Studies of the causes of absence from work tend to indicate uniformly that smokers miss more work than non-smokers. Some estimates of this increased smoker absenteeism range as high as 32%, or about 81 million additional lost days of work per year in the US alone. Unfortunately these studies, in the main, employ only a simple means difference test as a statistical basis for their deductions, and hence are incapable of determining whether smokers miss more work because they smoke or for other reasons common to smokers as a group. In this study, we posit a native model of absenteeism and employ Tobit analysis to estimate its response to various determinants using data from the 1968 PSID. We than empoly a Blinder-type decomposition technique to attempt to determine the amount of absenteeism attributable directly or indirectly to smoking behaviour of workers. Our results suggest that smokers miss no more work than non-smokers because they smoke. Rather smokers tend to be younger, heavier drinkers, blue collar workers, etc., and these groups will miss more work regardless of whether or not they smoke.