Why has Trump Gained in Polling Averages ?
Lifted from Robert’s Stochastic Thoughts
By Robert Waldmann
Why has Trump Gained in Polling Averages ?
Yesterday, the talk of the day was Democrats panicking (as always) because support for Trump is creeping up in polling averages (as are estimated probabilities of President Trump). Today’s talk is totally deplorable and I will ignore it.
I think/ believe the shift (from mid August to early September 2016) is mostly due to which polls have dropped (there are clear house effects) and sampling error.
The reason is that it doesn’t appear at all in the USA LA-Times panel poll. http://www.latimes.com/politics/
They poll the same people again and again. They weighted so that self reported 2012 voting corresponds to actual 2012 votes. This means that they underweight people who claim they voted for Obama (systematically more claim to have voted for the winner than actually voted for the winner). This makes the level of the poll uninteresting.
But changes show how actual people have changed their minds, and are interesting
Look at Trump gaining uh ooops What ! In this panel poll of the same set of people, Clinton has gained from mid August to early September. The sample isn’t huge, but the change isn’t due to any change in the sample. People trying to explain what has happened have to deal with these data.
The poll that dropped on Monday showing Trump with a 2 point lead where it had previously been 4 points HRC was not a major poll, had some methodological issues anyway, but most importantly had around half of its total change be the result of a change in top line reporting from ‘registered’ to ‘likely’. To their credit they also published the new ‘registered’, which showed HRC with a lead, albeit smaller.
The switch from “registered” to “likely” actually seems to be fairly standard around Labor Day, so there was nothing particularly nefarious. But it does mean the pollster has to interpose some weighting based on past elections. Normally this weighting shows that older voters and white voters HAVE BEEN more likely to actually vote than younger voters and POC, and since those demographics are historically lean more R then even if there is no actual movement in opinion you will get some movement in reporting.
So I certainly agree with Robert about “which polls have dropped” and “sampling error”, I just suggest that you take a hard look at “likely” vs “registered” in any before and after.
Also, and in a point I made in an e-mail to some Bears a couple weeks ago. There is a solid argument for devoting August to fundraising and backing off campaigning. Because while you might get dinged a little in what polling there is, historically most people are not really focusing before Labor Day. Which suggests piling up cash in August with plans to roll in out later in September and October with a focus on turnout and as necessary negative and positive advertising that will move numbers when it actually counts. If this was the decision made by the HRC people and if it was the right decision you would expect polls to start moving the opposite direction as this spending bit. And maybe it is just wishful thinking but I see early signs of this in just the short week since Labor Day.
Time will tell. I think the money advantage properly targetted to the right swing States will have an outsize impact on the Electoral Vote count. And while gross turnout and margin are important, not least because of follow on effects on down ballot races, I don’t blame the Clinton folk for keeping their eyes on the prize here.
Nate Silver points out that the state polls are much more predictive and important than national polls.
Beating Trump should have been a walk in the park, as they say. HRC is begining to look like an amateur at a game she is well versed in. I could not believe the “basket of deplorables” comment, and started checking property values in Toronto when I first heard it last evening on the news. It soulds like something that Trump would have said himself when trying to sound more Presidential. Except its dumb to the nth degree, especially considering all the things that could be said about Trump himself, his less than ethical business behavior, his idiotic comments, etc. We’re beginning to see why Obama stole the nomination right under her nose in her previous bid. She’s not running a sharp campaign. She wasn’t all that clever as Secty. of State and we know were she stood in the Senate regarding Bush’s lust for war. Exactly why do we think she’s so talented? I’m not saying Trump is better. I’m just wondering why the polls aren’t breaking the glass ceiling above her head as she runs against the biggest buffoon ever put up for the position.
As of 2012, African Americans actually have a slightly higher voting turnout than whites, 66% vs 64%. Hispanic Americans voted at 48%. This explains why states like Arizona and Texas are still trending Republican, although the demographics alone would argue for them trending Democratic. Perhaps the difference in voting rates, almost 20% points difference, between African Americans and Hispanics is due to religious affiliation.