OBJECTIVE: The objective of this study was to investigate the impact of psychosocial work environment factors on short and long absence spells. METHODS: Questionnaire data on work environment exposures and registered absence data during 2-year follow up were analyzed with Poisson regression for 1919 employees from the private and public sector. RESULTS: Short spells (1-10 working days) were predicted by low supervisor support, low predictability, and low meaning at work among men and high skill discretion among women. Long spells (>10 days) were predicted by low decision authority, low supervisor support, and low predictability among men and high psychologic demands and low decision authority among women. The variables predictability and meaning at work were developed for this study. CONCLUSION: Specific psychosocial work environment factors have both common and different effects on short and long absence spells. Effects also differ by gender.