A reply to “Further evidence for the bidimensionality of the components model of addiction: A reply to Amendola (2023)”

Fournier and colleagues have recently published a reply to my commentary on their original analysis. The authors disagreed with the conclusions of the commentary and noticed that confirmatory factor analysis of the bidimensional data-driven model was not tested for three of the four independent samples. However, the authors did not disagree with the concerns raised by the commentary, particularly on the weakness of questioning the validity of a model using a brief scale with only one item per component in participants from the general population recruited online. The present reply clarifies that the necessary information on not performing some confirmatory factor analysis was already transparently highlighted in the commentary. The results of exploratory factor analysis did not support further testing. Additionally, the scree plot of eigenvalues and parallel analysis were used to decide on the optimal number of factors to retain in the first stage of exploratory factor analysis. In this reply, the estimation of the optimal number of dimensions was advanced benefiting from the approach of exploratory graph analysis, a very accurate technique according to recent studies. The results of this further statistical analysis in psychometric networks were consistent with results from the previous commentary using scree plot and parallel analysis in demonstrating a stable one-dimension solution of the 6-item Bergen Social Media Addiction.### This reply was submitted to Addictive Behaviors on November 23, 2023, and to the partner journal Addictive Behaviors Reports on November 24, 2023. ###