A research paradigm involving real-world group memberships and experimental manipulations of relative socio-economic success is used to support a system justification theory of outgroup favoritism among low status groups (Jost & Banaji, 1994). Major findings from Studies 1 and 2 were that high success groups exhibited ingroup favoritism and low success groups exhibited outgroup favoritism, expecially when: (a) perceived legitimacy of the success differences was high rather than low; (b) dimensions of comparison were relevant to the success differences rather than irrelevant; and (c) participants were not asked to explain the success differences rather than when they were. Content analysis indicated that open-ended explanations for the success differences generated by high success in Study 2 refelcted ingroup favoritism, whereas explanations generated by low success groups reflected outgroup favoritism. When perceived legitmacy and ingroup identification were manipulated independently in Study 3, low success groups showed outgroup favoritism when perceived legitimacy was high (but not low) and on relevant (but not irrelevant) attributes. In Study 4, these effects were replicated in the context of a naturally occuring status difference. Taken as a whole, the evidence provides strong empirical support for a system justification approach to sterotyping and intergroup relations.