One pollster’s explanation for why the polls got it wrong • Hacker News Board

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

#511

Post by rayiner » Wed, Nov 11, 2020, 4:37 PM UTC

Dumblydorr wrote: post #486

lawnchair_larry wrote: post #367

Earlier quoted context omitted.

> I don't understand what exactly the people restricting speech in this way are trying to achieve. When your ideas can’t be sustained on their merits, your only option is to eliminate people who don’t share your world view. It’s a complete disaster and only breeds fragility.

I will attempt to speak their position, though I disagree with the left on identity politics. They might respond that, as a nation, US slaves and native americans and their descendants were exploited brutally by white English colonists. Over generations, this disparity was solidified by white male laws like jim crow, voter disenfranchisement, and overtly racist and harmful attitudes. Where they go wrong, IMO, is what…

> They might respond that, as a nation, US slaves and native americans and their descendants were exploited brutally by white English colonists. Over generations, this disparity was solidified by white male laws like jim crow, voter disenfranchisement, and overtly racist and harmful attitudes.

Very few people would deny this, and I highly doubt Shor disagrees with any of this.

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

#512

Post by Green_man » Wed, Nov 11, 2020, 4:38 PM UTC

nxmnxm99 wrote: post #506

Green_man wrote: post #325

Earlier quoted context omitted.

Trafalgar almost certainly got more states wrong (counts aren't all finished, but we do know likely winners in all states). Other sites like 538 predicted all states correctly, except NC and FL (50.5%/48.8% and 50.9%/48.4% respectively, both predictions wrongly favoring Biden [1]). Meanwhile, Trafalgar's polls put Trump up in NV, GA, PA, AZ, and MI (from their website, all late Oct / early Nov polls [2]). Once counts…

You're not looking at the deltas. > but it's important to note that the Trafalgar polls were mostly outliers, disagreeing with the consensus in many states Well, yeah?

We can't measure the deltas with as much confidence, because counting is far from finalized in many states (especially non swing states). But if trafalgar is the only group putting out polls where trump is ahead in several states, and then those states are lost, isn't that meaningful? Trafalgar was putting a "shy trump voter" bias into their polls, basically shoving the polls arbitrarily right. Some of that may have helped with polling errors that hurt other pollsters, but its clear they went too far, especially with many midwest states that ended up going blue.

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

#513

Post by fastball » Wed, Nov 11, 2020, 4:45 PM UTC

kaitai wrote: post #360

adventured wrote: post #333

Earlier quoted context omitted.

> The bulk of Trump voters are from rural areas, and tend to be less (or un-)educated. You're of course ignoring even worse problems on the Democrat side. The lowest education rates and highest drop-out rates in the US are not in rural areas, they're almost exclusively in the inner cities where Democrats tend to completely dominate the vote. Rural students at 87% are more likely to graduate from high school in four y…

I feel like your answer is sort of in bad faith. I live in an urban area, and people are quite concerned about the dropout rate (and more to the point that I think you're trying to imply but are dancing around for your own reasons, the racial achievement gaps). But I guess we put a different spin on it: we are ashamed of it, we should be ashamed of it, and we bring it up a lot in political races and on billboards and…

I think the point GC was making is also that rural people see how poorly democrats are running cities in many cases and don't want that to happen to them.

The logic being: better little-to-no government at all than one that does a terrible job.

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

#514

Post by Green_man » Wed, Nov 11, 2020, 4:47 PM UTC

zimablue wrote: post #488

Green_man wrote: post #325

Earlier quoted context omitted.

Trafalgar almost certainly got more states wrong (counts aren't all finished, but we do know likely winners in all states). Other sites like 538 predicted all states correctly, except NC and FL (50.5%/48.8% and 50.9%/48.4% respectively, both predictions wrongly favoring Biden [1]). Meanwhile, Trafalgar's polls put Trump up in NV, GA, PA, AZ, and MI (from their website, all late Oct / early Nov polls [2]). Once counts…

Number of states that you get wrong isn’t a great measure though right? If I call if for trump at 51-49 and you call it for Biden at 75-25 and Biden wins by 51-49 I was more correct than you were, to count “states called” is the wrong way to evaluate analysis in the same way that fptp is unrepresentatove of the popular vote, ironically.

I agree with you, and that's why i made effort to differentiate between margins vs winning. In my opinion, 2020's polls were a lot worse than 2016's, but because of Biden's much bigger lead (in the polls and in key states), the polling errors did not cross the threshold to be "wrong" in the winner take all system. Finally, there is an argument that even a Biden 75-25 call gives the correct outcome because each state is a winner take all system (with the exception of NE and ME, iirc). From a stats standpoint, the delta in that example make it very inaccurate, but from a model/communication standpoint, getting each state "right" (in binary terms) is very valuable.

Trafalgar had completely different results than the rest of the consensus, and once counts are more finalized in non swing states (places like CA and NY take forever, but we don't pay attention because they are always blue), we can accurately compare the polling and real results between different pollsters and polling methodologies. Some experts have cautioned against using exit polls for this purpose (what is usually done for a quick read of polling accuracy), because exit polls are only measuring election day in person votes, and thus trend really red this year. Trafalgar had unconventional methodologies like a "shy trump voter" bias where they arbitrarily shifted their numbers to the right, with weirdly consistent results of Trump +3 in a bunch of close/blue states. Perhaps there is more complicated justification of these numbers in the background, but I'm concerned that even if their deltas end up better than the "normal" pollsters, they are generating an inferior product with overbaked data.

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

#515

Post by Animats » Wed, Nov 11, 2020, 5:10 PM UTC

istjohn wrote: post #358

Animats wrote: post #186

Decades ago, phone polling got over a 50% response rate. Now it's under 10%. So there's at least a bias towards people who are willing to accept unsolicited calls.

The article says the response rate is just 1%.

That must be newer data. At 1%, the data is mostly noise.

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

#516

Post by astura » Wed, Nov 11, 2020, 5:11 PM UTC

mastazi wrote: post #443

vkou wrote: post #411

Earlier quoted context omitted.

Do you not think that the mainstream Republican party has moved radically rightward over the past 12 years? It's not spending its energy discussing Laffer curves and energy policy, and whether or not someone can be fired from their job for being gay, or if we should invade another Middle Eastern country anymore. It's spending its energy seriously talking about QAnon, and how the COVID pandemic is over, and and how th…

> But we no longer live in a sane world One way to deal with this type of dysphoria would be to spend some quality time with people who hold views different from the ones you have, and it would be much better if you didn't do that online, but in person - for reasons that I'm sure you understand well, since you are an HN reader. Or, you decide that everyone on the other side of the fence is completely bollocks. But wi…

I've had to end several friendships because after a while I couldn't stand the random racist rants that have nothing to do with the topic at hand. One person even shouted they want to "mow those niggers down" because there were people with darker pigmented skin were merely existing out in public and we were driving by them - they were doing absolutely nothing, only breathing. You complain about something sad or tragic happening to you and their response is to suggest it must be the fault of black people, no matter what the problem actually is.

I'm not talking about one or two people here either, I'm talking about several people, family and former friends, many who don't know each other and live in different geographic areas.

I just can't with that shit anymore.

Honestly, the more time I spend with people who have different beliefs than me the more I dislike them.

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

#517

Post by boomboomsubban » Wed, Nov 11, 2020, 5:14 PM UTC

ineedasername wrote: post #508

boomboomsubban wrote: post #387

Earlier quoted context omitted.

This is changing it from "they said to stop" to "they said it was wise to stop," and the distinction basically only matters if you want the Democrats to win. You're saying the same thing either way.

No, it really doesn't. You're making an extra leap in logic that isn't true for all people. Just because the protests could lose some votes doesn't mean everyone who acknowledge that possibility wanted them to stop. Pick any issue. A person can support the issue and recognize it may have negative consequences. From there I'm making the distinct that beliefs on the issue can go (at least) two ways: 1) The issue is too…

>A person can believe the lost votes are not important enough to override the importance of the protests.

That person will spend their time discussing the issues important to the protests, not spend their time discussing how the protests are actually hurting their greater cause.

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

#518

Post by Falling3 » Wed, Nov 11, 2020, 5:19 PM UTC

nl wrote: post #318

nxmnxm99 wrote: post #308

Earlier quoted context omitted.

You’re being downvoted because the “democrats are le smart and well read” trope is getting old

What does that mean? I think Democratic voters are vulnerable to a different kind of conspiracy theory. For example, huge numbers of left-leaning voters seem to believe that money is the determining factor in elections, whereas the evidence points towards you needing a certain amount, but beyond that it appears to be of marginal value.

How is overestimating the effect of money on election outcomes a conspiracy theory?

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

#519

Post by Falling3 » Wed, Nov 11, 2020, 5:24 PM UTC

thu2111 wrote: post #503

nl wrote: post #208

Earlier quoted context omitted.

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.usato…

You shouldn't be downvoted. Trump "taking on the deep state" isn't even a conspiracy theory, he literally made that a major part of his 2016 campaign. If deep state/swamp/etc is interpreted as a vague reference to the Washington/New York professional classes who populate government in huge numbers and tend to share very similar worldviews and beliefs, then calling this a QAnon pushed conspiracy theory would be a pret…

> Trump "taking on the deep state" isn't even a conspiracy theory, he literally made that a major part of his 2016 campaign

You're misframing this. The conspiracy theory is not that Trump is taking on the deep state; it's the existence of the deep state to begin with.

>If deep state/swamp/etc is interpreted as a vague reference to the Washington/New York professional classes who populate government in huge numbers and tend to share very similar worldviews and beliefs

Speaking of grotesque distortions of the truth, that is not what the "deep state" has ever referred to. It's always been, by definition, a conspiracy theory. It generally posits that the country is actually run by a group o unelected individuals in secret that orchestrate events in this country up to and including the selection of presidents.

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

#520

Post by rayiner » Wed, Nov 11, 2020, 5:27 PM UTC

mattkrause wrote: post #439

rayiner wrote: post #219

Earlier quoted context omitted.

Maybe it’s semantics, but to me that means something is wrong with the model. It’s not just a matter of an unlikely result being happening some percentage of the time. There is some hidden factor that could have been adjusted for ahead of time.

Both, I think. The model certainly ought to handle correlations, but the evaluation should take them into account too. You can't treat the election as fifty independent replicates: a miss in both Minnesota and Wisconsin is clearly worse than one error, but it's also not as bad as (say) getting Wisconsin and Rhode Island wrong, where it's more likely that two separate errors occurred.

I understand the point mathematically, I think I'm talking about more of how the presentation of the error term is unintuitive to me. It makes sense to me to say "the predications were right, Trump just got lucky" when he outperforms the 95th percentile error bars in a state. He just got lucky. But if he outperforms it in a bunch of states because the polls didn't account for the fact that Trump supporters don't answer pollsters, then it doesn't make sense to me to say "the predictions were right because they factored in the chance that the polls hard correlated errors." I get that you've quantified the possibility that the polls are wrong in a systematic way, but I don't think that's the kind of possibility people are thinking of when they hear "there is a 5% chance Trump could still win."