The goal of the DSM‐IV panic disorder field trial was to provide an empirical basis for choosing between alternate proposals (DSM‐III‐R and proposed DSM‐IV) for the diagnostic threshold for panic disorder, in particular the number and frequency of panic attacks required for diagnosis. The two criteria sets were compared with respect to their ability to identify individuals whose panic attacks were associated with distress, impairment, or help‐seeking. Subjects were a convenience sample screened in three geographically diverse primary care clinics for presence (past 6 months) or absence (lifetime) of panic attacks. Each underwent a clinician‐administered semistructured interview which included assessment of panic frequency, panic‐related impairment, psychiatric diagnosis, health services utilization, and medical illness. Self‐perceived health‐related quality of life was assessed using the Medical Outcome Study SF‐36 Health Survey Questionnaire. Although both proposals diagnosed the same proportion of panic‐impaired individuals, they were not completely overlapping. Twenty percent of subjects diagnosed by each criteria set were excluded by the other. Subjects who had been excluded by the DSM‐III‐R but included by the DSM‐IV proposal were those with fewer than 4 attacks in 4 weeks who also denied worry about the “next” attack. Broadening the worry criterion to include concerns about the health implications of attacks enabled diagnosis of this group. Subjects who met DSM‐III‐R, but not the proposed DSM‐IV criteria, had 4 attacks in 4 weeks but denied any panic related worry. Modification of the DSM‐IV proposal to include a month of worry or “a significant change in behavior related to the attacks” allowed inclusion of this group in the diagnostic category. These data suggest that the finalized DSM‐IV panic disorder criteria will diagnose a greater proportion of individuals whose panic attacks are associated with impairment without inflating the diagnostic category or significantly reducing specificity. Anxiety 2:157–166 (1996). © 1996 Wiley‐Liss, Inc.