Recent work on minority influence has led to a debate about whether majorities and minorities exercise different forms of influence. Nemeth (1986) has argued that consistent minorities induce different cognitive processes than do consistent majorities, with a resulting impact on the quality of the judgments rendered. Two experiments test this theory. In Experiment 1, Ss heard 3 tape-recorded lists of words and learned that either a minority or a majority differed in the category "first noticed." This feedback occurred either once or over 3 trials. When exposure was once, recall was not affected by the source; when it was consistent, Ss exposed to the minority view recalled more words than those exposed to the majority view. In Experiment 2, Ss were exposed to a minority view that was either consistent over time or inconsistent over time. Ss exposed to a consistent minority had better recall than control Ss. Exposure to an inconsistent minority did not improve recall. The results offer support for the Nemeth (1986) formulation.