I wouldn’t give much credence to the polls that measure Trump’s support. We’ve seen time and again that these polls underestimate Trump’s following by a noticeable margin.
What happens is that people realize that voting for Trump is not a socially approved choice, so they are not too eager to disclose their preference for him. I’d disregard these polls completely. They only measure how many people feel like disclosing they’ll vote for Trump, and that information is useless.
I agree that the polls are not reliable. However, I don’t see any reticence on the part of Trump supporters to say whom they support. The big issue with the polls is the adjustment pollsters make for likely voters and their inability to get representative samples of respondents.
Polls tend to get respondents who are most likely to be available online with time on their hands and an interest in politics — basically, connected seniors and college students. Others will respond, but at rates well below the proportion of the population they represent. Some won’t respond at all, such as the very wealthy, that physically and socially active, immigrants (most polls are only in English) and the alienated. So that’s the first mess: who does the poll miss and what bias does that cause in the results?
The PC issue is whether people will actually vote. That’s where people overstate their intentions, because there is a social norm that people should vote. In “normal” elections, you ask about whether people voted in past years and use that to develop an index about how likely they are to vote this year. (People lie less about what they did than about what they plan to do. Not perfect, of course.)
However, there’s nothing normal about this year. In a normal year, the heaviest turnout in the US is among seniors, and 20-somethings and minorities don’t show up. That could happen again in November, but it seems unlikely.
The other issue with early polls is that a lot of people haven’t decided what they want to do. Some won’t make that decision until the day of the election.
That’s why I didn’t pay much attention to the polls early in the Brexit campaign, but when I say the 52-48 split last weekend, I was worried. That’s within the typical margin of error for even a good poll. What the media should have been reporting was that the election was “too close to call.” A 2 point lead when the margin of error is 3 points means nothing.
LikeLike
Absolutely. That’s why even the first exit poll predicted Leave.
LikeLike
yep, you may already know this clarissa but there is a formal name for this and a pretty detailed body of research. The bradley effect. https://en.wikipedia.org/wiki/Bradley_effect This dealt with explicity of race, but amplifies bigger idea of acceptable expression of opinions.
LikeLike
No, I didn’t know this. Thank you for the link. I’m pretty sure this is what is happening with Trump’s support polls and why he keeps performing better than the polls suggest he should.
LikeLike
Meh, I don’t know about this. This so-called Bradley effect was trotted out millions of times in 2008 and 2012. ‘Oh, people really support McCain but they don’t want to appear racist so they choose Obama in the polls’. But nothing came out of it, of course.
Nobody knows if this purported bradley effect actually exists, there’s no proof of it working. It’s something ‘pundits’ use for narrative. Gotta fill the airtime!
LikeLike
Let’s hope so. I’m only afraid that all these recent “Hillary is ahead of Trump by double digits” polls will lull everybody into a summertime torpor and momentum will be lost. This is too important an election not to be observed very closely.
LikeLike
“Trump’s support polls and why he keeps performing better than the polls suggest he should.”
Primary polling this year suggested he would win handily. And he did. He led the field from the first poll to the last poll, an amazing achievement. There were no wings! It’s the pundit class who didn’t believe the polls.
General election polling suggests he’s losing to Clinton. What makes you think he’s ‘performing better’? By what metric?
Trump’s supporters are not the kind of people who are in any way shy about their support. This is not a Nixon ‘silent majority’ situation.
LikeLike
”..no swings..’
LikeLike
“General election polling suggests he’s losing to Clinton. What makes you think he’s ‘performing better’? By what metric?”
“Trump’s supporters are not the kind of people who are in any way shy about their support.”
LikeLike
All the “Bradley effect” says is that some people are hesitant to give answers that they know will be unpopular, or that will subject them to unwanted criticism. That’s an obvious part of human nature, and it’s been around forever.
There’s nothing new about that trait in people, and no doubt that it exists.
LikeLike
Well, there’s something to be said about the false-consensus bias, which is typically exhibited by less self-aware people who think their stupid opinions are quite normal and are shared by rest of the world.
“Women are emotional creatures. Everything knows that!”
That’s Bradley Effect in a nutshell.
“I lied on my poll question to appear non-racist, because I’m an insecure person. I’m sure there’s millions of people who think exactly like me. Ergo Obama will lose.”
LikeLike
This is precisely what I fear.
LikeLike
A conservative literary critic and college professor
Clarissa's Blog 
Leave a comment