First Sabato, now Nate Silver hedging his shaky bet
The headline on fivethirtyeight.com today is How Trump Could Win the White House While Losing the Popular Vote. He first makes clear that he is not predicting a Trump win and that he stands by his algorithm for predicting elections. Okay, got it. As I see it, he then proceeds to make that case that if his algorithm is somewhere around 50/50 in predicting a winner, he's really saying that Trump is the more likely winner.
If I'm right, we are now in a place where a 50/50 prediction is really more like a 55/45 or even a 60/40 bet for Trump, right? Wouldn't that suggest that Mr. Silver's algorithm has a flaw that perhaps he should adjust? Let's look deeper.
Mr. Silver explains why this electoral win, despite a popular vote loss, is more likely than normal. The reason makes sense. Clinton faces a big electoral problem because she holds a very large lead in some of the big states in which she leads, popular vote that will not manifest in the electoral vote. She is also too far behind in other smaller states where anti-Trump demographics are located for those NeverTrumpers and the like to make a difference.
Specifically, three of the groups where Trump is doing the worst, college educated whites, Latinos and Mormons, are concentrated in states that are not battlegrounds. College-educated whites exceed 35% of the population in only two battlegrounds, New Hampshire and Colorado, neither of which are states Trump needs. Latinos exceed 15% only in three battlegrounds, Florida, Nevada and Arizona. Silver makes the point that Arizona has too big of a GOP built-in advantage to overcome. He also notes that Trump doesn't need Nevada, although he is leading there, and all Trump really needs is to improve on Romney's performance among whites by 3% in Florida to the overcome Latino bump in Florida.
Last, Silver looks at Mormons. Trump is not doing well among Mormons, probably because prominent Mormons like Romney, Glenn Beck and Jeff Flake routinely attack Trump. Despite this fact, however, Mormons are not located in areas where they are likely to make an electoral difference. Most Mormons are concentrated in Utah, California and Idaho. The outcome in those three states is certain, Trump will win Utah and Idaho and Clinton will will California. The Mormon populations in Nevada and Arizona are too small to matter much.
So in summary, we can conclude from Mr. Silver's analysis a couple of things. First, the national polls don't matter that much because they include larger than normal leads in populous states that will not translate to additional electoral votes. California, New York and Illinois are examples. Those national polls also include states where Trump may have a smaller than the normal GOP lead because of such anti-Trump groups, but leads that are nearly certain to hold on election day anyway. Texas, Arizona, Utah an Idaho are good examples.
What to do about this dilemma? This question highlights the facts necessary to make a second conclusion. I would expect that Mr. Silver ought not show a 50/50 prediction if he thinks Trump's chances are better than 50/50, as he seems to suggest he would in this article. We can conclude from this that if he ends up throwing up his hands and going with the 50/50 prediction, or something close to it, it seems like a way for him to avoid predicting a Trump win while still getting credit for it in the end.
Nate Silver can point to this article and remind us that his 50/50 prediction was really a Trump-win prediction. For that matter, he could point to the article he wrote today to argue that his current 60/40 prediction was really more like a 50/50 bet, and he was just off a little on an assumption here or there. I'm starting to have may doubts. Remember, he already admitted that he screwed up on Donald Trump once. It looks like he's again unwilling to formally acknowledge the gains Trump is making.
This is what we are seeing with Sabato's Crystal Ball, a nod to the fact that Trump has better chances than the analysts are willing to formally state. It's hedging the bet or CYA, pick your description, but they both fit.
Follow my new account on Twitter for daily updates on the election!