Since Moscovici's seminal work on minority in¯uence (cf. Moscovici, 1976, 1985a, 1985b; Moscovici & Faucheux, 1972), a number of studies con®rmed that both majorities and minorities can induce a certain amount of in¯uence (see Chaiken & Stangor, 1987; De Dreu & De Vries, 2001; Maass & Clark, 1984; Moscovici, Mucchi-Faina, & Maass, 1994; Moscovici, Mugny, & Van Avermaet, 1985; PeÂrez & Mugny, 1993; Wood, Lundgren, Ouellette, Busceme, & Blackstone, 1994, for reviews). According to Moscovici (1980), when confronted with majority positions, targets focus on discrepancies between the majority's position and their own, and the in¯uence results from their motivation to reduce such discrepancies (a comparison process). Conversely, when confronted with minorities, targets' attention is focused on the minority's position, and in¯uence results from the careful processing of the situation and the relevant information (a validation process). These distinct processes result in differences in the nature of in¯uence: Majorities obtain more manifest in¯uence (i.e., on immediate, focal or public measures) than latent in¯uence (i.e., on delayed, indirect or private measures) in the form of compliance, while minorities obtain more latent than manifest in¯uence (i.e., conversion). This is a crucial point because the distinction between levels of in¯uence has received little attention in recent in¯uence approaches based on information processing.