Moscovici's model of minority influence and Hollander's model of idiosyncrasy credit were compared in an experiment conducted with I6 discussion groups composed of four male undergraduates (N = 64) and a confederate. Their task was to rank-order unanimously 10 proposals to remediate urban problems, to which the confederate attempted to have one more added, thus introducing a norm change. Two variables were manipulated: Status of the innovator (minority versus elected leader) and Ideological criterion for innovation (avant-garde versus reformist). It was predicted that the consistent behavioural style of the minority would make him more influential under the avant-garde condition than under the reformist condition whereas the idiosyncrasy credit of the leader would make him more influential under the reformist condition. Although the adoption hypotheses failed to be supported, the evaluations of the confederate were consistent with both the Moscovici and Hollander models. While the minority con federate's action gained him visibility and distinctive attributions of determination and assurance, the leader con federate's initiative cost him his competence and cooperation credits. Strategies devised to adopt the innovation or to reject it (five groups versus four) pointed to the importance of interactive networks within the majority and to the decisive role played by group leaders for the innovator's influence to be exerted. These findings lead to speculate that Hollander's theorizing could be complementary to Moscovici's to account for the diffusion of minority influence.