Sunday, November 6, 2016

Pols and Polls - the good, the bad, and the ugly

Summary for Nov. 6: Clinton still is forecast to win with chances at 98.4% (Huff), 84% (Upshot), and 65.5% (538). The same sources estimate her share of the electoral vote as 335, 322, and 292. Chances of flipping the Senate stand at 66% (Huff) and 53% (Upshot). There’s a late surge of Latino voting, a good thing for Clinton.

You can dig into the details after the break, including some reasons why Huff and 538 models make different predictions. We’ll know about which approach did better in two days.

*NY Times reports Latino surge*: Hillary Clinton Appears to Gain Late Momentum on Surge of Latino Voters

*Huffington Post explains Huff–538 differences*: Nate Silver Is Unskewing Polls — All Of Them — In Trump’s Direction

Nate Silver talks about uncertainty (via email): Here’s some of what he said.

  1. Who’s ahead in the polls right now?

Hillary Clinton is ahead in most national polls, as you can find every number from a 1-percentage-point Clinton lead to a 6-point lead in recent national surveys. There are also a couple of polls that still show a tied race or — in one case — Trump ahead. Overall, the range of national polls has narrowed a bit, although it remains wider than what we saw over the past few campaigns, with Clinton ahead by about 3 points on average.

One could argue about whether Clinton’s still ahead in the Electoral College, however. New Hampshire, Florida, North Carolina, Nevada and the 2nd Congressional District of Maine are all extremely competitive in recent polls. (Our forecast still has Clinton ahead in New Hampshire — by about 2 points — but there’s plenty of polling to support the notion of a small Trump lead there instead.) That means Clinton has 268 electoral votes in states where she’s clearly ahead in the polls — two short of the 270 she needs.

Thus, while Clinton’s a 76 percent favorite to win the popular vote according to our polls-only forecast, her odds are more tenuous — 64 percent — to win the Electoral College. (Her chances in the polls-plus forecast are identical.) It would not necessarily require a major polling error for Trump to be elected, though he would have to do so with an extremely narrow majority in the Electoral College.

  1. What’s the degree of uncertainty?

In some ways, our fundamental hypothesis about this campaign is that uncertainty is high, with both a narrow Trump win and a more robust Clinton win — in the mid-to-high single digits — remaining entirely plausible outcomes. The polls-plus model, which gives Trump a 36 percent chance, is basically the same one that gave Mitt Romney just a 9 percent chance on the eve of the 2012 election, so it isn’t inherently so cautious. But the still-high number of voters not committed to either Trump or Clinton — about 13 percent of the electorate says it’s undecided or will vote for a third-party candidate, as compared with just 3 percent in the final 2012 polling average — contributes substantially to uncertainty.

So does the unusually broad swing-state map, with the outcome in at least a dozen states still in some doubt. And it’s important to remember that the outcomes in each state are correlated with one another, so that if Clinton underperforms her polls in Wisconsin (for instance), she’ll probably also do so in Minnesota. Forecasts that don’t account for these correlations are liable to be overconfident about the outcome. It isn’t hard to find examples of candidates who systematically beat their polls in almost every competitive state, as President Obama did in 2012 and as Republican candidates for governor and senator did in 2014.

And that’s before accounting for some of the factors that the model doesn’t consider: the disagreement in the polls, the unusual nature of Trump’s candidacy and the demographic changes it is producing, Clinton’s superior turnout operation, the possibility of “shy Trump” voters, the fact that the news cycle is still somewhat fluid headed into the final weekend, the declining response rates to polls, and the substantial number of high-profile polling misses around the world over the past few years. We think this is a good year for a forecast that calls for more caution and prudence.

No comments:

Post a Comment