Levels and subtypes of depression should not be overlooked in research on neuroticism and cortisol.

DeSoto and Salinas (in press) examined the relaionship between neuroticism and the activity of the ypothalamic–pituitary–adrenal (HPA) axis. The authors observed positive correlation between neuroticism and the level of corisol in men (r = .29; n = 47) and a negative correlation between euroticism and the level of cortisol in women (r = −.18; n = 109). n this basis, the authors concluded that personality affects the ctivity of the HPA axis in a sex-specific manner. In our view, the nternal validity of the authors’ study is threatened by the absence f assessment and statistical control of depression. Neuroticism and depression are positively and strongly corelated variables (e.g., Jylha and Isometsa, 2006). This suggests hat depression is a key potential confounder in neuroticism esearch, requiring systematic consideration. In addition, depresion is known to be associated with alterations of the HPA xis, in a subtype-dependent way. Thus, melancholic depression as been related to hypercortisolism whereas atypical depresion has been related to hypocortisolism (Gold and Chrousos, 002; Hellhammer and Hellhammer, 2008). Interestingly, because elancholic depression has been found to be more prevalent n men and atypical depression in women (Blanco et al., 2012; ildebrandt et al., 2003; Schuch et al., 2014), this state of affairs s directly relevant to the sex-focused questioning of the authors. epression may actually account for the differences in cortisol levls imputed to neuroticism. Controlling for depression would have been all the more imporant that the associations between neuroticism and cortisol levels re relatively weak in the authors’ study, with neuroticism explainng about 3% of the variance in women’s cortisol level, and about % in men’s. The inclusion of depression in the tested models might ave rendered the reported associations between neuroticism and ortisol level statistically nonsignificant, leading the authors to raw different conclusions. Because of the noncontrol of key confounding variables, inconlusive studies are accumulating in neuroticism research. The orrection of this trend should be high on the agenda of neuroticism esearchers.