Although the media are regularly charged with bias, empirical evidence of media bias is variable. The aim of the current research was to explore the utility of an intergroup perspective to understanding media bias as it emerges in the context of intergroup conflict. Content analysis was conducted on accounts of ongoing Christian-Muslim conflict in Ambon, Indonesia, as reported in both Christian and Muslim newspapers. This revealed the operation of a 'naming bias', whereby both Christian and Muslim newspapers were more likely to explicitly name the religious outgroup as perpetrators of intergroup conflict than they were to attribute responsibility to their own group. The prevalence of this bias was, however, asymmetrical across the two groups: it was pronounced in the Muslim newspaper but minimised in the Christian one. This pattern was evident in a general sample of media reports, and in a sample of matched reports in which the same incident was covered by both papers. The naming bias and its variable operation is explained with reference to social psychological theorising about intergroup dynamics.