.001, = 0.15.”
    Quote in the paperSubstantial claim
  • Findingown data

    In the trained digit n-back task, there was a significant effect of group on improvement expectations.

    “In the trained digit n-back task there was, however, a significant effect of group, F(2, 113) = 4.62, p = 0.012.”
    Quote in the paperNot a substantial claim
  • Findingown data

    On the trained digit n-back task, participants using an Updating strategy performed significantly better than those not reporting any strategy, t(19) = 2.24, p = 0.003, d = 1.19.

    “Participants using either Updating, t(19) = 2.24, p = 0.003, d = 1.19, Grouping, t(26) = 3.66, p < 0.001, d = 1.49, Grouping and comparison, t(28) = 5.08, p < 0.001, d = 1.64, or Other strategy, t(21) = 2.49, p < 0.001, d = 1.19, performed significantly better as compared to those not reporting any strategies.”
    Quote in the paperSubstantial claim
  • Findingown data

    In the untrained 2-back task with letters, there was no significant main effect of group on d-prime.

    “ANCOVA on d-prime in the 2-back task showed no significant main effect of group, F(2, 111) = 2.43, p = 0.195, = 0.04”
    Quote in the paperNot a substantial claim
  • Findingown data

    In the untrained n-back tasks with letters, there were no significant main effects of group on RT in either the 2-back or 3-back task.

    “ANCOVAs on the RTs showed no main effect of group in either the 2-back task, F(2, 111) = 1.02, p = 0.545, = 0.02, or the 3-back task, F(2, 112) = 3.07, p = 0.124, = 0.05.”
    Quote in the paperNot a substantial claim
  • May combine two claims. 12
  • Classified as probably not a substantial claim. 15

  • May combine two claims.

    Findingown data

    In the untrained n-back tasks with colors, there were no statistically significant main effects of group on RT in either the 2-back or 3-back task.

    “Regarding the RT analyses, ANCOVA showed no statistically significant main effect of group in the 2-back task, F , or in the 3-back task, F(2, 110) = 3.99, p = 0.060, = 0.07.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Both the level of detail and the type of self-reported strategy (especially those similar to the externally provided one) were related to higher performance on all three n-back tasks at posttest.

    “both the level of detail and type of strategy (especially those similar to the externally provided one) was related to a higher performance on all three n-back tasks at posttest.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductown data

    The three groups did not differ at baseline on age, education, gender distribution, or pretest n-back performance.

    “Statistical comparisons by one-way ANOVAs and a Chi-square indicated that the three groups did not differ on age, education, gender distribution, or pretest n-back performance (Table 1).”
    Quote in the paperSubstantial claim

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

    Premisecitation

    Carretti and colleagues reported that imagery strategy instruction for word lists improved performance on a subsequent word list recall task and on a WM task requiring semantic categorization.

    “Carretti and colleagues reported that imagery strategy instruction for word lists improved performance on a subsequent word list recall task, and also on a WM task requiring semantic categorization of to-be-remembered words12.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [12] Carretti, B., Borella, E. & De Beni, R. Does strategic memory training improve the working memory performance of younger and older adults? Exp. Psychol. 54, 311–320 (2007). source

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    Training with n-back strategy instruction yielded positive transfer on untrained n-back tasks, resembling the transfer pattern typically seen after ordinary uninstructed 4-6-week working memory training.

    “training with n-back strategy instruction yielded positive transfer on untrained n-back tasks, resembling the transfer pattern typically seen after the ordinary uninstructed 4–6-week working memory training.”
    Quote in the paperSubstantial claim

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

    Premisecitation

    Turley-Ames and Whitfield found that for low WM span individuals, strategy instruction, especially rehearsal, improved complex span performance.

    “Turley-Ames and Whitfield11 found that for low WM span individuals, strategy instruction (especially a rehearsal strategy that was contrasted with an imagery strategy and a semantic strategy) improved complex span performance.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [11] Turley-Ames, K. J. & Whitfield, M. M. Strategy training and working memory task performance. J. Mem. Lang. 49, 446–468 (2003). source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Bailey and colleagues reported performance improvement on their complex span task after strategy practice with word list learning, but not on paired-associate cued recall or a memory-based shape selection task.

    “Bailey and colleagues13 reported performance improvement on their complex span task after strategy practice with word list learning, albeit paired-associate cued recall and a memory-based shape selection task did not show training benefits.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [13] Bailey, H. R., Dunlosky, J. & Hertzog, C. Does strategy training reduce age-related deficits in working memory? Gerontology 60, 346–356 (2014). source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    There is evidence for significant performance improvements on WM tasks already within the first two test sessions and even within a single session.

    “We employed short (single-session) training as there is evidence for significant performance improvements on WM tasks already within the first two test sessions and even within a single session17.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [17] Soveri, A. et al. Test-retest reliability of five frequently used executive tasks in healthy adults. Appl. Neuropsychol. Adult 25, 155-16 (2018). source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Dunning and Holmes found that in an adaptive WM training RCT, self-reported grouping strategy use at posttest increased for the adaptive training group compared to active and passive controls on all but one test, and increases in performance gains and grouping strategy use co-occurred on two tasks sharing structure and materials with Cogmed.

    “On all but one test, a significant increase in the self-reported use of a grouping strategy at posttest was observed for the adaptive training group as compared to active and passive controls. Increases in performance gains and grouping strategy use co-occurred on two of the four pre-post tasks. These two tasks shared both the task structure and stimulus materials with the Cogmed program used for”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [9] Dunning, D. L. & Holmes, J. Does working memory training promote the use of strategies on untrained working memory tasks? Mem. Cogn. 42, 854–862 (2014). source

      Not checked in the free run.

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

    Premisecitation

    Earlier studies with WM strategy intervention have not explicitly reported results from power analyses.

    “Earlier studies with WM strategy intervention have not explicitly reported results from power analyses10–13.”
    Quote in the paperNot a substantial claim0 of 4 sources checked
    • cites [10] McNamara, D. S. & Scott, J. L. Working memory capacity and strategy use. Mem. Cogn. 29, 10–17 (2001). source

      Not checked in the free run.

    • cites [11] Turley-Ames, K. J. & Whitfield, M. M. Strategy training and working memory task performance. J. Mem. Lang. 49, 446–468 (2003). source

      Not checked in the free run.

    • cites [12] Carretti, B., Borella, E. & De Beni, R. Does strategic memory training improve the working memory performance of younger and older adults? Exp. Psychol. 54, 311–320 (2007). source

      Not checked in the free run.

    • cites [13] Bailey, H. R., Dunlosky, J. & Hertzog, C. Does strategy training reduce age-related deficits in working memory? Gerontology 60, 346–356 (2014). source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    The final sample consisted of 116 participants (strategy training group n = 40, active control group n = 37, passive control group n = 39).

    “our final sample consisted of 116 participants (strategy training group n = 40, active control group n = 37, passive control group n = 39).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Study design and conductown data

    For ANCOVA analyses, the power was 0.97 for detecting a large effect size (f = 0.40) and 0.66 for a medium effect size (f = 0.25).

    “For ANCOVA analyses, the power was 0.97 for detecting a large effect size (f = 0.40) and 0.66 for a medium effect size (f = 0.25).”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    Strategies help to make the most out of the available WM resources.

    “According to this hypothesis, strategies help to make the most out of the available WM resources10.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [10] McNamara, D. S. & Scott, J. L. Working memory capacity and strategy use. Mem. Cogn. 29, 10–17 (2001). source

      Not checked in the free run.

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

    Study design and conductown data

    Two participants were excluded for being multivariate outliers at pretest (Mahalanobis distance value χ2(14, 118) = 36.12, p = 0.001).

    “two participants (active control group n = 1, passive control group n = 1) were excluded for being multivariate outliers at pretest (Mahalanobis distance value χ2 (14, 118) = 36.12, p = 0.001)24.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [24] Tabachnick, B. G. & Fidell, L. S. Using multivariate statistics (5th ed.) (Pearson/Allyn & Bacon, 2007).

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    Participants were not aware of the existence of two different training groups (the strategy group and the active control group) prior to randomization.

    “they were not aware of the existence of two different training groups (i.e., the strategy group and the active control group).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationcitation

    The present results using the n-back paradigm align well with previous findings that explicit WM strategy instruction can be effective and that self-generated WM strategies are associated with higher task performance.

    “The present results using the n-back paradigm align well with previous findings mainly from complex span studies that have indicated that explicit WM strategy instruction can be effective10–13, and that self-generated WM strategies are associated with higher task performance9,10,14–16.”
    Quote in the paperNot a substantial claim0 of 8 sources checked
    • cites [10] McNamara, D. S. & Scott, J. L. Working memory capacity and strategy use. Mem. Cogn. 29, 10–17 (2001). source

      Not checked in the free run.

    • cites [11] Turley-Ames, K. J. & Whitfield, M. M. Strategy training and working memory task performance. J. Mem. Lang. 49, 446–468 (2003). source

      Not checked in the free run.

    • cites [12] Carretti, B., Borella, E. & De Beni, R. Does strategic memory training improve the working memory performance of younger and older adults? Exp. Psychol. 54, 311–320 (2007). source

      Not checked in the free run.

    • cites [13] Bailey, H. R., Dunlosky, J. & Hertzog, C. Does strategy training reduce age-related deficits in working memory? Gerontology 60, 346–356 (2014). source

      Not checked in the free run.

    • cites [9] Dunning, D. L. & Holmes, J. Does working memory training promote the use of strategies on untrained working memory tasks? Mem. Cogn. 42, 854–862 (2014). source

      Not checked in the free run.

    • cites [14] Bailey, H., Dunlosky, J. & Kane, M. J. Why does working memory span predict complex cognition? Testing the strategy affordance hypothesis. Mem. Cogn. 36, 1383–1390 (2008). source

      Not checked in the free run.

    • cites [15] Dunlosky, J. & Kane, M. J. The contributions of strategy use to working memory span: A comparison of strategy assessment methods. Quart. J. Exp. Psychol. 60, 1227–1245 (2007). source

      Not checked in the free run.

    • cites [16] Kaakinen, J. K. & Hyönä, J. Strategy use in the reading span test: An analysis of eye movements and reported encoding strategies. Memory 15, 634–646 (2007). source

      Not checked in the free run.

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

    Study design and conductown data

    One participant was excluded because of completing the posttest a week too late.

    “one participant (passive control group n = 1) was excluded because of completing the posttest a week too late.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    Both the Strategy Mediation and Strategy-As-Effect hypotheses speak for the important role of strategies in WM performance.

    “both hypotheses speak for the important role of strategies in WM performance.”
    Quote in the paperSubstantial claim

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

    Study design and conductown data

    Five participants withdrew after randomization.

    “Five participants (strategy training group n = 1, active control group n = 3, passive control group n = 1) withdrew after the randomization.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    Multiple comparisons were corrected using Benjamini-Hochberg adjusted p-values for group comparisons on each pre-post outcome measure.

    “To correct for multiple comparisons, Benjamini-Hochberg adjusted p-values were applied for group comparisons on each pre-post outcome measure27.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [27] Benjamini, Y. & Hochberg, Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. Royal. Stat. Soc. 57, 289–300 (1995). source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Study design and conductown data

    Participants were randomly assigned to three groups using block randomization.

    “underwent block randomization into three groups.”
    Quote in the paperSubstantial claim

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

    Study design and conductown data

    One participant was excluded because of missing training data.

    “One participant (strategy training group n = 1) was excluded because of missing training data (this participant had trained on a tablet computer which our platform does not support)”
    Quote in the paperNot a substantial claim

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

    Study design and conductown data

    Two participants were excluded because they had participated in previous cognitive training studies administered by the lab.

    “Two participants (active control group n = 2) had participated in previous cognitive training studies administered by our lab, and were thus excluded.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    The present results are in line with the Strategy Mediation hypothesis rather than the Capacity hypothesis.

    “This is in line with the Strategy Mediation hypothesis (strategy employment can make the use of limited WM capacity more effective) rather than the Capacity hypothesis (WM training expands WM capacity).”
    Quote in the paperSubstantial claim

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

    Study design and conductown data

    The study protocol was preregistered prior to data collection.

    “Prior to the data collection, the study protocol was preregistered at AsPredicted.org”
    Quote in the paperNot a substantial claim

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

    Study design and conductreasoning

    Planned contrasts were used instead of post hoc tests as stated in the preregistration.

    “Note that unlike in the pre-registration where we stated that post hoc tests will be used, we instead chose to employ planned contrasts.”
    Quote in the paperNot a substantial claim
26 of 26

  • The paper relies on this without support.

    Self-reported strategy use accurately reflects participants' actual use of n-back strategies.

    “the use of internally generated strategies as measured by self-reports has been found to be related to better complex span performance in several studies9,10,14–16.”

  • The paper relies on this without support.

    The level of detail in participants' strategy descriptions reflects the quality or effectiveness of their strategy use.

    “The scoring procedure for the self-reported freely described n-back strategies was based on the level of detail in the strategy descriptions for each of the three n-back tasks at posttest.”

  • The paper relies on this without support.

    The higher improvement expectations in the strategy training group did not account for the observed training or transfer effects.

    “The expectations of the strategy training group were higher than those of the active control group, t(75) = 2.91, p = 0.005.”

  • The paper relies on this without support.

    The active and passive control groups can be combined for analyses of self-generated strategies.

    “the final sample of controls used here included 75 participants”

  • The paper relies on this without support.

    The externally provided visuospatial strategy is not relevant to the other, structurally different WM tasks.

    “but not on the other, structurally different WM tasks for which the externally provided visuospatial strategy would not be relevant.”

  • The paper relies on this without support.

    Ordinary adaptive WM training does not entail any strategy instruction.

    “While the ordinary adaptive WM training does not entail any strategy instruction”

  • The paper relies on this without support.

    More demanding tasks are typically more sensitive to changes.

    “As more demanding tasks are typically more sensitive to changes”
7 of 7

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?

  • Both training interventions, no matter how specific they have been designed, can trigger performance improvements which are not due to an increased WM capacity, but most likely due to non-specific cognitive training effects, such as improved attentional processing, or due to the acquisition of strategies such as chunking or visualization.

    “This implies that both training interventions, no matter how specific they have been designed, can trigger performance improvements which are not due to an increased WM capacity, but most likely due to non-specific cognitive training effects, such as improved attentional processing [19], or due to the acquisition of strategies such as chunking or visualization [80].”

    From Adaptive working memory training does not produce transfer effects in cognition and neuroimaging (2022) · cites this paper as [80]

  • Laine et al. only found transfer to structurally similar untrained tasks, not to structurally different tasks.

    “For example, Laine et al. [40] only found transfer to structurally similar untrained tasks (to which the same strategy could be applied), not to structurally different tasks.”

    From Nearest transfer effects of working memory training: A comparison of two programs focused on working memory updating (2019) · cites this paper as [40]

  • Laine et al. [40] demonstrated that teaching participants a strategy improved performance in trained n-back with digits and in structurally similar untrained n-back with letters and colors, but not in structurally dissimilar tasks such as running memory span.

    “Laine et al. [40] demonstrated that teaching participants a strategy improved their performance in trained tasks (n-back with digits), and in other structurally similar untrained tasks (n-back with letters and colors), but not in other, structurally dissimilar tasks (e.g., running memory span).”

    From Nearest transfer effects of working memory training: A comparison of two programs focused on working memory updating (2019) · cites this paper as [40]

  • Laine et al. [40] showed nearest transfer between very similar tasks when trained and transfer n-back tasks varied only on content (letters and colors instead of numbers).

    “Laine et al. [40] showed nearest transfer between very similar tasks, when they varied their trained and transfer n-back tasks only on content (letters and colors instead of numbers).”

    From Nearest transfer effects of working memory training: A comparison of two programs focused on working memory updating (2019) · cites this paper as [40]

  • The pattern of transfer to structurally similar tasks in Laine et al. [40] is comparable with the one seen here for arithmetical updating training, but not with our n-back training findings, since there was no evidence of transfer to the structurally similar task in the latter case.

    “The pattern of transfer to structurally similar tasks in the Laine et al. study [40] is comparable with the one seen here for arithmetical updating training, but not with our findings with the n-back training, since there was no evidence of transfer to the structurally similar task in the latter case.”

    From Nearest transfer effects of working memory training: A comparison of two programs focused on working memory updating (2019) · cites this paper as [40]

Statistics that don’t add up

Each reported p-value recomputed from its test statistic and degrees of freedom.

  • Inconsistent p-value: t(75) = −1.93, p = .056

    Minor

    The paper reports p = .056 here; recomputing from the test statistic and its degrees of freedom gives p = .057 (two-tailed). The check cannot tell whether the statistic or the p-value is the misprint.

    Reported p = .056 · recomputed p = .057

  • Inconsistent p-value: t(21) = 2.49, p < .001

    Minor

    The paper reports p < .001 here; recomputing from the test statistic and its degrees of freedom gives p = .021 (two-tailed). The check cannot tell whether the statistic or the p-value is the misprint.

    Reported p < .001 · recomputed p = .021

  • Inconsistent p-value: t(19) = 2.24, p = .003

    Minor

    The paper reports p = .003 here; recomputing from the test statistic and its degrees of freedom gives p = .037 (two-tailed). The check cannot tell whether the statistic or the p-value is the misprint.

    Reported p = .003 · recomputed p = .037

  • Inconsistent p-value: t(77) = −2.99, p = .003

    Minor

    The paper reports p = .003 here; recomputing from the test statistic and its degrees of freedom gives p = .004 (two-tailed). The check cannot tell whether the statistic or the p-value is the misprint.

    Reported p = .003 · recomputed p = .004

  • Inconsistent p-value: t(27) = 1.71, p = .092

    Minor

    The paper reports p = .092 here; recomputing from the test statistic and its degrees of freedom gives p = .099 (two-tailed). The check cannot tell whether the statistic or the p-value is the misprint.

    Reported p = .092 · recomputed p = .099

  • Inconsistent p-value: t(28) = 1.95, p = .056

    Minor

    The paper reports p = .056 here; recomputing from the test statistic and its degrees of freedom gives p = .061 (two-tailed). The check cannot tell whether the statistic or the p-value is the misprint.

    Reported p = .056 · recomputed p = .061

  • p-value fits an unstated one-tailed test: t(26) = 3.66, p < .001

    Minor

    The paper reports p < .001 here; recomputing from the test statistic and its degrees of freedom gives p = .001 (two-tailed). The reported value is what a one-tailed test gives, but the paper does not say near this result that a one-tailed test was used.

    Reported p < .001 · recomputed p = .001

  • Inconsistent p-value: F(5, 69) = 4.40, p < .001

    Minor

    The paper reports p < .001 here; recomputing from the test statistic and its degrees of freedom gives p = .002. The check cannot tell whether the statistic or the p-value is the misprint.

    Reported p < .001 · recomputed p = .002

  • Inconsistent p-value: t(27) = 2.55, p = .013

    Minor

    The paper reports p = .013 here; recomputing from the test statistic and its degrees of freedom gives p = .017 (two-tailed). The check cannot tell whether the statistic or the p-value is the misprint.

    Reported p = .013 · recomputed p = .017

9 of 9

Statistics recomputed

Every reported test result, recomputed from its test statistic and degrees of freedom. A result shows as not adding up only once an independent check has confirmed it.

  • Doesn’t add upt(75) = -1.93reported p = .056 · recomputed p = .057

    “t(75) = −1.93, p = .056”

  • Doesn’t add upt(77) = -2.99reported p = .003 · recomputed p = .004

    “t(77) = −2.99, p = .003”

  • Doesn’t add upt(19) = 2.24reported p = .003 · recomputed p = .037

    “t(19) = 2.24, p = .003”

  • Doesn’t add upt(26) = 3.66reported p < .001 · recomputed p = .001

    “t(26) = 3.66, p < .001”

  • Doesn’t add upt(21) = 2.49reported p < .001 · recomputed p = .021

    “t(21) = 2.49, p < .001”

  • Doesn’t add upF(5, 69) = 4.4reported p < .001 · recomputed p = .002

    “F(5, 69) = 4.40, p < .001”

  • Doesn’t add upt(27) = 2.55reported p = .013 · recomputed p = .017

    “t(27) = 2.55, p = .013”

  • Doesn’t add upt(27) = 1.71reported p = .092 · recomputed p = .099

    “t(27) = 1.71, p = .092”

  • Doesn’t add upt(28) = 1.95reported p = .056 · recomputed p = .061

    “t(28) = 1.95, p = .056”

  • Not confirmedF(2, 111) = 2.43reported p = .195 · recomputed p = .093

    “F(2, 111) = 2.43, p = .195”

  • Not confirmedF(2, 111) = 1.02reported p = .545 · recomputed p = .364

    “F(2, 111) = 1.02, p = .545”

  • Not confirmedF(2, 112) = 3.07reported p = .124 · recomputed p = .050

    “F(2, 112) = 3.07, p = .124”

  • Not confirmedF(2, 110) = 3.99reported p = .06 · recomputed p = .021

    “F(2, 110) = 3.99, p = .060”

  • Not confirmedF(2, 112) = 1.66reported p = .348 · recomputed p = .195

    “F(2, 112) = 1.66, p = .348”

  • Consistentt(75) = 3.98reported p < .001 · recomputed p < .001

    “t(75) = 3.98, p < .001”

  • Consistentt(75) = 2.91reported p = .005 · recomputed p = .005

    “t(75) = 2.91, p = .005”

  • ConsistentF(2, 113) = 1.75reported p = .179 · recomputed p = .178

    “F(2, 113) = 1.75, p = .179”

  • ConsistentF(2, 113) = 1.73reported p = .182 · recomputed p = .182

    “F(2, 113) = 1.73, p = .182”

  • ConsistentF(2, 113) = 2.81reported p = .064 · recomputed p = .064

    “F(2, 113) = 2.81, p = .064”

  • ConsistentF(2, 113) = 4.62reported p = .012 · recomputed p = .012

    “F(2, 113) = 4.62, p = .012”

  • Consistentt(74) = -1.02reported p = .311 · recomputed p = .311

    “t(74) = −1.02, p = .311”

  • ConsistentF(2, 112) = 17.09reported p < .001 · recomputed p < .001

    “F(2, 112) = 17.09, p < .001”

  • Consistentt(75) = -4.18reported p < .001 · recomputed p < .001

    “t(75) = −4.18, p < .001”

  • Consistentt(77) = -5.6reported p < .001 · recomputed p < .001

    “t(77) = −5.60, p < .001”

  • ConsistentF(2, 112) = 12.58reported p < .001 · recomputed p < .001

    “F(2, 112) = 12.58, p < .001”

  • Consistentt(75) = -2.97reported p = .013 · recomputed p = .004

    “t(75) = −2.97, p = .013”

    The paper states a correction for multiple tests

  • Consistentt(77) = -4.97reported p < .001 · recomputed p < .001

    “t(77) = −4.97, p < .001”

  • ConsistentF(2, 111) = 12.83reported p < .001 · recomputed p < .001

    “F(2, 111) = 12.83, p < .001”

  • Consistentt(75) = -4.52reported p < .001 · recomputed p < .001

    “t(75) = −4.52, p < .001”

  • Consistentt(76) = -4.2reported p < .001 · recomputed p < .001

    “t(76) = −4.20, p < .001”

  • ConsistentF(2, 110) = 9.85reported p < .001 · recomputed p < .001

    “F(2, 110) = 9.85, p < .001”

  • Consistentt(73) = -3.31reported p = .005 · recomputed p = .001

    “t(73) = −3.31, p = .005”

    The paper states a correction for multiple tests

  • Consistentt(75) = 7.83reported p < .001 · recomputed p < .001

    “t(75) = 7.83, p < .001”

  • ConsistentF(5, 70) = 9.23reported p < .001 · recomputed p < .001

    “F(5, 70) = 9.23, p < .001”

  • Consistentt(28) = 5.08reported p < .001 · recomputed p < .001

    “t(28) = 5.08, p < .001”

  • Consistentt(34) = 5.08reported p < .001 · recomputed p < .001

    “t(34) = 5.08, p < .001”

  • ConsistentF(5, 69) = 3.5reported p = .007 · recomputed p = .007

    “F(5, 69) = 3.50, p = .007”

  • Consistentt(31) = 3.95reported p < .001 · recomputed p < .001

    “t(31) = 3.95, p < .001”

38 of 38

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.

  • Regression degrees of freedom conflict with sample size

    Minor

    The reported regression uses 75 controls and one predictor. With an intercept, its t statistic should have 73 residual degrees of freedom, not 75.

  • Strategy-detail scale is described with two different lower bounds

    Minor

    The text gives the per-task strategy-detail range as 1 to 3, but also assigns zero points when no strategy is reported. The stated total range of 0 to 9 likewise requires a per-task minimum of zero.

2 of 2

Cites

Flagged first, then by citations

1 of the 31 references couldn’t be matched to a known paper.

Research works printed but not linked: 2

  1. Tabachnick, B. G. & Fidell, L. S. Using multivariate statistics (5th ed.) (Pearson/Allyn & Bacon, 2007).
  2. IBM SPSS statistics for Windows. Version 23.0, Armonk, N.Y. (2015).

Cited by

Sorted by citations

Related

Authors
Laine, M., Fellman, D., Waris, O., & Nyman, T. J.
Affiliations
Åbo Akademi University (FI)
Corresponding author
Laine, Matti
ORCID
Venue
Scientific Reports
Publisher
Springer Science and Business Media LLC
Published
2018-03-06
Volume
8
Issue
1
Pages
4045-4045
ISSN
2045-2322 (ISSN-L)
DOI
10.1038/s41598-018-22396-5
PMC
PMC5840432
OpenAlex
W2792886356
Semantic Scholar
c63711fd5a1f93dca07b70928f584bb29bd0afe7
Cited by
71
References
31
Citations by year
2026: 7 · 2025: 8 · 2024: 12 · 2023: 10 · 2022: 8 · 2021: 9 · 2020: 8 · 2019: 8 · 2018: 1
FWCI
8.51
Type
journal article
Language
en
Open access
Licence
Locations
  • Scientific Reports (publishedVersion, cc-by)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • CentAUR (University of Reading) (acceptedVersion)
  • UTUPub (University of Turku) (submittedVersion)
  • UTUPub (University of Turku) (submittedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
  • Åbo Akademi University Research Portal (submittedVersion, cc-by)
Funders
Academy of Finland (#260276)
Topics
  • Social Sciences › Psychology › Experimental and Cognitive Psychology › Cognitive Abilities and Testing
  • Social Sciences › Psychology › Experimental and Cognitive Psychology › Cognitive Functions and Memory
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Neural and Behavioral Psychology Studies
Keywords
working memory training, n-back task, strategy instruction, transfer effects, working memory updating, strategy adoption
MeSH terms
Adult, Female, Finland, Humans, Learning, Male, Memory, Short-Term, Psychological Techniques, Treatment Outcome, Young Adult

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 2026-10-08.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-08.

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