The carbon dioxide (CO(2)) challenge paradigm has been useful for modeling panic in the laboratory. While showing promise as a technique able to promote a better understanding of the etiology of panic disorder (PD), this goal has been impeded by the lack of standardization of the challenge methodology and by uncertainty concerning the optimal definition and assessment of laboratory panic. The purpose of this paper is to highlight the impact of method variance on laboratory findings and to present recommendations for future challenge research. We begin by reviewing studies that have employed CO(2) as a stimulus for panic provocation focusing on the status of key methodological parameters between the studies and the relationship of these parameters to findings. We then make pragmatic and theoretically-based recommendations concerning approaches to methodological standardization, the establishment of a valid laboratory panic definition and the desirability of using of additional outcome measures. We conclude that although further work is needed to improve the CO(2) challenge laboratory model of panic, this paradigm can play an important role in understanding the psychopathology of PD.