The author analyzes the relationship between unionization and absences due to sickness in Norway using register data covering the entire population employed in the private sector during 2003 to 2007. A few previous studies have shown much higher sickness absence among unionized than nonunionized employees in Canada, the United States, and the United Kingdom. The author shows this is also true for Norway, although primarily for employees in lowergrade jobs. The differential persists even when employees in the same detailed job category within the same establishment are compared using fixed effects models. Thus, differences between more or less unionized establishments, or differences in working conditions between members and nonmembers, fail to explain the differential. Analyses using individual fixed effects indicate that up to half of the union-nonunion differential in absence is due to selection on stable individual characteristics.