The dexamethasone suppression test (DST) can differentiate between endogenous and nonendogenous depression. Similarly, EEG sleep patterns can differentiate primary from secondary depression, and this technique has also been used to make the endogenous-nonendogenous discrimination. However, a number of physiological variables associated with this diagnostic distinction may also affect the DST results and sleep architecture. With the use of multivariate statistical procedures, we found that although age and weight loss affect the results of both tests, both the DST and sleep EEG differentiate endogenous from nonendogenous depression when these variables are taken into account. Severity of illness affected both proposed diagnostic markers, but did not account for the differences between diagnostic groups, alone or when added to the physiological variables. The DST was more sensitive in unipolar than in bipolar endogenous depression, but there were no significant differences in the sleep of unipolar and bipolar patients.