This study reports on research into the relationship between absence from work and smoking. A key feature of the study is the data, which come from the National Health Survey (NHS) undertaken in 1989/90 in Australia. Involving responses from more than 54,000 individuals, the NHS provides what is almost certainly the largest and most comprehensive data set available in the world today containing information on both absence and smoking behaviour. Moreover, the data permit controls to be applied for a large number of influences thought to have some bearing on work attendance. Logit models of absence incidence over a two week period are estimated, and smoking is consistently found to have a large and significant impact on absence. This impact, however, is not consistent across the sexes. The probability of a male smoker being absent from work is estimated to be 66% greater than that for a male who has never smoked. For females, the corresponding figure is just 23%. The findings also suggest that it is important to distinguish ex-smokers from other non-smokers, with the incidence of absence among ex-smokers being almost as high as that for current smokers. Finally, no evidence was uncovered to suggest that absence varied with the actual quantity of tobacco smoked, as measured by both the number of cigarettes smoked and estimated daily nicotine and tar intakes.