This commentary by In-Sue Oh and Huy Le, published in Industrial and Organizational Psychology, examines how researchers determine whether newly proposed constructs are genuinely distinct from existing ones. The authors evaluate a proposed checklist for addressing construct proliferation, focusing on its suggestion that a construct-level correlation of .85 or higher may indicate redundancy. They argue that this cutoff should not be applied universally because its meaning depends on the reliability and content of the constructs being compared. Estimating accurate construct-level correlations also requires correcting for multiple sources of measurement error, which involves complex and costly research designs. The authors further show that even a very high correlation between two constructs does not necessarily mean they are redundant if they relate differently to other variables. They recommend evaluating constructs across their broader network of relationships and relying on large samples, independent replications, or meta-analytic evidence. These recommendations can help researchers, reviewers, and editors make more accurate decisions about whether a new construct offers meaningful theoretical and empirical value.
The devil is in the details: Methodological nuances and challenges in evaluating construct redundancy
