18

Most of the polling for the 2016 US presidential election have been favourable to Hillary Clinton. One of the biggest questions I have right now is whether the polling is accurate. In particular, whether people are hiding the fact they intend to vote for Trump when polled for it. Have pollsters examined the possibility of a shy Tory factor / Bradley effect and seen whether it's affecting polling numbers for this election campaign? What have they concluded?

  • 2
    Imagine this poll: "1) Who would you vote for POTUS? 2) Who would you REALLY vote for POTUS? Don't lie." :-D. AFAIK, the only way to know how accurate a poll is is to contrast it with electoral results. Polling orgs develop models that correlate polls with results, and keep adjusting them as much as they can, but have difficulties accounting for situations that do not fit well into their model until there is enough historical data to improve the model. – SJuan76 Oct 30 '16 at 3:10
  • 2
    @Kris Online polls have two issues: they are new and their models are untested (nytimes.com/2015/11/28/upshot/…) and usually attract partisan people (who are more likely to click in the link and spend time with the poll), although some sites manage the later issue by not making specific political polls but adding some political questions to other kind of polls, to get to a broader demographic (bloomberg.com/politics/articles/2015-11-20/…) – SJuan76 Oct 30 '16 at 3:17
  • 1
    About your title: noone in the US will know the term "shy Tory factor", they won't even know what "Tory" means, and they won't understand it to mean "reluctance to self-identify by Trump supporters". The term "Bradley effect" is slightly known, but still obscure. – smci Nov 12 '16 at 13:53
  • 2
    I am from the US. I know what "shy Tory" means. I had to look up the "Bradley effect" to know what it meant when I first saw it on this site. I still forget who Bradley was and why it is similar to the "shy Tory" effect. And beyond that, not everyone here is from the US. Internationally, far more people are going to know what a Tory is than a Bradley. Heck, I suspect that there are more people in the US that know what Tory is than know about Bradley. – Brythan Nov 12 '16 at 17:19
2

538.com noted:

But James Lee of Susquehanna Polling & Research Inc. said his firm combined live-interview and automated-dialer calls, and Trump did better when voters were sharing their voting intention with a recorded voice rather than a live one.

This is a variant of the Shy Tory or Bradley effect. Of course, it could also be caused by other things. For example, live-interview may include cell phones and Clinton voters may be younger and more likely to use cell phones.

Note that immediately before that they said:

Several pollsters rejected the idea that Trump voters were too shy to tells pollsters whom they were supporting.

So we don't actually know if this is true or not. Professionals differ on their interpretation of the data.

It's also worth noting that the final popular vote result seems to have been within 1% of the polling. In early returns (when this question was posted), the results were farther away. That said, there were still some notable polling misses, e.g. in Wisconsin, where the final result differed from the polls by more than the margin of error in all polls.

0

All polls take this into account. I don't have any english source, I saw this in lessons of a french politics university. It has been applied for at least the last 25 years. There is a factor which is actualised at each election. It looks very surprising since it is never mentioned in the media, and it looks odd to be shy in an anonymous poll. But it is true.

If you want english sources, I suggest you look into politics courses. It is quite hard to find, if someone has something, please post it in comment.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.