In the catastrophic misinterpretation model of panic Clark [Behav. Res. Ther. 24(1986)1461] proposes that panic attacks result from the misinterpretation of autonomic arousal stimuli as precursors to a physical or psychological emergency. The model has been widely examined, with many researchers suggesting that this specific cognitive bias is implicated in both the phenomenon of panic, and the aetiology and maintenance of panic disorder. Various research methodologies have provided only partial or inconclusive support for the model as being uniquely associated with panic, and as a cognitive process underpinning the experience of panic. This paper reviews the body of existing evidence and its implications for the model and proposes future research directions. The influence of implicit operational definitions of key terms in the catastrophic misinterpretation literature (e.g. 'catastrophe', 'threat', 'anxiety-related') are examined, and clarifications proposed. Inconsistencies and limitations in the measurement of catastrophic misinterpretation are highlighted, and subsequently developments to measurement instruments are proposed.