The Trial-Heat and Seats-in-Trouble Forecasts of the 2016 Presidential and Congressional Elections
James E. Campbell, University at Buffalo, SUNY
In general, election forecasting has become easier in recent years. The increased polarization of the public and the hyper-competitively realigned parties has narrowed the range of plausible election outcomes. As mentioned in the introduction, over the last 30 years, neither party has received more than 54.7% of the two-party national popular vote (Bill Clinton in 1996) or less than 45.3% (Bob Dole in the same election). Three of the seven presidential elections since 1984 have been won with less than 52% of the two-party vote. A tight range of the vote is exactly what one would expect from a pair of highly competitive parties with adherents to each tightly dug into their ideological trenches. If we stick to this range and if its foundation is based on polarized partisanship, then forecasts cannot go as wrong as they could in the good old days when swing voters were (relatively) plentiful and landslides with over 60% of the vote were a real possibility. Presidential election forecasting, in general, ought to be easier (Campbell 2014a).
But forecasting the 2016 election is not forecasting in general. In many respects, 2016 does not look anything like a typical presidential election. The candidates of a typical presidential elections are not bombastic celebrity real estate tycoons opposing a former First Lady, Senator, Secretary of State, and defeated presidential nomination hopeful. Then there are the scandals and the significant internal party divisions, on both sides. Both candidates have sky-high unfavorables, perhaps indicating more late-deciding voters than usual. This election seems more out of a 1960s Allen Drury novel (Gen-Xers and Millennials may substitute House of Cards) than political reality. If ever there were a need for the forecasts’ systematic and objective bearings on an election, 2016 is it.