Background: The aim of the study is to investigate whether the gender composition in workplaces is related to long-term sickness absence (LSA). We start off with Kanter’s theory on ‘‘tokenism,’’ suggesting an increased risk of stress among minority groups (tokens), which, in turn, might increase the risk of ill health and LSA. Methods: The dataset consists of information obtained from the Swedish level of Living Survey (LNU) and the Swedish Establishment Survey (APU), linked to register-based data from the Swedish Social Insurance Agency. The longitudinal data is representative for the Swedish population and consists of 496 women and 566 men, aged 20—55 at baseline. Our study group consisted of employed persons in 1991 and we analyze, by means of piecewise constant intensity regressions, the first entry into LSA with a follow-up period of nine years. Results: Compared with women in gender-integrated workplaces, women’s risk of LSA is most elevated at both extremely male-dominated (0—20% females) and extremely female-dominated workplaces (80—100% females), although the result among women in the most male-dominated group did not reach statistical significance at the 5% level. Men’s risk seems less varied by gender composition. Conclusions: The present study suggests that the gender composition in the workplace has an impact on the risk of LSA, especially among women. Our findings lend no support for Kanter’s theory on the effects of being a token. Most likely, women’s and men’s different status positions have an impact on the different associations found.