Our paper, published in Pubic Opinion Quarterly, presents state-level estimates of both turnout and vote preferences in the 2016 presidential election using data from the ABC News/Washington Post tracking poll and multilevel regression with poststratification (MRP). The models, developed over the course of the campaign, successfully predicted the outcome in nearly all states using cumulative data from our ABC News/Washington Post polls.
The paper, co-authored by Chad P Kiewiet de Jonge, Gary Langer and Sofi Sinozich, reports how vote preferences changed over the course of the 18-day tracking period, compares subgroup-level estimates of turnout and vote preferences with the 2016 CPS Survey and National Election Pool exit poll, and summarizes the accuracy of the approach applied to the 2000, 2004, 2008, and 2012 elections. It includes discussion of how researchers can use MRP as an alternative approach to survey weighting as well as in forecasting elections.