Throw Down Week: An Elections Prediction Smackdown
Part 1 of a Three-Part Wonks’ Series
With less than a week to go before Election Day, the consensus among election prognosticators and pundits mainly revolve around a Hillary Clinton victory in both the popular vote and the Electoral College, albeit under varying degrees of uncertainty. With so many commentators predicting anywhere from a landslide win in the Electoral College to a narrow margin of victory for Clinton, the question becomes less about “Who Will Win?” to more of “By exactly how much?”
There are plenty of quality presidential prediction models out there but not all of them have the Wonks stamp of approval ®. At Wonks This Way, we made a last-minute decision to hop onto the predictions bandwagon with a head-to-head matchup between the two Wonks (e.g., myself and Kevin) in what we’ll deem “Throw Down Week.” In a three-part series, Kevin and I will compete by presenting our own election predictions along a set of three election stats:
- Popular vote percentage for each candidate
- State-by-state winners, and
- Final Electoral College tally
For Part 1, I laid out my predictions of the eventual outcome in the section below while my colleague Kevin will present his rebuttal tomorrow for Part 2. On Thursday next week, we’ll wrap up with Part 3, which will consist of a post-election write up centering on who came out on top along with commentary on the actual election results. Throw Down Week will also serve as our submission for DrivenData’s “America’s Next Top (Statistical) Model” competition.
And now, Part 1.
The Pastry Pundit vs. The Machine: A Process
What separates my gut-induced electoral forecast from Kevin’s forthcoming chicanery comes down to methodology; I’m going off of the pseudo science of punditry and what other people have been paid to write about for the news cycle while he’s using some fancy computer science data to create a made up algorithm. In other words, I’m mentally aggregating some of the analyses from FiveThirtyEight and whatever other sources I use to get my news for the substance of this piece.
With that said, my final predictions:
Popular Vote
Clinton: 50%
Trump: 46%
Others: 4%
Electoral College Map
In a tighter race than 2008 and 2012, Hillary Clinton wins the election with half the popular vote and 323 electoral votes from 25 of 50 states plus the District of Columbia. In losing Iowa and Ohio — two states that Obama carried in both of his elections — Clinton will surpass the desired 270 threshold by winning New Hampshire and Nevada, with both a narrow victory in Florida and North Carolina’s return to the Democratic column after the Tar Heel state went to Mitt Romney in 2012, inflating her total electoral vote count above 300.
Why will the five aforementioned swing states tilt the way that I assume? A couple of observations and rationales:
Iowa and Ohio
Hillary Clinton’s currently under-polling Barack Obama’s performance in these states and their presence in the heartland and Rust Belt correlate with higher concentrations of white voters without college degrees who are among the constituencies most likely to favor the Republican nominee. This fact should tip these states in Trump’s favor.
New Hampshire
Conversely, Trump should lose the Granite State despite its libertarian reputation for independent mindedness and its use of “Live Free or Die” as its official state motto. New Hampshire has a higher proportion of white college-educated voters than the previously mentioned swing states, which I believe will push Clinton over the top. As you may notice from the accompanying graphic, their use of the French language in “welcome” signs to visitors suggests that any cosmopolitan affectations among the New Hampshire electorate should outweigh the sentiments of its blue collar voters.
Florida and Nevada
Both these states have significant and growing Hispanic populations who are expected to turn out in larger numbers than in previous elections, largely among Mexican-Americans disgusted by Donald Trump’s comments on immigration and his proposal to build a wall along the U.S. border to Mexico, as well as the growing populations of non-Cuban Hispanics with fewer party loyalties to the GOP in Florida.
North Carolina
It’s a larger state with greater diversity; also see New Hampshire above.
While the final week of election coverage revolve around the FBI’s reopening of the Clinton email story and Donald Trump’s tax woes, I don’t think either story will have much of an impact in determining the results with early voting well under way for about a month now and most voters seemed to have already made up their minds by now about the candidates.
Your move, Kevin.
