Why Sanders' numbers are likely to keep going up in New York (and elsewhere)

[paulthompson]

As you may have heard by now, a poll released today has Clinton beating Sanders by 12 points in New York, 54% to 42%.

http://www.realclearpolitics.com/epolls/2016/president/ny/new_york_democratic_presidential_primary-4221.html

This is significantly better for Sanders than the three previous New York polls deemed reliable by Real Clear Politics, which had Clinton leads of 42 points, 21 points, and 21 points.

If Sanders has narrowed the lead this much with three weeks left to go, he's highly likely to narrow it more. When people have models of how many delegates Sander needs in each remaining state to win, they usually antiicipate that a slight loss in New York would be his best case scenario. But he might actually win there! At the very least, Clinton isn't going to have a decisive margin of victory that could effectively end the race.

Why? The Sanders late surge trend. I've been documenting this in state after state outside the South. (Unfortunately, the trend didn't happen in the South, but the South has finished voting, except for Kentucky.) I've posted these numbers a couple of times already, but I'm updating them and posting them again here, because I think they are so important to understanding poll results.

Here's a look at all the non-South results, in order of election date. (Note that most of the data comes from Real Clear Politics, including their final poll averages.)

Iowa - Sanders virtually tied after being down by about 20 points a month earlier.

New Hampshire - he beat the final poll average by about ten points.

Nevada - he lost by five points, but he did 18 points better than the one poll from a month and a half earlier.

Colorado - he beat the one prior poll by over 40 points!

Minnesota - he beat the latest poll by 50 points!

Massachusetts - he lost, but he beat the final poll average by five points. He also lost the early voting, but got a majority of the election day voting.

Oklahoma - he beat the final poll average by 12 points.

Vermont - his support in his home state was already basically maxed out - he won the state with 86% of the vote. But even so, that was seven points better than the one poll from a week earlier.

Kansas - he beat the one poll from a week earlier by 45 points!

Nebraska - apparently, no polls were done at all.

Maine - he bettered the one poll by 15 points.

Michigan - he beat the final poll average by 20 points!

Illinois - he lost by two points, which was in line with the last three polls. But the two polls from a few days prior to those had him down by 40 points!

Missouri - two polls had Clinton leading by five points and seven points respectively. One late poll had Sanders leading by one point, and the state ended up a virtual tie, with Clinton winning by less than a thousand votes.

Ohio - this is a bit of an anomaly. Clinton won by 14 points, which is six points better than the average of the last polls. But still, Sanders did ten points better than the average of the three polls from the week prior to that.

Idaho - he beat the one prior poll by 54 points!

Utah - he beat the last poll by 51 points! Plus, two poll prior to that one were off by 15 more points.

Arizona - he lost by 16 points, but that was still 14 points better than the final poll average. (Note that his numbers improved slightly a week after the election due to counting provisonal ballots.) He also lost early voting by a large margin, but won the election day voting.

Washington - no polls.

Hawaii - no polls.

Alaska - the lone poll from January showed Clinton winning by three points. In the actual election, Sanders had a 63 point victory margin!

Further evidence of this late surge trend can be seen by how he consistently loses big in early voting but does much better with election day voting. This has even happened in the South, where he has done so poorly overall. For instance, in Florida, Clinton won by 36 points by those who made up their minds one week or more prior to election day, whereas Clinton only won by 13 points among those who made up their minds in the final week. In North Carolina, Clinton was ahead in the final poll average by 24 points, but only won by 14 points. Why? Because Sanders lost the early voting there, but actually beat her with election day voting by four points.

---

The bottom line is, Clinton has a favorability rating of -14 nationwide, and Sanders has a favorability rating of +11 nationwide (according to the latest poll averages at Huffington Post). That means the more voters learn about Sanders, the more they realize they don't have to vote for someone they don't really like, so Sanders' numbers tend to go up.

Thus, you can be nearly certain that while Sanders might not win New York outright, his numbers there will keep trending better as we get closer to election day.

I've been keeping a close eye on the numbers for all the states, and I've only seen one time when Clinton actually outperformed the final polls, and that was in Ohio, where she beat the final poll average by five points. (And that was a strange situation, because many Democrats crossed over to vote for Kasich against Trump.) In every single other case, Sanders has equalled or bettered the final poll averages.

Oh, and New York and most of the states that follow a week later are closed primaries, which definitely favors Clinton (since Sanders is overwhelmingly popular with independent voters, who can't vote in those primaries). However, New York does not allow early voting (though it allows absentee balloting). Furthermore, of the states that vote a week later, Connecticut, Delaware, Rhode Island, and Pennsylvania also don't allow early voting, and only Maryland does. Most states that have voting so far have allowed early voting, so that's a factor that helps Sanders.