Objective: This study reports on the psychometric properties of a brief, self-administered screening questionnaire, the Columbia SuicideScreen[R] (CSS), intended to identify high school students at risk for suicide. Method: Seventeen hundred twenty-nine 9th- to 12th-grade students completed the CSS and Beck Depression Inventory during school hours in 1991 to 1994. Three hundred fifty-six students who screened positively and 285, group matched on age, gender, and ethnicity, who screened negatively were examined on the Diagnostic Interview Schedule for Children (DISC), version 2.3, to assess validity. The DISC-based suicide risk criterion was suicidal ideation or prior suicide attempt and a DSM-III-R diagnosis of major depression or dysthymia or substance use. Test-retest reliability was assessed in a subsample of 85. Results: The most balanced algorithm had a sensitivity of 0.75, specificity 0.83, and positive predictive value 16%. Suicidal ideation and prior attempt item reliabilities ([kappa]) were 0.48 and 0.58, respectively. Eight-day test-retest reliability for the most balanced scoring algorithm was 0.32. Conclusions: The CSS demonstrated good sensitivity and reasonable specificity identifying students at risk for suicide. A second-stage evaluation would be needed to reduce the burden of low specificity.