Skip to main content

The reliability, validity, and utility of brief depressive measures as compared to the Beck Depression Inventory-II

25 July 2003


Sixty-two undergraduate and graduate students completed packets of 6 brief depressive measures including: the Beck Depression Inventory-Short Form (BDI-SF), Patient Health Questionnaire-9 (PHQ-9), Patient Health Questionnaire-2 (PHQ-2), Geriatric Depression Scale-Short Form (GDS-SF), and the Center for Epidemiologic Studies Depression Scale-10-item (CESD-10). The Beck Depression Inventory-II was used as the comparison measure. Internal reliability, concurrent validity, criterion related validity (sensitivity and specificity), and general utility were explored. Internal reliability yielded strong alpha coefficients for all of the measures (BDI-II = 0.91; BDI-SF = 0.87; PHQ-9 = 0.84; PHQ-2 = .79; GDS-SF = 0.78; CES-D-10 = 0.81). Significant concurrent validity was obtained for all measures except the GDS-SF, using bivariate correlations with the BDI-II (PHQ-9, r = 0.89; PHQ-2, r = 0.69; BDI-SF, r = 0.86; CES-D-10, r = 0.83). Sensitivity measurements yielded only two elevated findings (PHQ-9 = 100.00%; BDI-SF = 88.97%), and specificity measures demonstrated similarly elevated results across the measures (PHQ-9 = 98.13%; PHQ-2 = 98.13%; GDS-SF = 92.40%; BDI-SF = 84.91 %). Overall utility rankings, based on reliability, validity, sensitivity, specificity, and completion time, highlighted the PHQ-9 as the best replacement for the BDI-II for use as a measure of severity or case-finding instrument.


Files are restricted to Pacific University. Sign in to view.