OBJECTIVE. To examine whether public support for government intervention to address health disparities varies when disparities are framed in terms of different social groups. METHOD. A survey experiment was embedded in a public opinion poll of Wisconsin adults. Respondents were randomly assigned to answer questions about either racial, economic, or education disparities in health. Ordered logit regression analyses examine differences across experimental conditions in support for government intervention to address health disparities. RESULTSs. Health disparities between economic groups received the broadest support for government intervention, while racial disparities in health received the least support for government intervention. These differences were explained by variation in how respondents perceived and evaluated health disparities between different social groups. CONCLUSION. Efforts to garner public support for policies aimed at eliminating health disparities should attend to the politics of social diversity, including the public’s disparate perceptions and evaluations of health disparities defined by different social groups.