One research question, two meta-analyses, three conclusions: Commentary on "A systematic review with meta-analysis of...

Recently two independent meta-analyses on the efficacy of Cognitive Bias Modification of Interpretation (CBM-I) to reduce aggressive behavior came to different conclusions: Ciesinski et al. (2023) concluded that "CBM demonstrates efficacy for the treatment of aggressive behavior" (Abstract), whereas our research team concluded that "findings show limited support for the efficacy of CBM-I to reduce aggressive behavior" (AlMoghrabi et al., 2023, Discussion). How can similar meta-analyses reach such different conclusions? In this commentary, we raise awareness concerning how 1) seemingly identical research questions can be based on meaningfully different definitions of the intervention and outcomes; 2) intervention efficacy conclusions can depend on outcome assessment type; and 3) the interpretation of underpowered moderator analyses should not depend on statistical significance. We end our commentary with a third, more nuanced conclusion that can reconcile the two disparate conclusions: that current CBM-I is an effective experimental manipulation to modify interpretation biases, but not an effective stand-alone treatment to reduce aggressive behavior.