Some limitations of the current study should be noted. First, the use of a cross-sectional design precludes us from determining causality. Although we posited that the mediational pathway goes from AS to negative urgency to negative reinforcement-related smoking expectancies, it is instead possible that AS mediates the relationship between negative urgency and negative reinforcement-related smoking expectancies, which is a possibility that we were unable to rule out altogether with reverse mediation analyses. However, reverse mediation analyses were not significant for smoking abstinence expectancies when adjusting for covariates. Furthermore, we see this possibility as being unlikely because AS is conceptualized as preceding and influencing the development of anxiety (see McLaughlin and Hatzenbuehler, 2009 for a review of the evidence), whereas negative urgency is the tendency to act impulsively in response to negative affective states, such as anxiety, and should therefore follow the development of negative emotions. One question that we were unable to address is the temporal order of smoking behavior and expectancies. It is possible that the relation of AS to smoking expectancies exists prior to smoking initiation and may
