The goal of the present meta-analysis was to ascertain the average effect size of Attention Bias Modification as a treatment for anxiety and depression, as well as to examine the effect of moderators on the intervention (i.e., clinical vs subclinical samples, computer vs. internet training, location of training, and age of sample). Measures of anxiety and depression were analyzed combined and independently, and for each outcome both between- and within-group effect sizes were addressed. Analyzes demonstrated that compared to no-treatment control groups, samples that received the intervention had significantly reduced symptom scores on measures of anxiety and depression when measures were considered together (d = .18). When considered independently, only reduction in anxiety scores remained significant (d = .24), though reduction in depression scores trended toward significance (d = .18, p = .05). The results were similar for within-group comparisons. All three outcome analyses (combined, anxiety, and depression) showed significant reduction in symptom scores from pre- to post-test (d = 1.00, d = 1.15, d = 1.05 respectively). Age, method of training (computer vs internet), and location of training were not shown to be a significant moderators of the effect. Samples using participants with clinical symptoms did not differ significantly from those with subclinical symptoms for between-group analyses, however, for within-group comparisons, samples using participants with clinical symptoms had significantly larger effects than those with subclinical symptoms when considering measures together (d = 1.34 vs d = .60) or measures of anxiety (d = 1.47 vs d = .64). The results indicate Attention Bias Modification shows promise as an intervention, though future research ought to investigate effects when compared to other forms of treatment (e.g., CBT), or as an adjunct to current treatments.
Files are restricted to Pacific University. Sign in to view.