What it’s about

In this paper, we investigate the psychometric assumptions behind very short scales and compare reflective versus formative modeling perspectives using the Ten-Item Personality Inventory.

Abstract

Very short scales are widely used when testing time is limited, but they raise strong modeling assumptions that are not always scrutinized. This article evaluates whether reflective latent-variable models are appropriate in such settings and presents formative-model proofs of concept with the Ten-Item Personality Inventory. The paper highlights implications for construct representation, model fit, and score interpretation when scale length is extremely constrained. It argues that modeling choice should be theory-driven and psychometrically explicit rather than automatic.

How to access the paper

Go to the journal page for this article here.

Download the PDF here.