Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

One pollster’s explanation for why the polls got it wrong

vox.com

Post reply on HN
Print view View original

121–130 of 565 posts

Re: One pollster’s explanation for why the polls got it wrong

#121

Post by ttamslam »

For no particular reason, I feel very strongly that you could achieve much more accurate results by asking people to estimate what _they_ think the final margin will be then average those results.

I'm inclined to think that community members know their community well enough that given a large enough sample, this average will trend towards the true result.

https://en.wikipedia.org/wiki/Wisdom_of_the_crowd

edit: A quick google search found: https://www.researchgate.net/publication/289826354_Predictin...

Re: One pollster’s explanation for why the polls got it wrong

#122

Post by fastball »

rayiner wrote: post #116

> I reached out to the smartest polling guy I know: David Shor, an independent data analyst who’s a veteran of the Obama presidential campaigns who formerly operated a massive web-based survey at Civis Analytics before leaving earlier this year Before being fired for pointing out that violent protests alienate people: https://nymag.com/intelligencer/2020/06/case-for-liberalism-... Schor’s firing is a tremendous examp…

"How did Trump get so many votes?"

The same people that attacked Schor have undoubtedly also asked themselves this question, without an ounce of self-awareness.

Except instead of recognizing their own culpability, the conclusion most such people are coming to seems to be "well, I guess nearly half of all Americans are bad people".

Re: One pollster’s explanation for why the polls got it wrong

#123

Post by dillondoyle »

evo_9 wrote: post #104

Who replies honestly to political question to random strangers? I know I don't. What's the benefit of letting anyone know where you stand politically at this point? None. It's much more likely - no matter which side you're for - that the person asking could have an adverse reaction. Things are just way too polarized now, not sure how that can change, but that's how it is. I mean outside friends/family I simply take a…

Research has confirmed that people don't lie to pollsters in any meaningful way when asked their head to head preference.

It's mentioned in the last graph of the article. The crux of this article is rather than 'shy trump voters' there is simply a large amount of these Trump voters who will refuse to take a survey flat out.

Re: One pollster’s explanation for why the polls got it wrong

#124

Post by mgraczyk »

ttamslam wrote: post #121

For no particular reason, I feel very strongly that you could achieve much more accurate results by asking people to estimate what _they_ think the final margin will be then average those results. I'm inclined to think that community members know their community well enough that given a large enough sample, this average will trend towards the true result. https://en.wikipedia.org/wiki/Wisdom_of_the_crowd edit: A quic…

You may find this specific technique interesting, in which you try to find the "expert minority" in the crowd, under the assumption that people who can better predict the crowd's belief can also better predict the true outcome:

https://en.wikipedia.org/wiki/Surprisingly_popular

Re: One pollster’s explanation for why the polls got it wrong

#125

Post by bra-ket »

mring33621 wrote: post #108
bra-ket wrote: post #82

Earlier quoted context omitted.

it's a bit hard to trust mainstream media that have been lying to you and called you 'racist' for 4 years.

Voting for racists makes you a racist.

define 'racist' and what makes you think Trump is a 'racist'

Re: One pollster’s explanation for why the polls got it wrong

#126

Post by ericb »

majormajor wrote: post #92
daxfohl wrote: post #16

Earlier quoted context omitted.

Before 2016, polls were much more accurate. And paper ballots were the norm. So there's that.

So you don't trust the results of the 2016 election? You think Trump stole it? (As far as I can tell, actually, 2012 pre-election polling was off by a few percentage points too, and showed Romney leading for most of October...)

> You think Trump stole it?

Not the parent, but I think there's a difference between looking at a few datapoints that make you curious and jumping to "Trump stole it." As for 2012, who's to say they back the same party each time, or even that it is the same attacker? Certainly most on Hacker News know software can have vulnerabilities.

In other words, there should be many scientific steps between "that's odd" to "hypothesis proven" otherwise it's just a conspiracy theory.

Re: One pollster’s explanation for why the polls got it wrong

#127

Post by 0xffff2 »

stanford_labrat wrote: post #91
EEMac wrote: post #34

Earlier quoted context omitted.

> People tell the truth, when you ask them who they’re voting for. They really do, on average. That might have been true in the past. From 2016-2020, Trump supporters have been loudly and repeatedly called racistNaziMisogynistHomophobe, or the polite version, "deplorable". That is not an environment that encourages honesty.

As a center right moderate I definitely feel this. Especially at work/with acquaintances where I feel compelled to just silently nod and agree with everything being said or else risk my career and social standing.

As a center right moderate, did you vote for Trump? I might use similar terms to describe my political beliefs if I was into labels, and I feel none of this. I've always been frank about my political leanings when it comes up, and no one I know has had any trouble talking about the issues while leaving the Candidates out of it. That said, given where the Republican party is right now, I can't imagine voting for any R on the ballot (and contra-wise Biden and Jerry Brown are the only two D votes I've ever cast, both entirely in opposition to the frankly scary Republican opponent, not in support of the candidate).

Re: One pollster’s explanation for why the polls got it wrong

#128

Post by dillondoyle »

mring33621 wrote: post #97
0x53 wrote: post #3

One of the biggest issues I think is that the polling is still done by just randomly calling people. I think I agree that there is a large skew introduced by the simple fact of only reaching people willing to answer a call from an unknown number.

Agreed. I would also suspect that there are some other factors: - people hoping for change are more willing to answer polls - people who are ashamed of their opinions are more unwilling to answer polls The second one, perhaps, gives too much credit for self-awareness, though.

It's less about 'ashamed of their opinions' or the 'shy trump voter.' The article talks about this. The crux of the problem - at least argued by this authored and from what I see working politics backed up generally - is that those Republican / Trump voters do not trust others and are less likely to take the poll in the first place. Those that do take survey generally do not lie about their support.

The article gives evidence. Here is the section:

"I use the General Social Survey’s question, which is, “Generally speaking, would you say that most people can be trusted or that you can’t be too careful in dealing with people?” ....

It turns out, in the GSS, that 70 percent of people say that people can’t be trusted. And if you do phone surveys, and you weight, you will get that 50 percent of people say that people can be trusted. It’s a pretty massive gap.

[Sociologist] Robert Putnam actually did some research on this, but people who don’t trust people and don’t trust institutions are way less likely to answer phone surveys. Unsurprising! This has always been true. It just used to not matter."

Re: One pollster’s explanation for why the polls got it wrong

#129

Post by ttamslam »

mgraczyk wrote: post #124
ttamslam wrote: post #121

For no particular reason, I feel very strongly that you could achieve much more accurate results by asking people to estimate what _they_ think the final margin will be then average those results. I'm inclined to think that community members know their community well enough that given a large enough sample, this average will trend towards the true result. https://en.wikipedia.org/wiki/Wisdom_of_the_crowd edit: A quic…

You may find this specific technique interesting, in which you try to find the "expert minority" in the crowd, under the assumption that people who can better predict the crowd's belief can also better predict the true outcome: https://en.wikipedia.org/wiki/Surprisingly_popular

This is very interesting, thank you!

Re: One pollster’s explanation for why the polls got it wrong

#130

Post by _AzMoo »

macintux wrote: post #90
_AzMoo wrote: post #67

Earlier quoted context omitted.

Exit polls were historically used as a method of detecting election fraud, because they've always been very accurate. It's only the last few elections where we've seen such a large disparity in polling and actual results.

Is exit polling no longer working? All of the discussion I’ve seen seems to focus on polling in advance of the election.

Yeah, since about 2000 it's been shown to be really unreliable in the US. Heavily weighted towards democrats for much of the same reasons that makes normal polling unreliable now.
Post reply on HN