They probably do to each other. I wouldn’t mention that I’m a Democrat in some of the barbecue joints I have patronized.
I’m having a hard time imagining a universe where Trump voters are capable of human emotions which are not greed, fear, and rage. Shame? Please.
A recent 538 podcast said there didn’t seem to be any evidence for it. (The last model talk podcast).
I live in a supremely liberal neighborhood in ultra-blue Chicago, and I work in academia. You can bet your liberal ass that I don’t advertise my strong support of President Trump. But I vote.
That is nice… not really. The issue here is if you also showed that shyness to pollsters that you don’t know on the telephone or the internet by telling them that you are not voting for Trump.
Well it’s a moot point as I’ve never been called by a pollster. Not sure what I would do if I was. Probably just hang up on them without saying a word.
In that case your reply would not had counted, the point remains, it is not likely that there are many shy Trump voters that are misleading pollsters.
You make a good point. There is a difference between those who lie, and those who merely refrain from replying. I’m not sure which count as “shy voters” for the purpose of this thread, but clearly both exist.
I’m sure some number of people lie because they are in the minority (or for other reasons), even to pollsters. The main issue though, is whether there is a polling error because one side or the other is lying to pollsters more.
As posted before, FOX news pollsters would be less likely to encounter shy conservative voters, and their polls are not good for Trump in several battleground states, they also match a lot with what other pollsters are reporting.
Yep. There doesn’t seem to be any good reason to think it’s happening.
There probably exist a few shy Trump voters, whatever that means. But there also probably exists a few shy Biden voters. There’s no evidence either group is particularly large, or larger than their counterparts on the other side.
I wonder if there are more apathetic trump voters who might not show up, but wouldn’t admit it to a pollster, and might even go out and vote if they are polled because they’d be psychologically committing themselves by telling the pollster they’ll be voting. There can’t be more newly discouraged biden voters than discouraged trump voters.
I thought that was funny. I think it was Nate who said, “Have you SEEN Trump voters? They don’t seem shy to me!”
Interesting idea! What’s that scientific fact, that simply observing (documenting) something changes it — so you can never do research without contaminating the data set?
In sociology I think it’s the Hawthorne effect. In physics it’s the Heisenberg Uncertainty Principle.
I’ve heard that in the 2016 election the problem was that pollsters did not anticipate the large differences in the voting behavior of college educated and non-college educated whites, so that pollsters did not correct their weighting. College educated people; for whatever reason are more likely to respond to polls, so need to be underweighted in polls to account for their actual percentage in the voting population.
Slate had an interesting article about that. The pollsters say they’ve fixed the problem and were spot on in 2018,
Thanks!
Exactly right. Weighting is always applied for demographics and region (rural/urban, black/white) but wasn’t typically used for education level (because there were not huge differences within other demographic groups just based on education level prior to 2016).
I think most pollsters have learned their lesson on this. But of course aggressive weighting can increase the error bars since you often are amplifying the responses of a relatively small sub-set of folks that answered the poll. To circle back to the OP, if you only got a handful of low-education-level white responses, and some of them were “shy Trump voters”, the weighting based on education level within the white sub-group could amplify the effect of those “shy” voters.
I still think for now we are well outside of the magnitude of error possible in states like MI and WI, but moderate tightening in the polls would be enough to make only a “normal polling error” (as Nate Silver famously put it in 2016) enough for Trump to win again. Right now we are in “abnormal polling error” territory.
Ok, so Trump voters believe that unfavorable polls are a tool of the Democrats to influence voters. Yet they won’t answer pollsters truthfully to help their own side “influence” voters.
Like so many other Republican things, this is just crazy shit to me.
That’s because he’s only looking at the non-shy ones, of course. Sampling error.