Since the TEMPS-A is a self-report questionnaire, we investigated several factors that may have an impact on a subject’s response as potential covariates, namely gender, current state, and illness severity. The sample was approximately 64% female, and weak, yet significant, correlations with gender were observed for the dysthymic and anxious temperaments only, as shown in Table 1. Since subjects may have been euthymic, depressed, or manic at the time of testing, we assessed the impact of a subject’s current state and level of impairment at the time of testing via the Global Assessment of Functioning (GAF) score, which had a mean of 64.3 (±14.7) in our sample and revealed moderately negative correlations (r=−0.20 to −0.25, p<0.001) with all temperaments except hyperthymic. Finally, we evaluated age at onset (AAO) of BD, defined as the age at onset of the first manic or depressive episode, as a proxy for illness severity, since an earlier AAO typically reflects a more severe form of illness. In this sample, earlier AAO was significantly correlated with an increase in the number of episodes and the frequency
