Intraparty Hostility: Social Identity, Subidentity, and the Hostile Media Effect in a Contested Primary

The hostile media effect (HME) has generally been tested in terms of in-groups and out-groups, with a neutral story in between. This ignores the nature of many social groups as comprising subgroups, often but not always sharing feelings of connectedness and purpose. In cases when bounded subgroups are at odds with one another, HME provides little guidance. A contested partisan primary provides such a case. This study takes identity centrality, candidate favorability, and perceived social network homogeneity as measures of partisanship and involvement, hypothesizing relationships between each and perceived bias against one's candidate and party. Findings show that markers of candidate-focused social identity predict greater perceived bias against one's candidate during the 2016 primary season, while party-focused identity fails to predict perceived bias against one's party. This suggests that candidate support identity overrides plain partisanship during primaries, supporting concern that a heated primary might damage general election party unity. Subsequent postconvention findings suggest that the salience of candidate-focused identity fades, while homogeneity of one's network regarding party support helps to make perceived hostility toward one's party identity more salient. However, as campaigns become more candidate-centered, the contestation between nested candidate and party identities may grow fiercer.