We assessed the reliability and validity of the Turkish version of the seven-item Panic Disorder Severity Scale (PDSS). We recruited 174 subjects, including 104 with current DSM-IV panic disorder with (n=76) or without(n=28)agoraphobia, 14 with a major depressive episode, 24 with a non-panic anxiety disorder, and 32 healthy controls. Assessment instruments were Panic Disorder Severity Scale, Panic and Agoraphobia Scale, both the observer-rated (P&Ao) and self-rating (P& Asr); Clinical Global Impression Scale (CGI); Hamilton Anxiety Scale, and Beck Depression Inventory. We repeated the measures for a group of panic disorder patients (n = 51) after 4 weeks to assess test-retest reliability. The internal consistency (Cronbach's alpha) of the PDSS was .92-94. The inter-rater correlation coefficient was .79. The test-retest correlation coefficient after 4 weeks was .63. In discriminant validity analyses, the highest correlation for PDSS was with P&Ao, P&Asr (r=.87 and.87, respectively) and CGI (r=.76) and the lowest with Beck Depression Inventory (r=.29). The cut-off point was six/seven, associated with high sensitivity (99%) and specificity (98%). This study confirmed the objectivity, reliability and validity of the Turkish version of the PDSS.