To authenticate to the full-text of this article, please visit this link: http://dx.doi.org/10.1111/j.1467-9566.2010.01260.x Byline: Christophe Vanroelen (1,2), Katia Levecque (3), Guy Moors (4), Fred Louckx (1) Keywords: occupational stress; modified LISREL; Belgium; social class; socio-economic health inequalities Abstract: Abstract In this study, focus is on the mechanisms linking credentialed skills and social class relations to five dimensions of occupational stressors and three self-reported health outcomes: persistent fatigue, musculoskeletal complaints and emotional wellbeing. We test for direct health associations of skills/class. Moreover, indirect health associations of skills/class, through differential exposure to occupational stressors and effect modifications of the occupational stressors, are tested. A modified LISREL analysis is applied to a representative cross-sectional sample of 11,099 Flemish wage-earners. The direct health effects of credentialed skills/class are limited, but they are clearly indirectly related to the self-reported health outcomes through differential exposure to occupational stressors. The indirect mechanisms point to both reinforcing and moderating effects on socio-economic health inequalities. Two cases of effect modification are also observed: social class modifying the association between control and persistent fatigue; and skills affecting the association between the quality of social relations and emotional wellbeing. Differential exposure to occupational stressors is a crucial mechanism linking skills/class to socio-economic health inequalities. Direct effects and effect modification of class/skills are relevant, but of limited importance. One of the effect modifications found suggests that a specific focus on contradictory class positions might be warranted in future research. Author Affiliation: (1)Department of Medical Sociology, Vrije Universiteit Brussel, Belgium (2)Health Inequalities Research Group (GREDS), Employment Conditions Knowledge Network (EMCONET), Universitat Pompeu Fabra, Barcelona, Spain (3)Research Foundation Flanders and Department of Sociology, Ghent University, Belgium (4)Department of Research Methods and Techniques, University of Tilburg, The Netherlands Article note: Address for correspondence: Christophe Vanroelen, Department of Medical Sociology, Vrije Universiteit Brussel, Laarbeeklaan 103, B-1090 Brussels, Belgium, e-mail: cvroelen@vub.ac.be