Do causal attributions serve the need to protect and / or enhance self-esteem? In a recent review, Miller and Ross (1975) proposed that there is evidence for self-serving effect in the attribution of success but not in the attribution of failure; and that this effect reflects biases in information-processing rather than self-esteem maintenance. The present review indicated that self-serving effects for both success and failure are obtained in most but not all experimental paradigms. Processes which may suppress or even reverse the self-serving effect were discussed. Most important, the examination of research in which self-serving effects are obtained suggested that these attributions are better understood in motivational than in information-processing terms.