CoChat
Sign Up
Cited by 235Anxiety, Depression, Psychometrics, Treatment, Cognitive ProcessesPsychological Treatments and AssessmentsPsychotherapy Techniques and Applications

How to boost positive interpretations? A Meta-Analysis of the effectiveness of cognitive bias modification for interpretation

Open access
Menne-Lothmann, Claudia, Viechtbauer, Wolfgang, Höhn, Petra, Kasanova, Zuzana, Haller, Simone P., Drukker, Marjan, van Os, Jim, Wichers, Marieke +1 more·2014·PLoS ONE
DOI 10.1371/journal.pone.0100925PMC PMC4072710
Open this paper in CoChat — free accountOpen PDF

Abstract

The current meta-analysis explores the strength of effects of cognitive bias modification training for interpretation bias (CBM-I) on positive (i.e., adaptive) interpretations and mood as well as the training and sample characteristics influencing these effects. Data-bases were searched with the key words "interpret* bias AND training" and "interpret* bias AND modif*". Reference lists of identified articles were checked and authors of identified articles were contacted for further relevant articles and unpublished data. Studies were reviewed for inclusion with eligibility criteria being that the study (a) aimed to target interpretation biases through any kind of training, (b) assessed mood and/or interpretation bias as outcome measures, (c) allocated individuals to training conditions at random, and (d) recruited adult samples. A meta-analytic multilevel mixed-effects model was employed to assess standardized mean changes in interpretation bias, negative mood, and emotional reactivity. In addition, several training and sample characteristics were explored for their potential to enhance benign training effectiveness. On average, benign CBM-I resulted in an increase in positive interpretation bias (p<.01) and a decrease in negative mood state (p<.001), but did not affect emotional reactivity. These effects were not consistently different from control conditions with no or neutral training. However, within benign training conditions imagery instructions and more training sessions were related to larger cognitive and mood effects, whereas feedback about training performance and inclusion of non-benign training items (instead of including benign items only) boosted cognitive effects only. Finally, training was more effective in women (cognitive and mood effects) and presumably samples with symptomatic emotional dysregulation (cognitive effects). Although the effects of emotional dysregulation and number of training sessions could not well be distinguished, there is an indication that when used with imagery instructions and more training sessions, benign CBM-I can be employed as a useful complementary treatment to usual psychotherapies.

AI snapshot

Written from the paper’s full text.

TL;DR

This meta-analysis of 42 articles (59 studies, 125 samples, 2,526 individuals) found that benign cognitive bias modification for interpretation (CBM-I) increased positive interpretation bias and decreased negative mood, but effects were not consistently different from neutral or no-training controls. Imagery instructions and more training sessions were the most consistent moderators of larger cognitive and mood effects.

Key findings

  • Benign CBM-I produced a large post-training difference favoring positive over negative interpretations (ES = 1.33, 95% CI 1.11–1.55, p < .001), significantly larger than neutral (ES = 0.49), negative (ES = −0.05), and no-training (ES = 0.28) conditions (all pairwise p < .001).
  • Pre- to post-training positive interpretation bias increased significantly only in benign training (ES = 0.43, 95% CI 0.17–0.69, p < .01), differing significantly only from negative training (ES = −0.22; pairwise ES = 0.65, p < .05).
  • Negative mood decreased significantly pre- to post-training in benign training (ES = 0.25, 95% CI 0.14–0.36, p < .001), but this did not differ significantly from neutral (ES = 0.22) or no-training (ES = −0.03) conditions; it did differ from negative training (ES = −0.20; pairwise ES = 0.44, p < .001).
  • Benign CBM-I did not attenuate negative mood reactivity after an emotional challenge: negative mood increased in all conditions (benign ES = −0.79, negative ES = −0.80, neutral ES = −1.03, no training ES = −0.77), with no significant pairwise differences.
  • Moderators: imagery instructions increased post-training positive-vs-negative bias from ES = 0.54 to 1.59 and pre-post positive bias by ES = 0.50; each additional session increased positive bias by ES = 0.08 and mood improvement by ES = 0.07; feedback increased pre-post positive bias by ES = 0.46; non-benign training items (ratio < 1) boosted cognitive effects; effects were larger in women and, for cognitive change, in symptomatic samples.

Methods

  • Systematic review and meta-analysis following PRISMA; databases searched through June 2013 with terms “interpret* bias AND training” and “interpret* bias AND modif*,” plus reference checks and author contacts.
  • Inclusion criteria: studies targeting interpretation bias through training, random allocation for multi-condition studies, interpretation bias and/or mood as outcomes, adult samples (healthy, subclinical/high-trait, or clinical anxiety/depression), and publication in English, German, or Dutch.
  • Multilevel mixed-effects meta-analytic models estimated standardized mean changes/differences for four contrasts: post-training positive vs negative interpretations, pre-post positive interpretation bias, pre-post negative mood, and pre-post emotional challenge negative mood; moderators were tested separately for benign training.

Limitations

  • The paper states that unpublished data and publication bias were addressed, with no indication of publication bias (funnel-plot asymmetry p-values = .72, .79, .46, .89), but it acknowledges that unpublished studies could still matter.
  • The paper states that it focused only on mood state, not the full symptom spectrum; effects on low mood were small, and effects on guilt, rumination, or other symptoms cannot be answered.
  • The paper states that moderators were tested separately, so their combined effects are unknown; repeated training sessions and symptomatic status were strongly associated (r = .73) and could not be well distinguished for pre-post positive interpretation bias.
  • Evident from the text: paradigm and imagery instructions were almost perfectly confounded (all AS paradigms used imagery; only one homograph paradigm did), so it is impossible to tell whether paradigm or imagery drove those effects; ratio < 1 was based on only three samples; neutral and no-training comparisons had much smaller sample counts than benign training, reducing power.

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

Integrity signals

Registry

Crossref · 10.1371/journal.pone.0100925

Cites

76 of 76 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 50 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    Within benign training conditions, more training sessions are related to larger cognitive and mood effects.

    “within benign training conditions imagery instructions and more training sessions were related to larger cognitive and mood effects”
    Quote in the paperSubstantial claim
  • Findingown data

    Within benign training conditions, feedback about training performance boosts cognitive effects only.

    “feedback about training performance and inclusion of non-benign training items (instead of including benign items only) boosted cognitive effects only.”
    Quote in the paperSubstantial claim
  • Findingown data

    Benign CBM-I does not affect emotional reactivity.

    “On average, benign CBM-I resulted in an increase in positive interpretation bias (p<.01) and a decrease in negative mood state (p<.001), but did not affect emotional reactivity.”
    Quote in the paperSubstantial claim
  • Findingown data

    Within benign training conditions, inclusion of non-benign training items (instead of including benign items only) boosts cognitive effects only.

    “feedback about training performance and inclusion of non-benign training items (instead of including benign items only) boosted cognitive effects only.”
    Quote in the paperSubstantial claim
  • Findingown data

    On average, benign CBM-I decreases negative mood state (p<.001).

    “On average, benign CBM-I resulted in an increase in positive interpretation bias (p<.01) and a decrease in negative mood state (p<.001), but did not affect emotional reactivity.”
    Quote in the paperSubstantial claim
  • May combine two claims. 77
  • Classified as probably not a substantial claim. 36
  • The quote only partly states this claim. 1

  • May combine two claims.

    Findingown data

    The effects of benign CBM-I on positive interpretation bias and negative mood are not consistently different from control conditions with no or neutral training.

    “These effects were not consistently different from control conditions with no or neutral training.”
    Quote in the paperSubstantial claim

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

    Findingown data

    No training produced a pooled within-group effect size of -0.77 (95% CI -1.49--0.05, p=.04) for difference in negative mood from pre- to post-emotional challenge.

    “| | | 1 | No training | −0.77 | −1.49; −0.05 | .04 |”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Both the study-level and effect-size-level variance components were statistically significant (p=.002 and p<.001, respectively).

    “Both the study level variance component and the effect size level variance component did reach statistical significance (p = .002 and p<.001, respectively – not depicted in table 2).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Across 75 independent samples from 34 studies, neutral training produced a significant post-training difference between endorsement of positive and negative interpretations, with effect size 0.49 (95% CI 0.06-0.93).

    “For the same difference in the neutral condition this effect was also significant but considerably smaller (i.e., 0.49, 95% CI 0.06–0.93).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    In benign training, negative mood decreased significantly from pre- to post-training (ES=0.25, 95% CI 0.14-0.36), indicating on average 60% showed reductions.

    “Overall negative mood decreased significantly (p<.001) in the benign training condition (a small effect size of 0.25 with 95% CI 0.14–0.36; table 2) indicating that, on average, 60% of individuals showed reductions in negative mood in the benign condition.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    In negative training, negative mood increased significantly from pre- to post-training (ES=-0.20, 95% CI -0.35- -0.05), indicating on average 58% showed increases.

    “In contrast, negative mood increased significantly (p<.01) for those receiving negative training (a small effect size of −0.20 with 95% CI −0.35– −0.05; table 2), indicating that, on average, 58% of individuals showed increases in negative mood after negative training”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Across 75 independent samples from 34 studies, benign training produced a significant post-training difference between endorsement of positive and negative interpretations, with effect size 1.33 (95% CI 1.11-1.55).

    “As shown in Table 2, across 75 independent samples (k = 75) (distributed over a total of 34 studies), benign and neutral training demonstrated a significant difference between endorsement of positive and negative interpretations after training. The effect size of this difference for the benign training was substantial (i.e., 1.33 with 95% CI 1.11–1.55).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    For pre-post change in positive interpretation bias, the effect-size-level variance component was significantly larger than zero (p<.001), but study-level heterogeneity was not significant (p=.58).

    “While the effect size level variance component was significantly larger than zero (p<.001), we did not find significant study-level heterogeneity (p = .58) for this outcome”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    For pre-post change in negative mood, only effect-size-level heterogeneity was significant (p<.001), not study-level (p=.86).

    “Again, only heterogeneity at the effect size level (p<.001) and not the study level (p = .86) was found”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    For pre-post emotional challenge change in mood, there was significant study-level heterogeneity (p<.01) but not effect-size-level heterogeneity (p=.18).

    “In contrast to the earlier outcomes, we found significant heterogeneity at the study (p<.01), but not at the effect size level (p = .18).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The change in negative mood for negative training differed significantly from neutral training (p<.01), but benign training did not differ significantly from neutral or no-training conditions (both p's >.05).

    “Moreover, while the change in the negative training condition varied significantly from the change in the neutral condition (p<.01; table 3), there was no significant difference in the amount of change in mood state between the benign and neutral or no-training conditions (both p’s >.05)”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The paradigm employed significantly impacted the post-training difference between positive and negative interpretation bias, with the AS paradigm (k=26) being significantly more effective.

    “The paradigm employed significantly impacted on the post-training difference between positive and negative interpretation bias in that the AS paradigm (k = 26) was significantly more effective (Table 4).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    No use of feedback (k=12) yielded a non-significant small effect (ES=.01, p=.96, 54% chance), but use of feedback (k=8) increased the effect significantly by ES=.46 (p<.05), resulting in a total effect of ES=.46 (68% absolute chance and 26% higher chance).

    “Additionally, while no use of feedback (k = 12) yielded a non-significant small effect (ES = .01, p = .96, corresponding to a 54% chance for an increase in positive interpretations), the use of feedback (k = 8) increased this effect significantly by ES = .46 (p<.05). This results in a total effect of feedback administration of ES = .46 (68% absolute chance and 26% higher chance of increase in”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingreasoning

    The AS paradigm advantage corresponds to a 19 percentage point absolute difference and a 25% higher relative chance of endorsing positive over negative interpretations.

    “In absolute terms, that is a 94%−75% = 19% percentage points difference. Relative, that is 94/75 = 1.25, that is a 25% higher chance of endorsing positive interpretations more strongly than negative interpretations.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    For pre-post change in positive interpretation bias, paradigm, imagery instructions, and sex emerged as significant moderators.

    “For the change in positive interpretation bias from pre- to post-training, again paradigm, imagery instructions, and sex emerged as significant moderators.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    No use of imagery instructions (k=8) resulted in an effect size of .54 (p<.01), indicating a chance of about 71% for endorsing more positive than negative interpretations.

    “No use of imagery instructions (k = 8) resulted in an effect size of.54 (p<.01) indicating a chance of about 71% for endorsing more positive than negative interpretations.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Paradigm and imagery-use were almost perfectly confounded; all AS paradigms employed imagery instructions, whereas only one of the nine homograph paradigms did.

    “However, it needs to be noted that paradigm and imagery-use were almost perfectly confounded; all AS-paradigms employed imagery instructions, whereas only one of the nine homograph paradigms did.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The effect increased by ES=1.05 (p<.001) when imagery instructions were employed (k=27), resulting in an overall ES of 1.59, indicating a chance of almost 94% of endorsing more positive than negative interpretations, a 33% higher chance than when no imagery instructions were employed.

    “This effect increased by ES = 1.05 (p<.001), when imagery instructions were employed (k = 27)(see Table 4), resulting in an overall ES of 1.59 indicating a chance of almost 94% of endorsing more positive than negative interpretations, that is a 33% higher chance than when no imagery instructions were employed.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    In all-female samples, the post-training difference between positive and negative interpretation biases was large and significant (ES=2.03, p<.001, 98% chance), and this difference decreased by ES=-.02 (p<.05) per percentage point more males in the sample.

    “Finally, while in all-female samples the difference between positive and negative interpretation biases was large and significant (ES = 2.03, p<.001, 98% chance of endorsing more positive than negative interpretations after training), this difference decreased by ES = −.02 (p<.05) per percentage point more males in the sample (see table 4).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The effect was significantly increased by ES=.95 (p<.001) when other paradigms (k=6) were employed, resulting in an overall ES=1.08, corresponding to a 54% higher chance of an increase in positive interpretation bias.

    “This effect was then, however, significantly increased by ES = .95, p<.001) when ‘other’ paradigms (k = 6; 1 homograph; 3 WSAT, and 2 other paradigms) were employed. This results in an overall ES = 1.08 corresponding to a 54% higher chance of an increase in positive interpretation bias.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    For post-training endorsement of positive versus negative interpretations, use of imagery instructions and male sex within the sample correlated by r=-.60, indicating that samples including fewer male participants were more often instructed to use imagery.

    “For the endorsement of positive versus negative interpretations after training, the use of imagery instructions and male sex within the sample correlated by r = −.60 indicating that samples including fewer male participants were more often instructed to use imagery.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    All other paradigms combined (k=9) yielded a significant effect size of .69 (p<.001), indicating a 75% chance on average for endorsing positive interpretations more strongly than negative interpretations.

    “All other paradigms combined (k = 9) yielded a significant effect size of.69 (p<.001), indicating that there is a 75% chance (on average) for endorsing positive interpretations more strongly than negative interpretations.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingreasoning

    For AS paradigms, the effect increased by ES=.90, yielding an overall effect of 1.59, corresponding to a 94% chance for endorsing positive interpretations more strongly than negative interpretations.

    “For AS-paradigms this effect increased by ES = .90 (see table 4), yielding an overall effect of.69+.90 = 1.59, which corresponds to a 94% chance for endorsing positive interpretations more strongly than negative interpretations.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    No use of imagery instructions (k=6) did not yield a significant effect (ES=-.20, p=.21, 42% chance), but imagery instructions (k=12) significantly enhanced the effect by ES=.50 (p<.01), resulting in a net effect of ES=.30 (62% absolute chance and 48% higher chance).

    “Whereas no use of imagery instructions (k = 6) did not yield a significant effect (ES = −.20, p = .21, 42% chance of increase in positive interpretation bias), this effect was significantly enhanced by ES = .50 (p<.01) when imagery instructions were employed (k = 12). This results in a net-effect of imagery instructions of ES = .30 (62% absolute chance and 48% higher chance of increase in positive”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingreasoning

    In a sample with 40% males, the total effect converges to 1.28, corresponding to an absolute chance of 90%, a 9% decreased chance of endorsing more positive than negative interpretations after training.

    “Thus, in a sample with for example 40% males this converges to a total effect of 1.28, corresponding to an absolute chance of 90%, which is a 9% decreased chance of endorsing more positive than negative interpretations after training.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    More than one training session was only administered to individuals with symptoms, whereas one training session was administered to both healthy and symptomatic samples.

    “However, it is important to mention that more than one training session was only administered to individuals with symptoms whereas one training session was administered to both healthy and symptomatic samples (also see table 1)”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Adding one session significantly increased the decrease in negative mood by ES=.07 (p<.05), resulting in a total ES=.16 (56% absolute and 4% higher chance when administering two instead of one session).

    “Adding one session (k one session = 40; k four sessions = 2; k six sessions = 1; k seven sessions = 2; k eight sessions = 2; k twelve sessions = 1) resulted in a significant larger decrease in negative mood by increasing the effect by ES = .07, p<.05, resulting in a total ES = .16, which corresponds to a 56% absolute and 4% higher chance of decrease in negative mood when administering for example”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingreasoning

    The effect of 10 sessions of training would result in an ES=.80, with a higher chance of approximately 49%.

    “Although this may seem a very small effect, the effect of 10 sessions of training would already result in an ES = .80, with a higher chance of approximately 49%.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    All-female samples showed a large and significant effect (ES=1.30, p<.001, 90% chance), and this effect was significantly reduced by ES=.03 (p<.01) per percentage point males; for 40% males, the chance for an increase would be lowered by 61%.

    “Conform above results, while all-female samples showed a large and significant effect (ES = 1.30, p<.001, 90% chance of increase), this effect was significantly reduced by ES = .03 (p<.01) per percentage point males in the sample. Thus, a sample with for example 40% males, the chance for an increase in positive interpretation bias would be lowered by 61%.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The ratio of benign training items to all training items significantly moderated the pre-post change in positive interpretation bias; ratio .69 yielded ES=0.94 (83% chance), and ratio 1.00 significantly reduced the effect (p<.01), resulting in ES=.07 (53% absolute chance and 57% lower chance).

    “Also, the ratio of benign training items versus all training items (ratio 0.69 (k = 2), 0.80 (k = 1), 1.00 (k = 17)) showed a significant effect. While the ratio of.69 yielded a large effect (ES = 0.94, corresponding to a 83% chance of an increase in interpretation bias) this effect was significantly reduced when the ratio became 1.00 (p<.01). This results in an absolute effect of ES = .07 (53%”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    No imagery instructions resulted in a non-significant effect (k=11; ES=-0.25, p=.083, 40% chance), but adding imagery instructions (k=33) increased the effect significantly by ES=.50 (p<.01), resulting in a total ES=.25 (60% absolute and 50% higher chance).

    “The use of no imagery instructions resulted in a non-significant effect (k = 11; ES = −0.25, p = .083, 40% chance of decrease in negative mood). However, adding imagery instructions (k = 33) increased this effect significantly by ES = .50 (p<.01), resulting in a total ES = .25, which corresponds to a 60% absolute and 50% higher chance of decrease in negative mood.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Adding one training session significantly increased the effect of training by ES=.08 (p<.01; total ES=.01; absolute chance of 51% and higher chance of 6%).

    “Finally, adding one training session significantly increased the effect of training by ES = .08 (k one session = 12; k four sessions = 2; k six sessions = 1; k seven sessions = 2; k eight sessions = 2; k twelve sessions = 1, p<.01; total ES = .01; absolute chance of 51% and higher chance of 6%).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    None of the moderators demonstrating a significant effect correlated strongly with another (all r's ≤.50), except for number of training sessions and health status (r=.73).

    “None of the moderators demonstrating a significant effect correlated strongly with another (all r’s ≤.50), except for the number of training sessions and health status of individuals (r = .73) (also see Table S3).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Post-hoc analyses including both number of training sessions and health status showed ES=.09 (p<.05) for number of sessions and ES=-.01 (p=.99) for health status.

    “Post-hoc analyses including both ‘number of training sessions’ and ‘health status’ revealed the following effects: ES = .09, p<.05 for ‘number of sessions’ and ES = −.01, p = .99, for ‘health status’”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The effect was significantly increased by ES=.29 (p=.05) when other paradigms were employed (k=10), with a total effect of ES=.48 (68% absolute chance and 17% higher chance).

    “This effect was significantly increased by ES = .29 (p = .05) when other paradigms were employed (k = 10). The total effect when other paradigms were employed therefore was ES = .48, with an absolute chance of 68% and a higher chance of 17% of decrease in negative mood.”
    Quote in the paperSubstantial claim

  • The quote only partly states this claim.

    Findingown data

    The correlations between the significant moderators for change in negative mood were small (all r's <.36).

    “The correlations between the significant moderators here were small (all r’s <.36) (also see Table S3).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    There was no significant correlation between post-training endorsement of positive versus negative bias or pre-post change in positive bias and increase in negative mood in response to an emotional challenge (r=-.0001, CI -0.41-0.41, k=23 and r=-.15, CI -0.77-0.62, k=8, respectively).

    “There was no significant correlation between the endorsement of positive versus negative bias after training or the change in positive bias from before to after training and the increase in negative mood in response to an emotional challenge (r = −.0001, CI −0.41–0.41, k = 23 and r = −.15, CI −0.77–0.62, k = 8, respectively).”
    Quote in the paperNot a substantial claim

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

    Findingown data

    Some additional analyses possibly hint at asymmetry in the pre-training versus post-training endorsements of positive interpretations outcome.

    “(see Figure S1 for some additional analyses possibly hinting at some asymmetry in the pre-training versus post-training endorsements of positive interpretations outcome).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    While benign training effects differed reliably from negative training effects, the difference relative to neutral or no-training conditions remained insignificant for the change in positive bias and mood.

    “While these benign training effects differed reliably from negative training effects, the difference relative to neutral or no-training conditions remained insignificant for the change in positive bias and mood.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    In the current analyses, benign CBM-I resulted in large post-training endorsements of positive relative to negative interpretations, small to medium changes in positive interpretational style from pre- to post training, and small decreases in negative mood states from pre- to post training.

    “In the current analyses, benign CBM-I resulted in large post-training endorsements of positive relative to negative interpretations, in small to medium changes in positive interpretational style from pre- to post training, and small decreases in negative mood states from pre- to post training.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Benign CBM-I training resulted in significant but small decreases in negative mood, which significantly differed from changes in the opposite direction in the negative condition, but not from changes in neutral and no-training conditions.

    “Benign CBM-I training resulted in significant, but small decreases in negative mood, which significantly differed from changes in the opposite direction in the negative condition, but not from changes in neutral and no-training conditions.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The benign condition showed significant but small changes in positive interpretation bias from pre- to post-training, which differed from negative training only.

    “The benign condition showed significant but small changes, which differed from negative training only.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Whereas benign CBM-I seems to be the only condition to significantly affect interpretational biases, mood was affected in all but the no-training condition.

    “Whereas benign CBM-I seems to be the only condition to significantly affect interpretational biases, mood was affected in all but the no-training condition.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Benign CBM-I resulted in a large post-training difference between positive and negative interpretation bias, which was also significantly larger than in any comparison condition.

    “Benign CBM-I resulted in a large post-training difference between positive and negative interpretation bias, which was also significantly larger than in any comparison condition.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Imagery instructions were able to significantly boost improvements in both positive interpretation bias and negative mood.

    “Most convincingly, imagery instructions were able to significantly boost improvements in both positive interpretation bias and negative mood.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Repeated training sessions were most often employed in samples with mood-symptoms, and mood-symptoms also showed significant influence on the increase in positive interpretation bias.

    “repeated training sessions were most often employed in samples with mood-symptoms (with the latter also showing significant influence on the increase in positive interpretation bias)”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The effect sizes for adding one additional training session were small: .08 for increase in positive interpretation bias and .07 for decrease in negative mood.

    “Although the effect sizes were small (.08 for increase in positive interpretation bias and.07 for decrease in negative mood)”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Both sex and imagery yielded significant effects for each of the other outcome measures as well, where only very weak correlations were observable, supporting the robustness of the imagery effect.

    “However, both sex and imagery yielded significant effects for each of the other outcome measures as well (where only very weak correlations were observable) supporting the robustness of the imaging effect.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingreasoning

    The increase in positive interpretation bias from pre- to post-training was particularly prominent in samples with anxiety and mood symptoms/diagnoses as compared to healthy non-symptomatic controls, which showed a non-significant change.

    “The increase in positive interpretation bias from pre- to post-training was particularly prominent in samples with anxiety and mood symptoms/diagnoses as compared to healthy (non-symptomatic) controls (showing a non-significant change).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Particularly women tend to benefit more from benign CBM-I, both cognitively (interpretation bias) and emotionally (mood), with significant and large effect sizes.

    “Our findings indicated that particularly women (who are more vulnerable to develop mood symptoms) tend to benefit more from benign CBM-I, both cognitively (interpretation bias) and emotionally (mood) with significant and large effect sizes.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Women are more vulnerable to develop mood symptoms.

  • May combine two claims.

    Findingreasoning

    Administering feedback about response accuracy and having a moderate ratio of benign training items and total items within the training was related to a larger increase in positive interpretation bias with moderate to large effect sizes, while not influencing any other outcome measures.

    “While not influencing any of the other outcome measures, administering feedback about response accuracy and having a moderate ratio of benign training items and total items within the training was further related to a larger increase in positive interpretation bias with moderate (feedback) to large (ratio) effect sizes.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingreasoning

    Post-hoc analyses did not support the influence of emotional symptoms, but all studies administering more than one training session were conducted in symptomatic samples, possibly suggesting that repeated sessions are of added value within symptomatic samples.

    “Although post-hoc analyses did not support the influence of emotional symptoms, it is important to notice that all studies administering more than one training session were conducted in symptomatic samples, possibly suggesting that repeated sessions are of added value within symptomatic samples.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Benign CBM-I was associated with an increase in positive interpretation bias and a decrease in negative mood.

    “It was associated with an increase in positive interpretation bias and a decrease in negative mood.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Employing imagery instructions in particular and to a lesser extent repeated training sessions, a moderate ratio of benign training items and total items, and administering feedback during training can significantly increase these effects.

    “employing imagery instructions in particular and to a lesser extent repeated training sessions, a moderate ratio of benign training items and total items, and administering feedback during training can significantly increase these effects.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The increase in positive interpretation bias and decrease in negative mood were significantly correlated, indicating that an improvement in cognitive style goes hand in hand with immediate improvements in mood.

    “These effects (i.e., increase in positive interpretation bias and decrease in negative mood) were significantly correlated indicating that an improvement in cognitive style goes hand in hand with immediate improvements in mood.”
    Quote in the paperSubstantial claim

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

    Study design and conductcitation

    The review protocol adhered to the PRISMA guidelines.

    “The protocol for reviewing the articles was developed by the first author based on the suggested strategy by Lipsey and Wilson [22] and adhered to the PRISMA guidelines (Table S1).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [22] MW Eysenck, 1991, Bias in interpretation of ambiguous sentences related to threat in anxiety, Journal of Abnormal Psychology source

      Not checked in the free run.

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

    Study design and conductnone given

    The whole screening procedure was conducted independently by two researchers.

    “The whole screening procedure was conducted independently by two researchers (CML and PH for the first two searches, and CML and SH for the last search).”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    Women present more often with emotional problems than men.

    “women present more often with emotional problems than men, e.g. see [21]”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [21] MM Husky, 2009, Past depression and gender interact to influence emotional reactivity to daily life stress, Cognitive Therapy and Research source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    Allocation to training groups was done at random in studies with more than one training group.

    “for studies with more than one training group, allocation to training groups was done at random to prevent confounding of the data by any third variables such as personal preference.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    Risk of bias in individual studies was minimized by making randomization to training condition an inclusion criterion for reports with more than one training condition.

    “Risk of bias in individual studies was attempted to be kept at a minimum by making randomization to training condition an inclusion criteria for those reports including more than one training condition.”
    Quote in the paperSubstantial claim

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

    Study design and conductnone given

    For studies with training across several sessions, attrition within each training condition was coded as a proxy for risk of selected attrition in the benign condition.

    “For those studies that administered training across several sessions, attrition (percentage of drop-out) within each training condition was coded as a proxy for risk of selected attrition in the benign condition.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductown data

    All included articles were independently coded by at least two researchers using a standardized coding protocol.

    “All included articles were independently coded by at least two researchers (CML, PH, ZK, and SH) using a standardized coding protocol and appropriate forms (available upon request from the first author; also see ‘protocol’).”
    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

    All missing data were requested from authors via e-mail, and all but five out of 35 contacted responded (response rate = 86%).

    “All missing data were requested from authors via e-mail. As most primary articles report between-group differences at post-training rather than within-group comparisons, almost all authors were contacted. All but five out of 35 of those contacted responded positively (response rate = 86%).”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    To avoid non-independent effect size estimates, only one measure was selected based on an a priori established preference list.

    “To avoid the problem of non-independent effect size estimates in these samples, we selected only one measure based on an a priori established preference list [26].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [26] Cooper H (2010) Research synthesis and meta-analysis: A step-by-step approach. Thousands Oaks, CA: Sage.

      Not checked in the free run.

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

    Study design and conductown data

    Summary statistics necessary to compute effect sizes were retrieved by at least two independent coders, compared, and if necessary adapted.

    “The summary statistics necessary to compute the effect sizes (see below) were retrieved from the articles by at least two independent coders (CML, PH, ZK, and SH), compared, and if necessary adapted.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    If the same group of subjects underwent more than one training, only d for the first training condition the group was exposed to was computed.

    “if the same group of subjects underwent more than one training (whether it be a different training condition or a variation of the same training condition) we only computed d for the first training condition the group was exposed to.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    For studies where r was unknown and SDdiff had to be computed, the mean correlation was employed to impute r separately for each of the four training conditions per contrast.

    “For studies where r was unknown and SDdiff had to be computed in order to obtain d, the mean correlation was employed to impute r (this was done separately for each of the four training conditions per contrast).”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    To address whether random assignment and attrition affect results, these variables were coded for each study.

    “To address the question whether random assignment and attrition affects results these variables were coded for each study.”
    Quote in the paperSubstantial claim

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

    Study design and conductreasoning

    The main analyses were repeated excluding studies with only one training condition and excluding conditions with an attrition rate >15%.

    “To address the potential risk of bias within studies, the main analyses were repeated excluding studies with only one training condition and therefore no random assignment to training condition and excluding conditions that had an attrition rate of >15%.”
    Quote in the paperNot a substantial claim

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

    Study design and conductcitation

    Articles were retrieved according to the PRISMA guidelines.

    “Articles were retrieved according to the PRISMA guidelines [24].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [24] D Moher, 2009, Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement, PLoS Med 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 conductown data

    Publication bias was examined visually by inspecting funnel plot asymmetry and by testing the inverse of sample size as a covariate.

    “We also examined the presence of publication biases visually (by inspecting the presence of asymmetry in funnel plots) and by including and testing the inverse of the sample size as a potential covariate in the models [30].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [30] JL Peters, 2006, Comparison of two methods to detect publication bias in meta-analysis, Journal of the American Medical Association source

      Not checked in the free run.

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

    Premisecitation

    For an effect size of 0.2, 58% of individuals are expected to show an increase in positive interpretation bias (and a decrease in negative mood).

    “For an effect size of 0.2 (a “small” effect), the positive interpretation bias is expected to increase (and negative mood is expected to decrease) for φ(0.2) = .58 (i.e., 58%) of individuals.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [31] SB Morris, 2002, Combining effect size estimates in meta-analysis with repeated measures and independent-group desings, Psychological Methods source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    For effect sizes of 0.5 and 0.8, 69% and 79% of individuals, respectively, are expected to show an increase in positive interpretation bias (and a decrease in negative mood).

    “Effect sizes of 0.5 (a “medium” effect) and 0.8 (a “large” effect) correspond to an increase in positive interpretation bias (and a decrease in negative mood) in 69% and 79% of individuals, respectively [31].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [31] SB Morris, 2002, Combining effect size estimates in meta-analysis with repeated measures and independent-group desings, Psychological Methods source

      Not checked in the free run.

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

    Premisecitation

    A significant relationship between the inverse of the sample size and the observed d-values may be suggestive of publication bias.

    “A significant relationship between the inverse of the sample size and the observed d-values may be suggestive of publication bias.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [30] JL Peters, 2006, Comparison of two methods to detect publication bias in meta-analysis, Journal of the American Medical Association source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Study design and conductown data

    Moderators were entered separately because entering them simultaneously would have lost the majority of data due to missing values and drastically reduced power.

    “Each of the moderators was added to the meta-regression separately (we were unable to enter the moderators simultaneously as the majority of data would be lost due to missing values (also see table 1), which would have drastically reduced the power to find any relationships).”
    Quote in the paperSubstantial claim

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

    Premisecitation

    For a standardized mean change d, φ(d) estimates the proportion of individuals for whom the difference or change scores reflect a preferable outcome.

    “Letting φ() denote the cumulative density function of a standard normal distribution, the interpretation of the standardized mean change can also be facilitated by noting that φ(d) estimates the proportion of individuals for which the difference or change scores reflect a preferable outcome [31].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [31] SB Morris, 2002, Combining effect size estimates in meta-analysis with repeated measures and independent-group desings, Psychological Methods source

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    When analyzing specific training characteristics, the type of training paradigm was controlled for whenever it had a significant moderating effect.

    “However, since training paradigm (AS, homograph, and WSAT) is a rather complex variable possibly differentially influencing the effect of other training characteristics, we controlled for type of paradigm (whenever training paradigm resulted in a significant moderating effect) when analyzing the more specific training characteristics.”
    Quote in the paperSubstantial claim

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

    Interpretationown data

    The pooled effects for pre-post emotional challenge change in mood reflect averages across studies.

    “As before, the pooled effects reflect averages across studies.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    The effects of repeated number of training sessions and mood symptoms for the increase in positive interpretation bias were strongly associated and therefore need to be viewed with caution.

    “Although most correlations between the significant moderators were far from perfect (≤.5), the effects of repeated number of training sessions and mood symptoms for the increase in positive interpretation bias were strongly associated and therefore need to be viewed with caution.”
    Quote in the paperSubstantial claim

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

    Premisecitation

    CBM-I has recently been considered a promising clinical tool with potential for boosting positive thoughts and thereby improving emotional symptoms.

    “Cognitive bias modification training for interpretation bias (CBM-I) has recently been considered a promising clinical tool, e.g., see [33] with potential for boosting positive thoughts (i.e., positive or benign interpretations of ambiguous situations) and thereby improving emotional symptoms.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [33] SE Blackwell, 2010, Modifying interpretation and imagination in clinical depression: a single case series using cognitive bias modification, Applied Cognitive Psychology source

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    Training paradigm, imagery instructions, number of training sessions, and sex moderated the impact of benign training on at least two different outcome measures.

    “To sum up, training paradigm, imagery instructions, the number of training sessions, and sex moderated the impact of benign training on at least two different outcome measures.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationcitation

    These findings partly support but also extend results from the earlier meta-analysis which combined assessment of CBM-I effects with another training program, CBM-A.

    “These findings partly support but also extend results from the earlier meta-analysis which combined assessment of CBM-I effects with another training program, CBM-A [11].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [11] LS Hallion, 2011, A meta-analysis of the effect of cognitive bias modification on anxiety and depression, Psychological Bulletin source

      Not checked in the free run.

  • May combine two claims.

    Study design and conductown data

    Compared to benign training, no-training and neutral training were administered less frequently (three and six samples vs 20 samples), so power for these comparisons might have been reduced.

    “However, it also needs to be taken into account that compared to the benign training no-training and neutral training were administered less frequently (three and six samples as compared to 20 samples), hence power for these comparisons might have been reduced.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    Only benign training showed a significant change in interpretational style, indicating its potential as an interventive tool.

    “Yet, it is still notable that only benign training showed a significant change in interpretational style, indicating its potential as an interventive tool.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationcitation

    Watching stressful videos, trying to solve an unsolvable anagram task under time pressure, or giving a speech should be stressful for most individuals.

    “For example, watching stressful videos [39], trying to solve an unsolvable anagram task under time pressure ‘knowing’ that ‘most people have no problems solving it in time’, e.g., see [9] or giving a speech, e.g. see [55] should be stressful for most individuals.”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites [39] Hertel PT, Vasquez E, Benbow A, Hughes M (2011) Recollection is impaired by the modification of interpretation bias. Journal of Abnormal Psychology Advance online publication. doi:10.1037/a0023974. source

      Not checked in the free run.

    • cites [9] E Salemink, 2009, Effects of positive interpretive bias modification in highly anxious individuals, Journal of Anxiety Disorders source

      Not checked in the free run.

    • cites [55] H Standage, 2009, Comparing visual and auditory presentation for the modification of interpretation bias, Journal of Behavior Therapy and Experimental Psychiatry source

      Not checked in the free run.

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

    Study design and conductreasoning

    From the current results it is impossible to tell whether the AS-paradigm would still be superior to the homograph paradigm if no imagery instructions had been employed.

    “From the current results it impossible to tell whether the AS-paradigm would still be superior to the homograph paradigm if no imagery instructions had been employed.”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    The AS-paradigm was developed to increase stimuli realism and is a task with high face validity.

    “Overall, the AS-paradigm was developed to increase stimuli realism [7], [64] and is a task with high face validity.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [7], [64]

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationreasoning

    Assuming individuals are confronted with ambiguous situations repeatedly in their daily lives, the cumulative effect of benignly interpreting these situations might result in clinically significant improvements in the long run.

    “assuming that individuals are confronted with ambiguous situations repeatedly in their daily lives, the cumulative effect of benignly interpreting these situations might result in clinically significant improvements in the long run”
    Quote in the paperNot a substantial claim
    • The authors state this as an assumption: Individuals are confronted with ambiguous situations repeatedly in their daily lives.

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    Training effects are more visible when training and test are of increased similarity.

    “So, perhaps training effects are more visible when training and test are of increased similarity.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    The findings for the post-training endorsement of positive versus negative bias may have been solely driven by the use of imagery instructions, which were almost only employed in studies with the AS-paradigm.

    “Alternatively, the findings for the post-training endorsement of positive versus negative bias may have been solely driven by the use of imagery instructions which were (almost) only employed in studies with the AS-paradigm.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    Previous research has shown that, while the endorsement of positive versus negative interpretation style after training showed significant training effects with the AS-paradigm, other interpretation bias assessment tools did not.

    “previous research has shown that, while the endorsement of positive versus negative interpretation style after training showed significant training effects with the AS-paradigm, other interpretation bias assessment tools did not [53]”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [53] E Salemink, 2007, Trained interpretive bias: Validity and effects on anxiety, Journal of Behavior Therapy and Experimental Psychiatry source

      Not checked in the free run.

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

    Premisecitation

    Imagery has long been assumed to have a special link with emotions.

    “Imagery has long been assumed to have a special link with emotions [18].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [18] EA Holmes, 2006, Positive Interpretation Training: Effects of Mental Imagery Versus Verbal Training on Positive Mood, Behavior Therapy source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    The use of imagery in CBM-I has been proposed to increase training effects and primary studies have supported this idea.

    “The use of imagery in CBM-I has been proposed to increase training effects and primary studies have supported this idea [10], [18].”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [10], [18]

      Not checked in the free run.

  • May combine two claims.

    Interpretationown data

    The effect of imagery for post-training endorsement of positive versus negative interpretations may be partly attributable to the effect of sex and vice versa.

    “This may indicate that the effect of imagery for this particular outcome is partly attributable to the effect of sex and vice versa.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductreasoning

    It is difficult to determine which of the two moderators, repeated training sessions and mood-symptoms, drove the effect, although post-hoc analyses including both factors simultaneously supported the robustness of the influence of more training sessions.

    “This makes it difficult to determine which of the two moderators drove this effect although post-hoc analyses including both factors simultaneously supported the robustness of the influence of more training sessions.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    One pioneering study already demonstrated the feasibility and effectiveness of a multi-session internet-based CBM-I.

    “One pioneering study [34] already demonstrated the feasibility and effectiveness of a multi-session internet-based CBM-I.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [34] JO Bowler, 2012, A comparison of cognitive bias modification for interpretation and computerized cognitive behavior therapy: effects on anxiety, depression, attentional control, and interpretive bias, Journal of Consulting and Clinical Psychology source

      Not checked in the free run.

  • May combine two claims.

    Interpretationcitation

    There was no indication for publication bias, so even had there been other unpublished data, these would not have changed the results.

    “as there was no indication for publication bias [31], we cautiously suggest that even had there been other unpublished data, these would not have changed the results.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites [31] SB Morris, 2002, Combining effect size estimates in meta-analysis with repeated measures and independent-group desings, Psychological Methods source

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    Exposure to one or more sessions of benign CBM-I may result in only moderate changes in healthy populations as they already possess a substantial positive interpretation bias to begin with, resulting in a ceiling effect.

    “Exposure to one or more sessions of benign CBM-I may result in only moderate changes in healthy populations as they already possess a substantial positive interpretation bias to begin with, resulting in a ceiling effect.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Healthy populations already possess a substantial positive interpretation bias to begin with.

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

    Premisecitation

    Interpretation biases represent only one potential bias in information processing, which, in turn, represents only one possible cause of emotional disturbance.

    “Second, interpretation biases represent only one potential bias in information processing [64], which, in turn, represents only one possible cause of emotional disturbance [11].”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [64] A Mathews, 2004, On the malleability of emotional encoding, Behaviour Research and Therapy source

      Not checked in the free run.

    • cites [11] LS Hallion, 2011, A meta-analysis of the effect of cognitive bias modification on anxiety and depression, Psychological Bulletin source

      Not checked in the free run.

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

    Premisecitation

    Low mood is a common symptom of mood- and anxiety disorders.

    “Low mood is a common symptom of mood- and anxiety disorders [15].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [15] APA (2000) Quick reference to the diagnostic criteria from DSM-IV-TR. Arlington, VA: American Psychiatric Association.

      Not checked in the free run.

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

    Study design and conductreasoning

    Whether other symptoms like feelings of guilt or rumination can also be influenced by benign CBM-I cannot be answered.

    “whether other symptoms like feelings of guilt or rumination can also be influenced cannot be answered.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductown data

    Most of the correlations between the moderators were small to moderate, reducing the likelihood that the results were confounded by collinearity.

    “Most of the correlations between the moderators were small to moderate reducing the likelihood that the results were confounded by collinearity.”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    Studies are currently emerging showing that training more than one benign cognitive bias, such as interpretation and attention bias, and repeated training sessions of CBM-I result in clinically relevant improvements.

    “Studies are currently emerging showing that training more than one benign cognitive bias (e.g., interpretation and attention bias) [68], [69] and repeated training sessions of CBM-I [20], [34] result in clinical relevant improvements.”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [68], [69]

      Not checked in the free run.

    • cites [20], [34]

      Not checked in the free run.

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

    Premisecitation

    Adolescents are particularly vulnerable to emotional disturbance.

    “as adolescents are particularly vulnerable to emotional disturbance [70]”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [70] Lewinsohn PM, Hops H, Roberts RE, Seeley JR, Andrews JA (1993) Adolescent psychopathology: I. Prevalence and incidence of depression and other DSM-III-R disorders in high school students. Journal of Abnormal Psychology 102.

      Not checked in the free run.

  • May combine two claims.

    Study design and conductreasoning

    The current results depict the effects of the moderators above and beyond the effect of paradigm but not above and beyond the presence of any other moderators.

    “Therefore, the current results depict the effects of the moderators above and beyond the effect of paradigm but not above and beyond the presence of any other moderators.”
    Quote in the paperSubstantial claim
105 of 105

  • The paper relies on this without support.

    CBM-I was first introduced in 2000.

    “To the authors’ knowledge, CBM-I was first introduced in 2000; additional searches for the years 1992 through 1999 were conducted to ensure that no earlier and possibly less popular accounts were missed.”

  • The paper relies on this without support.

    It is currently unclear whether one of the various CBM-I paradigms used is more effective than others.

    “It is currently unclear whether one of the various CBM-I paradigms used (i.e., homograph, AS, and WSAT) is more effective than others.”

  • The paper relies on this without support.

    The administration of feedback about response accuracy has not yet been systematically tested.

    “However, to the best of the authors’ knowledge, this has not yet been systematically tested.”

  • The paper relies on this without support.

    Additional training characteristics such as mode of stimuli presentation and ratio of training items have not yet been systematically investigated.

    “Additional training characteristics, such as the mode of stimuli presentation (visual or auditory), and the ratio of training items in the benign training direction to the total number of presented items (i.e., does the fraction of benign items matter?) also frequently vary between CBM-I studies but have not yet been systematically investigated.”

  • The paper relies on this without support.

    Questions remain on the degree to which CBM can actually ‘correct’ biases and reduce symptomatology.

    “While recent individual studies appear promising, questions remain on the degree to which CBM can actually ‘correct’ biases and reduce symptomatology.”

  • The paper relies on this without support.

    The emotional challenges used in the included studies were universally stressful and not under the direct influence of interpretation biases.

    “For example, watching stressful videos [39], trying to solve an unsolvable anagram task under time pressure ‘knowing’ that ‘most people have no problems solving it in time’, e.g., see [9] or giving a speech, e.g. see [55] should be stressful for most individuals.”

  • The authors state this as an assumption.

    Individuals are confronted with ambiguous situations repeatedly in their daily lives.

    “assuming that individuals are confronted with ambiguous situations repeatedly in their daily lives”

  • The paper relies on this without support.

    Interpretation biases only start to play a role in reducing mood when people are confronted with ambiguous situations in their own life.

    “First, interpretation biases only start to play a role in reducing mood when people are confronted with ambiguous situations in their own life.”

  • The paper relies on this without support.

    Women are more vulnerable to develop mood symptoms.

    “particularly women (who are more vulnerable to develop mood symptoms)”

  • The paper relies on this without support.

    Healthy populations already possess a substantial positive interpretation bias to begin with.

    “as they already possess a substantial positive interpretation bias to begin with”
10 of 10

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?

  • A greater number of training sessions and instructions for participants to imagine themselves in the social scenarios produce larger effects on social anxiety disorder symptoms.

    “While the effect on social anxiety symptoms has not always been greater than control conditions, a greater number of training sessions and instructions for participants to imagine themselves in the social scenarios have been shown to produce larger effects on social anxiety disorder symptoms (Menne-Lothmann et al., 2014).”

    From Acceptability and Co-Development of an Online Cognitive Bias Modification Intervention for Emerging Adults With Hazardous Alcohol Use and Social Anxiety: A Mixed Methods Study (2020) · cites this paper as Menne-Lothmann et al., 2014

    The Results section reports that imagery instructions and additional training sessions were associated with larger decreases in negative mood. It does not establish larger effects on social anxiety disorder symptoms specifically, or limit the imagery instructions to imagining oneself in social scenarios. The paper also cautions that repeated sessions were closely associated with participants’ symptom status. (LLM, conf=0.88)

    Experimental — accuracy not yet measured

  • Research evidence has demonstrated the efficacy of CBM-I in reducing pathological beliefs and symptoms within the affective disorder spectrum.

    “Research evidence has demonstrated the efficacy of CBM-I in reducing pathological beliefs and symptoms within the affective disorder spectrum [10].”

    From A qualitative study of the acceptability of cognitive bias modification for paranoia (CBM-pa) in patients with psychosis (2019) · cites this paper as [10]

    The Discussion’s Conclusions report that benign CBM-I increased positive interpretation bias and decreased negative mood, but note that effects did not consistently differ from control training. The Discussion’s Limitations and future directions section says the effect on low mood was small and that effects on other symptoms, such as guilt or rumination, could not be determined. The paper therefore does not establish efficacy in reducing pathological beliefs and symptoms across the affective disorder spectrum. Page numbers are unavailable in our copy. (LLM, conf=0.88)

    Experimental — accuracy not yet measured

  • Meta-analytic findings suggest that IBM training has a small effect on reducing negative mood.

    “Metaanalytic findings suggest that IBM training has a small effect on reducing negative mood (Cristea, Kok et al., 2015;Hallion & Ruscio, 2011;Menne-Lothmann et al., 2014).”

    From Anxiety and attention to threat: cognitive mechanisms and treatment with attention bias modification (2016) · cites this paper as Menne-Lothmann et al., 2014

Self-check

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

  • No-training effect is not centered in its reported confidence interval

    Minor

    The reported 95% confidence interval has a midpoint of 0.35, not the reported effect size of 0.28. This is inconsistent with a symmetric confidence interval for the estimated effect.

    Reported 0.28 · recomputed 0.35

  • Mood-analysis session counts exceed the reported sample total

    Minor

    The session-frequency counts for the benign-training mood analysis sum to 48 samples, but Table 2 reports 47 benign samples for that outcome.

    Reported 47 · recomputed 48

2 of 2

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Related

Authors
Menne-Lothmann, C., Viechtbauer, W., Höhn, P., Kasanova, Z., Haller, S. P., Drukker, M., van Os, J., Wichers, M., & Lau, J. Y. F.
Affiliations
  • Maastricht University (NL)
  • University of Oxford (GB)
  • King's College London (GB)
ORCID
Venue
PLoS ONE
Publisher
Public Library of Science (PLoS)
Published
2014-06-26
Volume
9
Issue
6
Pages
e100925
ISSN
1932-6203 (ISSN-L)
DOI
10.1371/journal.pone.0100925
PMC
PMC4072710
OpenAlex
W2036630076
Semantic Scholar
15404a3bbb3ea3729574533942385823a36d63c1
Cited by
235
References
75
Citations by year
2026: 14 · 2025: 10 · 2024: 13 · 2023: 26 · 2022: 14 · 2021: 30 · 2020: 27 · 2019: 24 · 2018: 16 · 2017: 19 · 2016: 16 · 2015: 21 · 2014: 5
FWCI
2.75
Type
journal article
Language
en
Open access
Licence
Locations
  • PLoS ONE (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Research Publications (Maastricht University) (publishedVersion, cc-by)
  • RePEc: Research Papers in Economics (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
  • Figshare (submittedVersion, cc-by)
  • Research Portal (King's College London) (submittedVersion, cc-by)
  • Research Publications (Maastricht University) (submittedVersion, other-oa)
Funders
Universiteit Maastricht
Topics
  • Social Sciences › Psychology › Experimental and Cognitive Psychology › Anxiety, Depression, Psychometrics, Treatment, Cognitive Processes
  • Social Sciences › Psychology › Clinical Psychology › Psychological Treatments and Assessments
  • Social Sciences › Psychology › Clinical Psychology › Psychotherapy Techniques and Applications
Keywords
cognitive bias modification for interpretation, interpretation bias, negative mood, emotional reactivity, imagery instructions, gender differences, emotional dysregulation
MeSH terms
Affect, Humans, Cognitive Behavioral Therapy, Outcome Assessment, Health Care

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 2026-10-09.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-09.

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