A longitudinal study (N = 109) of interschool contact and attitudes was conducted to test Allport's (1954) Contact Hypothesis and Brown and Hewstone's (2005) addendum to it on the moderating role of typicality in the contact-attitude relationship. Three different measures of intergroup attitude were employed, including a new measure of infrahumanisation (Leyens et al., 2000). Support for the hypotheses was found on all three measures in the longitudinal analyses. Quantity of contact with a member of the outgroup was consistently associated with more favourable attitudes towards the outgroup as a whole. Importantly, contact was found to predict attitude but the reverse causal path was not significant. Also, on one measure there was an indication that the effects of contact quality were more beneficial when the contact persons were seen as typical of the outgroup than when they were not. It is concluded that, as originally hypothesised by Allport, contact with members of an outgroup can improve intergroup attitudes, but especially if those people can be seen as representative of their group. Copyright © 2006 John Wiley & Sons, Ltd.