Purpose -The purpose of this paper is to use data from the 2008 and 2012 US Senate elections to examine the relationship between candidate size (obese, overweight, normal weight) and candidate selection and election outcomes. Design/methodology/approach -Using pictures captured from candidate web sites, participants rated the size of candidates in the primary and general US Senate elections. w 2 analyses, t-tests and hierarchical multiple regressions were used to test for evidence of bias against overweight and obese candidates and whether gender and election information moderate that relationship. Findings -Obese candidates were largely absent from the pool of candidates in both the primary and general elections. Overweight women, but not overweight men, were also underrepresented. Supporting our hypothesis that there is bias against overweight candidates, heavier candidates tended to receive lower vote share than their thinner counterparts, and the larger the size difference between the candidates, the larger the vote share discrepancy. The paper did not find a moderating effect for gender or highinformation high vs low-information elections on the relationship between candidate size and vote share. Research limitations/implications -Further research is needed to understand the process by which obese candidates are culled from the candidate pool and the cognitions underlying the biases against overweight candidates. Social implications -Because of the bias against obese political candidates, as much as one-third of the adult US population are likely to be excluded or being elected to a major political office. Originality value -This study is the first to use election data to examine whether bias based on size extends to the electoral process.