Temperament as a predictor of symptomatology in children: addressing contamination of measures

Abstract

Temperament has been conceptualized as an important predictor of children's psychological adjustment. However, even with reliable and valid measures, there is the additional problem of overlapping item content across measures of temperament and symptoms that threatens the interpretability of such associations. This study assessed this possible confounding using both confirmatory factor analysis (CFA) and expert ratings. A number of items from temperament measures of negative and positive emotionality, impulsivity, and attention focusing were shown to overlap with items measuring depressive and conduct problem symptoms. CFAs demonstrated that temperament could be reliably measured after eliminating overlapping items. Negative emotionality and impulsivity showed a positive relation to symptom measures, whereas positive emotionality and attention showed a negative relation to symptom measures. The pattern of associations indicated consistent relations between negative emotionality and depression and between impulsivity and conduct problems. The results show that even after removal of the threat to validity presented by overlap in measures, there continue to be significant, interpretable relations between temperament and symptoms.