General Behavior Inventory identification of unipolar and bipolar affective conditions in a nonclinical university population.

Validated the General Behavior Inventory (GBI), revised to identify unipolar as well as bipolar affective conditions, in a nonclinical sample (n = 201) against naive, interview-derived diagnoses. For bipolar and unipolar conditions, respectively, the GBI had high positive (.94, .87) and negative (.99, .93) predictive power with the effect of prevalence considered, adequate sensitivity (.78, .76), high specificity (.99, .99), and adequate selection ratios for sampling of affective and nonaffective subjects from nonclinical populations for research purposes. The utility of the GBI in several different research contexts is discussed.