Projection in the face of centrism: voter inferences about candidates’ party affiliation in low-information contexts

When are voters more likely to project their own political position onto a candidate for office? We investigate this question by examining the assumed partisanship of a (self-declared) centrist politician, using data from an original survey experiment fielded in Canada, the UK, and the US. In doing so, we build on the Social Categorization Model as well as recent US-focussed political science research on projection and in-group/out-group racial divides – extending our analysis to incorporate racial and class similarities/differences across three countries where these divides likely vary in salience. We thus seek to: (1) contribute to research on the inferences citizens draw in nonpartisan elections and low-information contexts generally; and (2) highlight some potential methodological complications of using partisanship-less candidates in vignette experiments. Results suggest that even in the face of a self-declared centrist, voters from across the political spectrum tended to assume shared partisanship in Canada, the UK, and the US. Examining projection by in-group/out-group divisions indicated that class appears to shape projection across all three countries, but that the racial divide only mattered in the US. Finally, we also find evidence of counterprojection toward out-group members – but once again only in the American context.