IA-Emerson: Trump +5

[pbrower2a]

If Emerson College polls indicate how America will vote in November, maybe I do not want to take my blood pressure pills.

They are out of line with national tracking polls. It;s not that some states do better for the Republican in a few states and worse in others; it is that the polls are consistently off. There could be problems of sampling.

I expect more polls of Iowa within the next couple of weeks.

[pbrower2a]

Let's take a look at the overall activity of the pollster. It hasn't done any polling in a long time and suddenly gives us a large number of polls that show Donald Trump doing much better, and very suddenly, for no apparent reason for him to be doing better.  Every poll looks like veritable wish-fulfillment for a Trump fan.

There is no indication that Donald Trump is doing anything better than he has since the primaries. Really, one could hardly dream up more sympathetic polls after some indication that Hillary Clinton is on a pace to win at least as strongly as Barack Obama in 2008. After one week? Where is the smoking gun?


Yes, PPP did deliver a large number of polls on behalf of an organization whose objective is the raising of the minimum wage, and those suggest no real change. But my rules indicate that because they are for a special-interest group with a partisan bias and they do not fit a need to fill in a poll (in accordance with the principle that beggars can't be choosers, so I would have to accept just about anything involving some states that get polled rarely).

Either there is an inexplicable surge by Donald Trump or there is something screwy with these polls. They have some odd weighting and are land-line only; they do not offer an option for a response in Spanish (important for reaching people who might consider any call in English by a complete stranger as junk).

There likely will be little polling over the weekend, but a poll from New Hampshire shows Hillary Clinton winning the state by a landslide margin while landslide margins by which she was allegedly winning in states like Michigan, Pennsylvania, and Virginia have all but vanished.
   

[pbrower2a]

So you CHOOSE to believe the polls you like, right?

There are A nationall polls.
There are A WI polls.
There are plenty of B polls.

Almost all showing the race tightening (some slightly, some considerably).

But muh, NH poll with 400 sample size knows best!

The poll of New Hampshire does show a little slipping by Hillary Clinton. So did the Marquette University poll. New Hampshire is homogeneous enough that a poll of 400 people can get a good picture of statewide politics. One would need more for Ohio, and even more (like about 4000) for Texas or California

A recently-inactive pollster flooding the media with several new polls at once? That is itself suspicious.

I rejected polls because they come from special-interest groups, like a bunch of polls by PPP for a group seeking to raise the federal minimum wage. If I don't do that then I must accept polls by  trade associations and other right-leaning groups. I have shown polls that I have not used, and shown why I refuse to use them.

Has Hillary Clinton slipped some from last week? Probably -- because Donald Trump hadn't said anything egregious for a week or so. Hillary Clinton fares well in the polls to the extent that Donald Trump makes a fool of himself.