CoChat
Sign Up
Cited by 127Cognitive Abilities and TestingDevelopmental and Educational NeuropsychologyCognitive and developmental aspects of mathematical skills

The cognitive and academic benefits of Cogmed: A meta-analysis

Open access
Aksayli, N. Deniz, Sala, Giovanni, Gobet, Fernand·2019·Educational Research Review
DOI 10.1016/j.edurev.2019.04.003

Published version of a preprint

Open this paper in CoChat — free accountOpen PDF

Abstract

Cogmed Working Memory Training (CWMT) is a commercial cognitive-training program designed to foster working-memory capacity. Enhanced working-memory capacity is then supposed to increase one's overall cognitive function and academic achievement. This meta-analysis investigates the effects of CWMT on cognitive and academic outcomes. The inclusion criteria were met by 50 studies (637 effect sizes). Highly consistent near-zero effects were estimated in far-transfer measures of cognitive ability (e.g., attention and intelligence) and academic achievement (language ability and mathematics). By contrast, slightly heterogeneous small to medium effects were observed in memory tasks (i.e., near transfer). Moderator analysis showed that these effects were weaker for near-transfer measures not directly related to the trained tasks. These results highlight that, while near transfer occurs regularly, far transfer is rare or, possibly, inexistent. Transfer thus appears to be a function of the degree of overlap between trained tasks and outcome tasks.

AI snapshot

Written from the paper’s full text.

TL;DR

This meta-analysis of 50 studies (637 effect sizes; 3,059 participants) finds that Cogmed Working Memory Training produces near-zero far-transfer effects on cognitive ability and academic achievement, while producing small-to-medium near-transfer effects on memory tasks. The authors conclude that transfer depends on overlap between trained and outcome tasks, and that CWMT cannot be recommended as an educational tool.

Key findings

  • Far transfer was consistently near zero: immediate post-test \( \bar{g} = 0.048 \), 95% CI [-0.017; 0.113], \( p = .135 \); follow-up \( \bar{g} = 0.051 \), 95% CI [-0.018; 0.119], \( p = .131 \).
  • Near transfer was significant: immediate post-test \( \bar{g} = 0.444 \), 95% CI [0.337; 0.551], \( p < .001 \); follow-up \( \bar{g} = 0.365 \), 95% CI [0.242; 0.487], \( p < .001 \).
  • Effects were larger for very-near-transfer memory tasks (\( \bar{g} = 0.566 \) immediate; \( \bar{g} = 0.487 \) follow-up) than for lesser-near-transfer tasks (\( \bar{g} = 0.246 \) immediate; \( \bar{g} = 0.176 \) follow-up).
  • Baseline differences and task similarity (Criterion) were the main moderators; far-transfer models showed no true heterogeneity after controlling for baseline differences.

Methods

  • Systematic search following PRISMA across MEDLINE, PsycINFO, Science Direct, and ProQuest Dissertation & Theses, using Boolean terms for Cogmed/working-memory/cognitive training; 2,761 records retrieved, yielding 50 studies from 2005–2017 that met five inclusion criteria.
  • Effect sizes were Hedges’s \( g \), analyzed with robust variance estimation (RVE) with hierarchical weights and small-sample corrections in Robumeta; near transfer was defined as memory tasks (Gsm) and far transfer as all other cognitive/academic measures.
  • Moderators included allocation, control-group type, baseline difference, age, population, measure type, and criterion (very near vs. lesser near transfer); sensitivity analyses used influential-case diagnostics and publication-bias analyses (funnel plots, trim-and-fill, PET-PEESE).

Limitations

  • The paper states that the sampling-error variance formula used is an approximation because pre-post-test correlations were rarely reported in primary studies (only Honoré & Noël, 2017, reported them).
  • The paper notes that follow-up analyses were based on a subsample of studies, and post-test-to-follow-up attrition may affect comparisons.
  • The paper states that future studies should test effects on latent factors rather than observed variables; evident from the text, the meta-analysis is limited to CWMT and cannot directly establish effects of other commercial training programs.

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

Integrity signals

Registry

Crossref · 10.1016/j.edurev.2019.04.003

Cites

75 of 76 references verified1 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

    Baseline and Criterion were significant moderators of near-transfer immediate post-test effects (Baseline b = -0.487, p = .002; Criterion b = 0.330, p < .001).

    “The significant moderators were Baseline (b = -0.487,SE = 0.137, p = .002)and Criterion (b = 0.330, SE = 0.054, p < .001).”
    Quote in the paperSubstantial claim
  • Findingown data

    Slightly heterogeneous small to medium effects were observed in memory tasks (i.e., near transfer).

    “By contrast, slightly heterogeneous small to medium effects were observed in memory tasks (i.e., near transfer).”
    Quote in the paperSubstantial claim
  • Findingown data

    Moderator analysis showed that near-transfer effects were weaker for measures not directly related to the trained tasks.

    “Moderator analysis showed that these effects were weaker for near-transfer measures not directly related to the trained tasks.”
    Quote in the paperSubstantial claim
  • Findingown data

    The overall lesser-near-transfer immediate post-test effect size was g = 0.246, SE = 0.069, 95% CI [0.102; 0.391], p = .002.

    “The overall effect size was g ¯ = 0.246, SE = 0.069, 95% CI [0.102; 0.391], m = 33, k = 93, df = 19.49,p = .002,ω 2 = 0.000, τ 2 = 0.091.”
    Quote in the paperSubstantial claim
  • Findingown data

    The degree of overlap between CWMT training tasks and memory tasks was a significant moderator.

    “The degree of overlap between the training tasks of CWMT and the memory tasks was a significant moderator.”
    Quote in the paperSubstantial claim
  • May combine two claims. 37
  • Classified as probably not a substantial claim. 20
  • The quote only partly states this claim. 3

  • May combine two claims.

    Findingown data

    Baseline and Age were significant moderators of lesser-near-transfer effects (Baseline b = -0.598, p = .006; Age children b = 0.318, p = .018).

    “Baseline was a significant moderator (b = -0.598,SE = 0.169, p = .006).Also, Age was a significant moderator, with the effect of the training significantly higher in the children than in the adults and older adults (b = 0.318, SE = 0.117, p = .018).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    In total, 50 studies conducted from 2005 to 2017 met all inclusion criteria, including 637 effect sizes and 3059 participants.

    “In total, we found 50 studies, conducted from 2005 to 2017, that met all the inclusion criteria.These studies included 637 effect sizes and a total of 3059 participants.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    No appreciable effect of CWMT was found on far-transfer tasks; no estimate was significantly different from zero.

    “While the training regimen increased the performance on memory tasks, no appreciable effect was found in far-transfer tasks (no estimate significantly different from zero).”
    Quote in the paperSubstantial claim

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

    Findingown data

    In the lesser-near-transfer follow-up analysis, one influential case (g=-0.235, ID=551) was found.

    “We found one influential case (g = -0.235,ID = 551).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    For lesser-near-transfer measures at follow-up, the PET estimate was g=0.089 (SE=0.078, p=.270) and the PEESE estimate was g=0.154 (SE=0.044, p=.004).

    “The PET estimate was g ¯= 0.089 (SE = 0.078, p = .270),and the PEESE estimate was g ¯= 0.154 (SE = 0.044, p = .004).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Trim-and-fill estimates for lesser-near-transfer measures at follow-up were g=0.189 (SE=0.047, p<.001), g=0.149 (SE=0.043, p<.001), and g=0.178 (SE=0.046, p<.001) with L0, R0, and Q0 estimators, respectively.

    “The trim-and-fill estimates were g ¯= 0.189 (SE = 0.047, p < .001),g ¯= 0.149 (SE = 0.043, p < .001),and g ¯= 0.178 (SE = 0.046, p < .001)with the L0, R0, and Q0 estimators, respectively.”
    Quote in the paperSubstantial claim

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

    Findingown data

    In the very-near-transfer follow-up analysis, one influential case (g=1.998, ID=99) and one outlier (g=3.062, ID=553) were identified and excluded.

    “We found one influential case (g = 1.998,ID = 99) and another effect was excluded because, like in the previous model, it was a considered an outlier (g = 3.062, ID = 553).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    For lesser-near-transfer measures at follow-up, no moderator was significant because no true heterogeneity was observed.

    “Again, since no true heterogeneity was observed, no moderator was significant.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Baseline, Criterion, and Age were significant moderators of near-transfer follow-up effects (Baseline b = -0.398, p < .050; Criterion b = 0.359, p = .001; Age children b = -0.298, p = .029).

    “Baseline and Criterion were significant moderators (b = -0.398,SE = 0.181, p < .050and b = 0.359, SE = 0.085, p = .001).Also, Age was a significant moderator, with the children performing worse than the adults and older adults (b = -0.298,SE = 0.099, p = .029).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    After excluding the influential case and outlier, the overall very-near-transfer effect at follow-up was g=0.445, SE=0.072, 95% CI [0.290, 0.600], m=21, k=58, df=13.48, p<.001, ω2=0.047, τ2=0.038.

    “After excluding these effects, the overall effect size was g ¯= 0.445, SE = 0.072, 95% CI [0.290; 0.600], m = 21, k = 58, df = 13.48,p < .001,ω 2 = 0.047, τ 2 = 0.038.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    For very-near-transfer measures at follow-up, the PET estimate was g=0.239 (SE=0.134, p=.090) and the PEESE estimate was g=0.326 (SE=0.076, p<.001).

    “The PET estimate was g ¯= 0.239 (SE = 0.134, p = .090),and the PEESE estimate was g ¯= 0.326 (SE = 0.076, p < .001).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Publication-bias-corrected estimates were about g=0.450 at immediate post-test and ranged between g=0.234 and g=0.380 at follow-up.

    “While the corrected estimates were about g ¯= 0.450 at immediate post-test, they ranged between g ¯= 0.234 and g ¯= 0.380 at follow-up.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Trim-and-fill estimates for very-near-transfer measures at follow-up were g=0.380 (SE=0.060, p<.001), g=0.295 (SE=0.068, p<.001), and g=0.234 (SE=0.068, p<.001) with L0, R0, and Q0 estimators, respectively.

    “The trim-and-fill estimates were g ¯= 0.380 (SE = 0.060, p < .001),g ¯= 0.295 (SE = 0.068, p < .001),and g ¯= 0.234 (SE = 0.068, p < .001)with the L0, R0, and Q0 estimators, respectively.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    At follow-up, results were very similar to immediate post-test, with the only difference being effect size.

    “The results at follow-up were very similar to the ones observed immediately after the post-test.Once again, the only difference was represented by the size of the effects.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Effect sizes at follow-up were smaller than at immediate post-test for lesser-near-transfer measures, both uncorrected (g=0.171 vs g=0.248) and publication-bias-corrected (approximately g=0.150-0.160 vs g=0.250-0.260).

    “Once again, the effect sizes at follow-up were smaller than the ones obtained at immediate post-test.This applied both to the uncorrected estimates (g ¯= 0.171 and g ¯= 0.248, respectively) and publication-bias-corrected estimates (approximatively g ¯ = 0.150-0.160and g ¯= 0.250-0260, respectively).”
    Quote in the paperSubstantial claim

  • The quote only partly states this claim.

    Findingown data

    The funnel plot for lesser-near-transfer measures at follow-up looked slightly asymmetrical (Fig. 7).

    “The funnel plot looked slightly asymmetrical (Fig. 7).”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). The quote only partly states this claim.

    Findingown data

    The funnel plot for very-near-transfer measures at follow-up looked slightly asymmetrical (Fig. 6).

    “The funnel plot looked slightly asymmetrical (Fig. 6).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingreasoning

    The effect sizes observed in lesser-near-transfer measures are quite small and tend to diminish a few months after the end of training.

    “First, the effect sizes observed in lesser-near-transfer measures are quite small and tend to diminish a few months after the end of the training.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The improvement on memory tasks remained significant after several months, although it slightly decreased (g=0.365 at follow-up).

    “The training program increased performance on memory tasks immediately after post-test (g ¯= 0.444) and this improvement remained significant after several months, although it slightly decreased (g ¯= 0.365 at follow-up).”
    Quote in the paperSubstantial claim

  • The quote only partly states this claim.

    Findingown data

    Some evidence of publication bias was found at follow-up for lesser-near-transfer effects (e.g., PET estimate g=0.089).

    “Nevertheless, some evidence of publication bias was found at follow-up (e.g., the PET estimate was g ¯= 0.089).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Lesser-near-transfer effects remained significantly different from zero at follow-up (g=0.176, p=.005), although some decrease was observed.

    “The overall effect sizes were small but significantly different from zero; they remained significantly different from zero for a few months after the end of the training, although some decrease was observed (g ¯= 0.246, p = .002and g ¯= 0.176, p = .005,at post-test and follow-up, respectively; see also Tables S1-S4 in the Supplemental material available online).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Lesser-near-transfer overall effect sizes were small but significantly different from zero at post-test (g=0.246, p=.002).

    “The overall effect sizes were small but significantly different from zero; they remained significantly different from zero for a few months after the end of the training, although some decrease was observed (g ¯= 0.246, p = .002and g ¯= 0.176, p = .005,at post-test and follow-up, respectively; see also Tables S1-S4 in the Supplemental material available online).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    The very-near-transfer overall effect was medium (about g = 0.450 at least), while the lesser-near-transfer effect was significantly smaller (about g = 0.250 at most).

    “While the very-near-transfer overall effect was medium (about g ¯ = 0.450 at least), the lesser-near-transfer effect was significantly smaller (about g ¯= 0.250 at most).”
    Quote in the paperSubstantial claim

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

    Premisecitation

    The studies investigating the effects of CWMT are the most numerous and best designed in the category of commercial cognitive-training programs.

    “The studies investigating the effects of CWMT are undoubtedly the most numerous and best designed (e.g., often including active controls) in this category (Simons et al., 2016, pp. 143-148).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Simons et al., 2016 source

      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

    Two authors coded each effect size independently and reached 100% agreement.

    “Two authors coded each effect size independently and reached 100% agreement.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisenone given

    The lack of active controls (or any controls), non-random allocation of participants to groups, and small sample sizes are major flaws that may bias the results of CWMT studies and introduce spurious variability in the pool of data.

    “The lack of active controls (or any controls), non-random allocation of the participants to the groups, and small sample sizes are some of the major flaws that may bias the results of CWMT studies and introduce spurious variability in the pool of data.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    No difference with the follow-up assessment was observed for far-transfer (no appreciable effect or true heterogeneity).

    “Simply put, no difference with the follow-up assessment was observed (no appreciable effect or true heterogeneity).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    While near transfer occurs regularly, far transfer is rare or possibly inexistent.

    “These results highlight that, while near transfer occurs regularly, far transfer is rare or, possibly, inexistent.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Study design and conductcitation

    Robust variance estimation (RVE) with hierarchical weights and small-sample corrections was employed to calculate overall effect sizes and perform meta-regression analysis.

    “We employed robust variance estimation (RVE) with hierarchical weights and small-sample corrections to calculate the overall effect size and perform meta-regression analysis (Hedges, Tipton, & Johnson, 2010;Tanner-Smith & Tipton, 2014;Tanner-Smith, Tipton, & Polanin, 2016).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Hedges, Tipton, & Johnson, 2010 source

      Not checked in the free run.

    • cites Tanner-Smith & Tipton, 2014 source

      Not checked in the free run.

    • cites Tanner-Smith, Tipton, & Polanin, 2016

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    The findings regarding CWMT have been mixed, which has kept researchers from reaching a definite conclusion on the topic.

    “As in most cognitive-training programs, the findings regarding CWMT have been mixed, which has kept researchers from reaching a definite conclusion on the topic.”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Shinaver, Entwistle, & Söderqvist, 2014

      Not checked in the free run.

    • cites Simons et al., 2016 source

      Not checked in the free run.

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

    Study design and conductcitation

    Influential case analysis was performed to test the robustness of the results.

    “To test the robustness of the results, we performed Viechtbauer and Cheung's (2010) influential case analysis (run with Metafor R package; Viechtbauer, 2010).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Viechtbauer and Cheung's (2010)

      Not checked in the free run.

    • cites Viechtbauer, 2010

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    A systematic search strategy was employed to find the relevant studies.

    “A systematic search strategy was employed to find the relevant studies (PRISMA statement; Moher, Liberati, Tetzlaff, & Altman, 2009).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Moher, Liberati, Tetzlaff, & Altman, 2009 source

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    The near-transfer effect was consistent; most true heterogeneity was accounted for by influential/extreme effects, Baseline, and Criterion.

    “Finally, the effect was consistent.In fact, most of the observed true heterogeneity was accounted for by a few influential/extreme effects, the Baseline moderator (i.e., regression to the mean), and the degree of overlap between the trained task and the transfer task (moderator Criterion).”
    Quote in the paperSubstantial claim

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

    Study design and conductown data

    The first two authors coded each effect size for moderator variables independently with high inter-rater agreement (Cohen's kappa = 0.98; 98% agreement).

    “The first two authors coded each effect size for moderator variables independently.The Cohen's kappa was κ = 0.98.”
    • Also stated in: § 3.4. Moderators · plain
      “The first two authors coded each effect size for moderator variables independently.The inter-rater agreement was 98%.”
    Quote in the paperNot a substantial claim

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

    Premisereasoning

    Improving fluid intelligence would benefit other cognitive and academic skills.

    “In turn, improving fluid intelligence would benefit other cognitive and academic skills.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductnone given

    Our inclusion criteria are stricter and led to exclusion of studies whose experimental design does not meet a minimum standard of quality (e.g., no control group, subjective measures).

    “Second, our inclusion criteria are stricter and have led to the exclusion of those studies whose experimental design does not meet a minimum standard of quality (e.g., inclusion of a control group and exclusion of subjective measures of cognitive/academic skill).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    The pattern of results for near-transfer effects at follow-up was the same as at immediate post-test: significant overall effects and true heterogeneity mainly explained by Baseline and Criterion moderators.

    “The pattern of results regarding near-transfer effects at follow-up was thus the same as that at immediate post-test: significant overall effects and some true heterogeneity mainly explained by Baseline and Criterion moderators.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    If WM capacity is increased, then an improvement in fluid intelligence tasks is expected to occur.

    “If WM capacity is increased, then an improvement in such tasks is expected to occur (Jaeggi et al., 2008).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Jaeggi et al., 2008 source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation

    Attentional processes play a role in both fluid intelligence and working memory tasks.

    “Another complementary explanation of the presumed broad generalization of skills following WM training refers to the role played by attentional processes in both fluid intelligence and working memory tasks (Engle, 2018;Gray, Chabris, & Braver, 2003).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Engle, 2018 source

      Not checked in the free run.

    • cites Gray, Chabris, & Braver, 2003 source

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    The inclusion of the most recent studies has increased the precision of the estimated effects and corrected the early optimistic findings.

    “The inclusion of the most recent studies has increased the precision of the estimated effects and corrected the early optimistic findings.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S1 (asserted).

    Premisecitation

    WM and fluid intelligence have a shared capacity constraint.

    “The hypothesis underlying this claim is that WM and fluid intelligence have a shared capacity constraint (Halford, Cowan, & Andrews, 2007).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Halford, Cowan, & Andrews, 2007 source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation

    Performance on fluid intelligence tasks is constrained by the amount of information that can be manipulated by WM.

    “The performance on fluid intelligence tasks (e.g., Raven's progressive matrices) is constrained by the amount of information (i.e., number of items) that can be manipulated by WM.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationown data

    The PET estimator overcorrected (one-tailed p<.100), so the PEESE estimator was more trustworthy.

    “As seen above, the PET estimator overcorrected (one-tailed p < .100).Thus, the PEESE estimator was more trustworthy.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    Enhanced WM capacity is supposed to make information processing more efficient, which should bring a wide set of benefits in academic, professional, and social contexts.

    “Enhanced WM capacity is supposed to make information processing more efficient, which, in turn, should bring a wide set of benefits in academic, professional, and social contexts (see Pearson, 2016).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Pearson, 2016

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Premisecitation

    Human cognition is the product of a biological system.

    “As pointed out by some scholars (e.g., Engle, 2015), human cognition is the product of a biological system.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Engle, 2015

      Not checked in the free run.

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

    Interpretationreasoning

    Very-near transfer indicates that acquired skills can be used in highly similar tasks.

    “The findings regarding far-transfer and very-near-transfer effects are easily interpretable: far-transfer does not exist with CWMT and very-near transfer indicates that the acquired skills can be used in highly similar tasks.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    CWMT does not foster WM capacity, any other core cognitive mechanism, or academic skills.

    “In line with Shipstead et al. (2012a, b), our opinion is that CWMT does not foster WM capacity, any other core cognitive mechanism, or academic skills.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Shipstead et al. (2012a, b)

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Interpretationreasoning

    Improvements on a latent factor extracted from a broad set of memory tasks would represent far more compelling evidence of cognitive enhancement than that often provided in the reviewed primary studies, which are based on few observed measures.

    “Improvements on a latent factor extracted from a broad set of memory tasks would represent far more compelling evidence of cognitive enhancement than that often provided in the reviewed primary studies, which are based on few observed measures.”
    Quote in the paperNot a substantial claim

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

    Interpretationreasoning

    People undergoing CWMT training may simply acquire the ability to perform such tasks slightly more efficiently than controls, which would explain the small observed effect sizes in near-transfer measures and the concurrent absence of far transfer.

    “Thus, people undergoing CWMT training may simply acquire the ability to perform such tasks slightly more efficiently than controls.That would explain the small observed effect sizes in the near-transfer measures and the concurrent absence of far transfer.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    If CWMT were enhancing participants' WM capacity, improvements in other cognitive and academic tasks should have been observed at post-test or follow-up; however, this was not the case.

    “Thus, if CWMT training program were enhancing the participants' WM capacity, improvements in other cognitive and academic tasks should have been observed at either post-test assessment or follow-up assessment.However, this was not the case.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: WM capacity is a major predictor of academic achievement and is highly correlated with fluid intelligence.

  • May combine two claims.

    Interpretationcitation

    Near-transfer effects should not be interpreted as evidence of memory enhancement; rather, they denote improvement in the ability to perform the trained tasks.

    “these effects should not be interpreted as evidence of memory enhancement.Rather, such effects denote improvement in the ability to perform the trained tasks.”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Shipstead et al., 2012a, b

      Not checked in the free run.

    • cites Simons et al., 2016 source

      Not checked in the free run.

    • cites Brehmer, Westerberg, & Bäckman, 2012

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable).

    Premisecitation

    Cognitive skills are commonly defined as the shared variance between many different cognitive tasks.

    “Cognitive skills are commonly defined as the shared variance between many different cognitive tasks (e.g., Strata II and III of the CHC model; McGrew, 2009).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites McGrew, 2009 source

      Not checked in the free run.

52 of 52

  • The authors state this as an assumption.

    If WM capacity could be enhanced by training, the benefits would spread across many other skills positively related to WM.

    “Given these premises, it is natural to assume that if WM capacity could be enhanced by training, the benefits would spread across many other skills positively related to WM (Jaeggi et al., 2008).”

  • The authors state this as an assumption.

    WM training fosters domain-general mechanisms such as WM capacity and attentional control, which in turn enhances other cognitive and academic skills.

    “Therefore, like with the other cognitive-training programs, the essential assumption underlying these hypotheses is that WM training fosters domain-general mechanisms such as WM capacity and attentional control, which in turn enhances other cognitive and academic skills.”

  • The paper relies on this without support.

    WM capacity is a major predictor of academic achievement and is highly correlated with fluid intelligence.

    “Second, and most crucially, WM capacity is a major predictor of academic achievement and is highly correlated with fluid intelligence.”
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?

  • When true heterogeneity is present, it is explained by control group type, statistical artifacts such as regression to the mean, and a few extreme effect sizes.

    “When present at all, true heterogeneity is accounted for by the type of control group used (active or non-active), statistical artifacts such as pre-post-test regression to the mean due to baseline differences between the experimental and control groups, and, to a lesser extent, a few extreme effect sizes.”

    From Working memory training in typically developing children: A multilevel meta-analysis (2020) · cites this paper as Aksayli et al., 2019

    The Results (section 4.1.1) attribute observed far-transfer heterogeneity to regression to the mean, and sections 3.7 and 4.2.2 describe excluding extreme effect sizes. But section 3.4 only identifies active versus non-active control groups as a moderator considered; the reported near-transfer heterogeneity in section 4.2.2 is mainly explained by baseline differences and the type of memory task, not control-group type. The claim therefore overstates what the paper says explains heterogeneity. (LLM, conf=0.85)

    Experimental — accuracy not yet measured

  • Aksayli et al. reported a consistent lack of far transfer, small to medium effect sizes for near transfer, and that these transfer effects depended on the similarity of the transfer tasks to the WM training paradigm.

    “Similar findings were reported by Aksayli et al. [79], who evaluated the effects of adaptive Cogmed WM training, a commercial training programme, relative to non-adaptive Cogmed training. The authors reported a consistent lack of far transfer, small to medium effects sizes for near transfer and that these transfer effects depended on the similarity of the transfer tasks to the WM training paradigm”

    From Adaptive working memory training does not produce transfer effects in cognition and neuroimaging (2022) · cites this paper as [79]

    The Results (section 4.2.1) report a medium near-transfer effect moderated by task overlap, and the Discussion (section 5) reports consistently negligible far transfer. The comparison is narrower than the paper's design: the Inclusion criteria (section 3.2) and Moderators (section 3.4) show that the meta-analysis included several kinds of control group, not only non-adaptive Cogmed training. (LLM, conf=0.88)

    Experimental — accuracy not yet measured

  • Working memory training does not enhance children's domain-general cognitive skills or academic achievement.

    “WM training does not enhance children’s domain-general cognitive skills or academic achievement (Aksayli, Sala, & Gobet, 2019; Melby-Lervåg et al., 2016; Sala & Gobet, 2020).”

    From Cognitive and academic benefits of music training with children: A multilevel meta-analysis (2020) · cites this paper as Aksayli, Sala, & Gobet, 2019

    The Discussion reports no generalized effects on cognitive or academic outcomes regardless of age, which includes children; Section 5.1 likewise concludes that Cogmed does not foster core cognitive mechanisms or academic skills. The meta-analysis evaluates Cogmed specifically, so it does not establish the claim for working-memory training programs generally. (LLM, conf=0.85)

    Experimental — accuracy not yet measured

  • Multiple meta-analyses have reported that true heterogeneity in working memory training, both within and between studies, is null or low.

    “Interestingly, several meta-analyses have estimated within-and between-study true heterogeneity in WM training to be null or low, both with near-transfer and far-transfer effects.”

    From Working memory training in typically developing children: A multilevel meta-analysis (2020) · cites this paper as Aksayli et al., 2019

  • The results are in line with and corroborate Sala and Gobet (2017) and previous meta-analytic syntheses on working memory training in the general population.

    “These results are also in line with Sala and Gobet (2017; a re-analysis with RVE of the data used in that study yields similar results; for the details, see the supplemental materials) and, more broadly, the conclusions of previous meta-analytic syntheses concerning WM training in the general population (Aksayli et al., 2019;Melby-Lervåg et al., 2016;Sala et al., 2019a).”

    From Working memory training in typically developing children: A multilevel meta-analysis (2020) · cites this paper as Aksayli et al., 2019

Cites

Flagged first, then by citations

Research works printed but not linked: 1

  1. Pearson

Cited by

Sorted by citations

Related

Authors
Aksayli, N. D., Sala, G., & Gobet, F.
Affiliations
  • The University of Osaka (JP)
  • University of Liverpool (GB)
Corresponding author
Sala, Giovanni
ORCID
Venue
Educational Research Review
Publisher
Elsevier BV
Published
2019-06
Volume
27
Pages
229-243
ISSN
1747-938X (ISSN-L)
DOI
10.1016/j.edurev.2019.04.003
OpenAlex
W2921463501
Cited by
123
References
75
Citations by year
2026: 16 · 2025: 19 · 2024: 18 · 2023: 13 · 2022: 19 · 2021: 20 · 2020: 14 · 2019: 4
FWCI
2.13
Type
journal article
Language
en
Open access
Licence
Locations
  • Educational Research Review (publishedVersion, cc-by)
  • London School of Economics and Political Science Research Online (London School of Economics and Political Science) (acceptedVersion)
  • London School of Economics and Political Science Theses Online (London School of Economics and Political Science) (submittedVersion)
  • PsyArXiv (OSF Preprints) (submittedVersion)
  • PsyArXiv (OSF Preprints) (submittedVersion, cc-by)
Funders
Japan Society for the Promotion of Science
Topics
  • Social Sciences › Psychology › Experimental and Cognitive Psychology › Cognitive Abilities and Testing
  • Social Sciences › Psychology › Neuropsychology and Physiological Psychology › Developmental and Educational Neuropsychology
  • Physical Sciences › Mathematics › Statistics and Probability › Cognitive and developmental aspects of mathematical skills
Keywords
Cogmed, working memory training, cognitive training, far transfer, near transfer, academic performance, attention, intelligence, memory tasks, language skills, mathematics achievement

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