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

201–210 of 565 posts

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

#201

Post by rsynnott »

jonplackett wrote: post #19

When and where was the last poll that predicted an election accurately? Polling is starting to feel more like a weather forecast - presented with absolute certainty, only right about half the time.

I mean, if you're looking for sub-1% accuracy, you're not going to get it, but polls are usually accurate to within a few percent. In the US, the last couple of elections were quite close, so fairly small misses had a big impact (at least in 2016; in 2020 the only state that was clearly a wrong call based on polls was Florida, and arguably Georgia).

Notably, polls were generally fairly accurate for the US 2018 midterms.

In general, polls work better in electoral systems where national polls are useful; the US presidential election in particular is challenging.

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

#202

Post by throwaway32849 »

throwaway32849 wrote: post #145
x87678r wrote: post #24

I think no one really knows what the different parties represent any more. Most people I know voted Republican were low-wage working people, where my highly paid educated friends voted Democrat. Isn't it supposed to be other way around? No wonder everyone is confused.

> Most people I know voted Republican were low-wage working people, where my highly paid educated friends voted Democrat Biden actively courted high-earning, educated voters with his free college and student loan forgiveness policies, which amount to a wealth transfer to people who already have higher lifetime average earnings and are at a competitive advantage in the employment marketplace. Not very progressive: htt…

I'm not sure why this was downvoted. The degree divide is growing as an electorate bifurcator, with the GOP fast becoming the party of the college-less:

> https://www.csmonitor.com/USA/Politics/2020/1027/As-college-...

And as I pointed out, it's not solely for the reasons college-educated progressives are eager to suggest ("ignorant, uneducated people vote republican"). Student loan forgiveness is a flatly a regressive policy, amounting to a handout to people who are already economically advantaged.

> https://www.cnbc.com/2020/11/09/student-loan-forgiveness-cou...

> Biden laid out a narrower plan, saying he would forgive $10,000 in student debt for all borrowers, and the rest of the debt for those who attended public colleges or historically Black colleges and universities and earn less than $125,000 a year.

Imagine being a non-college educated voter earning $60-70k a year, and reading that people earning upto $125k a year will get a handout.

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

#203

Post by x87678r »

y-c-o-m-b wrote: post #178
x87678r wrote: post #24

I think no one really knows what the different parties represent any more. Most people I know voted Republican were low-wage working people, where my highly paid educated friends voted Democrat. Isn't it supposed to be other way around? No wonder everyone is confused.

I think exit polls ( https://www.nytimes.com/interactive/2020/11/03/us/elections/... ) - which again could be wildly inaccurate especially due to their small size - showed people making over 6 figures were mostly voting for Trump.

yup looks like I'm wrong.

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

#204

Post by mjevans »

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...)

I think the Electoral College 'stole' it, from the popular vote.

I blame a combination of:

advertisements :: all political ads should probably just be banned period

misinformation on social media :: probably _some_ of which are foreign powers trying to divide the population of other nations

Lack of IRV :: I'd prefer any IRV system over what we have now, though https://en.wikipedia.org/wiki/Schulze_method seems like a winner focused variation of eliminate the 'biggest looser' first and yield the winner IRV.

The Electoral College :: for distorting the entire process and being non-necessary in the modern world. In the late 1700s when it was created the speed of communication was horseback letters and trans Atlantic sailing ships. It made sense to select a representative who would then travel in person to the core of politics, learn more, and possibly change their mind if the interests they represented were better served by another choice. The hyper-politicized, party focused, blindly following electors of today may as well not exist.

To solve voter representation and choice I believe the US needs:

0) Go all mail in ballots everywhere, like Washington State https://www.sos.wa.gov/elections/faq_vote_by_mail.aspx

1) IRV, preferably "beatpath method, beatpath winner, path voting, and path winner" https://en.wikipedia.org/wiki/Schulze_method

2) Remove the electoral college. Step A.1) https://en.wikipedia.org/wiki/National_Popular_Vote_Intersta... Step A.2) Require all votes are counted everywhere. B) Constitutional amendment removing the Electoral College; and while we're at it, mandating IRV of some sort (and hopefully enumerating academic criteria of the method rather than enumerating specific methods).

3) A ban on all political ads and limits on campaign finance and PACs.

4) More voter (general populace) education, an informed public is an implicit requirement for a functional democracy. Probably also better info-graphics presenting 'the books' (money in, money out) about who's paying for what results.

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

#205

Post by zepto »

SV_BubbleTime wrote: post #193
zepto wrote: post #2

This presumably also affects polls on things that measure support for individual issues related to trust in society - for example Gun control.

You can also look at previous votes on initiatives as semi-accurate polling. Michael Bloomberg / Everytown for Gun Safety / Moms Demand Action / The Trace / Mayors Against Illegal Guns which are all the same entity repeatedly pushed out the statistic that "90% of guns owners want to end private gun sales" of course this is totally misleading, but... When it came a vote in Washington state, after outspending the NRA o…

Previous votes generally aren’t semi-accurate.

For example ask people to vote on ‘universal background checks’ alone, you’ll probably get a majority voting in favor.

If you ask if a gun owner should give their guns to a trusted friend of family member for safekeeping if they feel depressed or suicidal, most people would also say that was a good idea.

Of course they wouldn’t realize they had just criminalized it.

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

#206

Post by drilldrive »

Frankly, this has an obvious answer: The people who answer polls are not representative of the population at large. Yet this was known decently enough since 2016. But more at the heart of the issue here is pollsters are providing a service not to inform but to persuade (both the public and donors); internal-polling is actually what matters. And instead of reformulating the service with more accurate questions for total populace (who will your neighbor vote for?), the pollsters doubled down on the standard methods and made (roughly doubly!) a larger polling error than 2016. I mean the Maine senate race was on average ~15 points off the final result[0]. Frankly this was the race for the pollsters to prove themselves that they can self-correct, as they have demonstrated to be unable to do this, it is clear they just should be abolished.

[0] https://www.realclearpolitics.com/epolls/2020/senate/me/main...

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

#207

Post by kube-system »

joshuamorton wrote: post #164
Damorian wrote: post #159

Earlier quoted context omitted.

"Pollsters say people don't lie to pollsters" is a bad argument. I ended up on some list and lied to every caller out of spite. These people are effectively telemarketers for corrupt propaganda companies interrupting my family at dinner. They don't deserve my honest opinion.

The important thing isn't people lying to pollsters, its people lying to pollsters in a way that's coordinated with a particular demographic or voting block. If 10% of people surveyed, at random, lie to the pollsters and choose the "opposite" candidate out of spite, you'll actually see a sort of reversion to the mean compared to the actual opinions.

Assuming that people are lying at random.

I would bet there is a significant correlation with voting habits -- particularly in this election cycle. Why wouldn't someone who holds the view that polls are corrupt and biased have their attitude towards answering them affected by that?

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

#208

Post by nl »

kjeetgill wrote: post #188
flowerlad wrote: post #183

Earlier quoted context omitted.

Not bad people, but gullible people. Anyone that believes QAnon conspiracy theories such as Hillary Clinton, Bill Gates, and Oprah Winfrey are raping and drinking the blood of kidnapped children in pursuit of the “fountain of youth” is gullible.

Do you think the bulk of voters for Trump believe one of these things? He won a considerable number of votes.

I think a large proportion of Trump voters believe parts of QAnon-pushed conspiracy theories - especially that he is "taking on a Deep State conspiracy", and data indicates some support for that[1][2].

Surely they don't believe the "blood drinking paedophile" part though? It seems outrageous, and yet I know a bunch of people on HN believed PizzaGate back in 2016, so maybe I'm underestimating it.

[1] https://www.usatoday.com/story/news/politics/2020/10/22/qano...

[2] https://www.forbes.com/sites/tommybeer/2020/09/02/majority-o...

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

#209

Post by noncoml »

drilldrive wrote: post #206

Frankly, this has an obvious answer: The people who answer polls are not representative of the population at large. Yet this was known decently enough since 2016. But more at the heart of the issue here is pollsters are providing a service not to inform but to persuade (both the public and donors); internal-polling is actually what matters. And instead of reformulating the service with more accurate questions for tot…

> it is clear they just should be abolished

I am not sure I understand what you mean. Abolished how?

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

#210

Post by Der_Einzige »

kjeetgill wrote: post #188
flowerlad wrote: post #183

Earlier quoted context omitted.

Not bad people, but gullible people. Anyone that believes QAnon conspiracy theories such as Hillary Clinton, Bill Gates, and Oprah Winfrey are raping and drinking the blood of kidnapped children in pursuit of the “fountain of youth” is gullible.

Do you think the bulk of voters for Trump believe one of these things? He won a considerable number of votes.

I'm guessing but I bet that about 30-40% of his voters believe it.
Post reply on HN