Forecasting with faces forthcoming in International Journal of Public Opinion Research

Can schoolgirls who do not recognize potential candidates do a better job than polls of making an early prediction of who will win an election? A paper by Scott Armstrong, Kesten Green, Randy Jones, and Malcolm Wright, now forthcoming in the International Journal of Public Opinion Research, presents evidence from research to predict the outcome of the previous U.S. presidential election and previous studies that they can. The New Zealand schoolgirls and other research participants were shown only pictures of potential candidates faces and were asked to rate their competence. The candidates with the highest facial competence ratings in each of the two parties received the highest popular vote. For the Republicans, McCain had the highest facial competence rating at a time when Giuliani was well ahead in the polls. For the Democrats, Clinton had the highest rating and Obama was in second place; both had higher ratings than McCain. Clinton received slightly more popular votes but lost the nomination to Obama in the vote by party delegates. A working paper version of the paper is available here (http://mpra.ub.uni-muenchen.de/9150/1/MPRA_paper_9150.pdf).