Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
This study tests for taste-based discrimination in a high-stakes popularity contest. Data is taken from audience voting in six countries on the reality television show Big Brother, a setting where statistical discrimination can play no role. The audience votes as to which contestants remain on the show, the winner of which earns a large cash prize; I test the grounds on which voters discriminate, whether by gender, race, or age. Results show a striking taste for discrimination against women: being female makes an eligible contestant significantly more likely to lose an audience vote in five of the seven versions of Big Brother analysed. There is also evidence of a taste for discrimination against non-white contestants amongst audiences in Germany, Italy and the UK. However, little support is found for taste-based age discrimination. I present evidence that the levels of discrimination identified are robust to differences in the types of contestant appearing on Big Brother.