Over Fifteen Years of Confusion: How the Big Five Inventory Mismeasures Openness

Abstract

This study reports that the openness subscale of the 10-item Big Five Inventory is affected by a linguistically ambiguous item: ‘I see myself as someone who has few artistic interests.’ Across nine datasets, responses to this item showed weak or inconsistent alignment with the other openness item, unlike patterns observed for the other Big Five dimensions. The authors tested a revised wording using ‘hardly any artistic interests,’ which improved inter-item correlations, test-retest reliability, criterion validity, and statistical power. Simulation analyses showed that the original item can reduce power to detect true changes in openness, and additional simulations indicated that regression-based adjustment methods can partially reduce bias in existing datasets. The findings call for re-evaluation of prior openness findings and more careful wording in future personality measurement.