The effects of group size, group status and trait valence (positive or negative stereotypes of in-and outgroup) on intergroup bias was studied in nation-wide probability samples of majority and minority groups in Finland and Sweden, (N ¼ 2479). Ethnolinguistic vitality was used as a proxy for status. It is argued that the specific history of real-life intergroup relations has to be duly acknowledged when predicting main and interactive effects on intergroup bias in natural contexts. Supporting the predictions made, numerical group size showed a stable main effect; members of numerical minorities showed more bias than members of numerical majorities, regardless of trait valence. While status had no main effect, there was a significant interaction between status and size as well as between status and trait valence: intergroup bias was highest in the high status minority, and low status groups showed less bias than high status groups on negatively valenced traits. In fact, minority members showed the reverse of PNAE. In addition, majority members favoured the outgroup on negatively valenced traits, but favoured their ingroup on positively valenced traits. Different explanations for these results are discussed.