- Techniques from social science-- like social-circle polling and Bayesian Truth Serum-- could render better election forecasts than traditional polls.
- A cross-institutional group of social scientists applied these successfully during the 2016 US presidential election (results published in Nature Human Behavior). Now, they are using them again to forecast the November midterms for the House of Representatives.
- Mirta Galesic, Cowan Chair of Human Social Dynamics at the Santa Fe Institute, is available to discuss the group's predictions based on the data so far.
Quotes
“None of the novel methods bring good news for Democrats.” -Mirta Galesic, Santa Fe Institute
“Overall, our predictions based on social circles and the Bayesian Truth Serum suggest more of a 'blue splash' than a 'blue wave.'” -Mirta Galesic, Santa Fe Institute
Further reading
Psychology Today "And the winner is..." (forthcoming)
Nature Human Behavior "Asking about social circles improves election predictions" (February 26, 2018)
RSS