The social identity-intergroup differentiation hypothesis is a hotly debated issue among social identity researchers (Brown, 2000; Turner, 1999); it states that individuals having a stronger in-group identification will perceive greater differences between their in-group and a relevant out-group. This study examines the importance of three factors when testing this hypothesis: the strength and salience of in-group identification, the relevance of the out-group for social comparison, and the relevance of the dimension of social comparison. The hypothesis was examined in relation to the national identity of a sample of Canadian students. Perceptions of the in-group and out-groups were measured at Time 1 (N =171). The same measures were given at Time 2 (N = 77), along with a variety of measures of social identity. It was predicted that this hypothesis would be supported when the dimension of social comparison was of high relevance and only for an important social comparison group (i.e. Americans). Finally, the ability of identity to predict differentiation at another point in time was examined in order to examine the issue of identity salience and stability. Results generally supported the hypotheses and are discussed in relation to prior research and the conceptualization of a minority identity.