Is general intelligence responsible for differences in individual reliability in personality measures?

Abstract

One possible hypothesis for personality differentiation is the higher reliability of high-ability individuals in typical response measures. This differential reliability has been explained as resulting from different verbal abilities as a consequence of the difficulties that low-ability individuals have in understanding items, or as the effect of response bias, or due to higher precision in the answers of high-ability individuals. The lack of an estimation of individual reliability has made it difficult to test these hypotheses. However, recent psychometric advances have made it possible to measure person reliability and thus address the issue. The present study analyses the relationships between person reliability measures and the response bias of different personality measures in measurements of intelligence in a sample of 532 adolescents. The results show that person reliability is more closely related to general intelligence than to specific abilities and that the results for low-ability individuals cannot be explained by verbal deficits or by higher levels of acquiescence or social desirability. The differential reliability of measures across ability levels therefore seems to be related to higher levels of traitedness in high-ability individuals, i.e. traits are represented in them with greater strength and clarity.