We used meta-analytic techniques in an attempt to clarify the strength and direction of the association between smoking status and personality, which narrative reviews have indicated remains a largely inconsistent literature. Included were cross-sectional studies that reported personality data for healthy, adult smokers and nonsmokers using measures of personality traits derived from Eysenck's tripartite taxonomy of human personality. Of the 25 studies that contributed to the meta-analysis, 22 reported data on smoking status and extraversion and 22 reported data on smoking status and neuroticism. Meta-analysis using a fixed-effects framework indicated a significant difference between smokers and nonsmokers on both extraversion (p<.001) and neuroticism (p<.001) traits, which remained significant when a random-effects framework was used to accommodate significant between-study heterogeneity. These data from cross-sectional observational studies published between 1972 and 2001 indicate that both increased extraversion and increased neuroticism are associated with an increased likelihood of being a smoker rather than a nonsmoker, although in both cases the effect sizes indicated by the meta-analysis were small. We found no evidence that the strength of these associations varied with year of publication.