There is some debate whether job strain or working hours is more prominent in explaining work-family conflict. We tested a multi-group structural equation model and the results suggested the model was equally applicable to academic and administrative staff employed at an Australian regional university. After controlling for demographic and work-related factors the main predictors of work-family conflict were: job strain, total work hours, job satisfaction, employment as an academic and having dependent children. Social support was negatively related with work-family conflict but the association was not significant. Despite greater job control, academics reported greater job strain and work hours. The results suggest that strategies aimed at decreasing job strain and work hours may reduce the extent of work-family conflict.