The rejection–identification model (RIM; Branscombe, Schmitt, & Harvey, 1999) is supported by a number of previous studies (e.g., Schmitt, Branscombe, Kobrynowicz, & Owen, 2002; Schmitt, Spears, & Branscombe, 2003). This suggests that rejection by an outgroup can lead minority group members to identify more with their ingroup, thereby buffering them from the negative effects of discrimination. However, contradictory findings have been produced by other research (e.g., Eccleston & Major, 2006; Major, Quinton, & Schmader, 2003; McCoy & Major, 2003; Sellers & Shelton, 2003), suggesting that the relationship between rejection and identification is far from being completely understood. In the present study, we followed a cohort of 113 international students for a period of 2 years. The study sought to extend the previous work in two important ways. First, it examined the RIM within a longitudinal perspective. Second, building on important work on the multidimensionality of social identification (e.g., Ellemers, Kortekaas, & Ouwerkerk, 1999; Jackson, 2002), it tested the RIM using a three-dimensional approach to group identification. Results supported the predictions of the RIM and indicated that perceived discrimination causes minority group identification and not the reverse. The multidimensional approach also served to reveal a specific effect of discrimination on the cognitive components of identification.