U.S. presidential election modelling

[iNow]

4 hours ago, ScienceNostalgia101 said:

All this math and so little emphasis on the possibility that poll respondents were simply lying through their teeth?

The more parsimonious explanation involves the move away from landline telephones and the way people don’t answer calls from unknown numbers.

When people move from one state to another, they tend to take their phone number with them and those numbers don’t match local area codes. Consequently, they get missed in random samples dialed by pollsters.

Sure, some people lie when reached and asked questions, but that’s less common and bleeds out as noise with a sufficiently large sample size.

Maybe you should go post this as a YouTube comment somewhere so the entire world may benefit from the correction. 

[ScienceNostalgia101]

On 12/31/2020 at 11:04 PM, iNow said:

The more parsimonious explanation involves the move away from landline telephones and the way people don’t answer calls from unknown numbers.

When people move from one state to another, they tend to take their phone number with them and those numbers don’t match local area codes. Consequently, they get missed in random samples dialed by pollsters.

Sure, some people lie when reached and asked questions, but that’s less common and bleeds out as noise with a sufficiently large sample size.

Maybe you should go post this as a YouTube comment somewhere so the entire world may benefit from the correction. 

I'm pretty sure anyone really committed to figuring out people's real opinions from online comments wouldn't so easily mistake an individual for the crowd. Especially someone who doesn't claim to speak for everyone... and on occasion, actively claims not to.

 

To whatever extent phone polls can be blamed for getting 2016 wrong, that leaves behind the question of why pollsters didn't bother to find better sampling methods. Could it be that, deep down, no matter how strong their sample, they have no way of predicting an imbalance between people who tell one lie and people who tell the opposite?

[iNow]

10 hours ago, ScienceNostalgia101 said:

Could it be that, deep down, no matter how strong their sample, they have no way of predicting an imbalance between people who tell one lie and people who tell the opposite?

Unlikely.

 

10 hours ago, ScienceNostalgia101 said:

To whatever extent phone polls can be blamed for getting 2016 wrong, that leaves behind the question of why pollsters didn't bother to find better sampling methods

What do you recommend is a better sampling method?

[J.C.MacSwell]

I suspect many, even in an anonymous poll, would be unwilling to admit they could vote for Trump, even if it was due to disagreeing with the positions of his opponents.

This might lead many to lie or be unwilling to participate in polling (and this unwillingness alone would skew the results, even if everyone being polled was honest)

I can't see the equivalent effect for eventual Biden voters.

 

[iNow]

6 hours ago, J.C.MacSwell said:

I suspect many, even in an anonymous poll, would be unwilling to admit they could vote for Trump <...> This might lead many to lie or be unwilling to participate in polling

Yes, definitely in my experience these last several years it’s become clear that most Trump supporters are a rather shy bunch of wallflowers who prefer keeping quiet and hiding their feelings about Trump from others 🙄

[ScienceNostalgia101]

13 hours ago, iNow said:

Unlikely.

 

What do you recommend is a better sampling method?

I would recommend watching any potential source of spontaneous sincerity like a hawk; whether through Internet culture or otherwise; to circumvent people's tendency to lie, and, keeping all plausible interpretations of what one observed in mind, use evolutionary reasoning to narrow down the interpretations as best as possible.

 

Would that still be open to interpretation? Sure. But so is "3 separate photographs of 3 separate crowds of Trump supporters, any of which might be doing this in a district that was known to have voted Trump last election, and none of which have any reason to believe that this would make Biden let his guard down as easily as lying to a pollster did for Clinton in 2016."

[J.C.MacSwell]

6 hours ago, iNow said:

Yes, definitely in my experience these last several years it’s become clear that most Trump supporters are a rather shy bunch of wallflowers who prefer keeping quiet and hiding their feelings about Trump from others 🙄

My bad. I forgot all 70+ million Trump voters all went out and rallied.

Fortunately Biden could draw crowds as well:

[iNow]

7 hours ago, ScienceNostalgia101 said:

I would recommend watching any potential source of spontaneous sincerity like a hawk; whether through Internet culture or otherwise; to circumvent people's tendency to lie, and, keeping all plausible interpretations of what one observed in mind, use evolutionary reasoning to narrow down the interpretations as best as possible.

Thanks for confirming that you have no idea how sampling works with this vague inoperative reply which itself ignores how many of the things you suggest are already being done.

5 hours ago, J.C.MacSwell said:

My bad. I forgot all 70+ million Trump voters all went out and rallied.

You seem to be suggesting that one must physically attend a rally or else qualify for you’re “too shy to anonymously tell a pollster they’re voting for Trump” population. I’m sure we both agree that’s silly. 

I’m sure some fraction of people polled chose to hide their support for Trump when asked. Nobody disputes that. However, you’re suggesting that enough people did this in the last 2 elections to totally skew the results of polls across the board conducted by different organizations. To me, that’s self-evidently absurd and I’ve already offered a more parsimonious explanation for the failure in polls above. The people who lie are marginal at best, and their responses tend to average out as noise that has minimal to no impact on the sample.

Also, IMO the narrative that polls are “way off base“ is itself questionable. The polls are actually quite good overall.

 

https://fivethirtyeight.com/features/the-polls-werent-great-but-thats-pretty-normal/

Quote

On average, the final national polls were off by 2.3 points. That’s pretty close, but there are also several examples of polling error comparable to this year. There were 4-point polling misses in 1996 and 2012 (as mentioned, the national polls weren’t so good that year), a 5-point error in 2000 and a whopping 8-point miss in 1980, when Ronald Reagan beat Jimmy Carter by far more than polls predicted.

<...>

Evaluating all this data, our model estimates that the final national polls will miss by about 3 percentage points in an average year. (That’s why we call a 3-point miss a “normal-sized polling error.”) In other words, a 4-point polling error is somewhat par for the course

<...>

The main reason that polls aren’t going to provide you with the certitude you might desire is because polls have always come with a degree of uncertainty. In a highly polarized era, most elections are going to be close — close enough as to exceed the ability of polls to provide you with a definitive answer. Say the final polling averages miss by a bit more than 3 points on average, as our forecast assumes. That means the margin of error is closer to 7 or 8 points. And every presidential election so far this century has fallen within that range.1 So if you’re coming to the polls for strong probabilistic hints of what is going to happen, they can provide those — and the hints will usually lead you in roughly the right direction, as they did this year. But if you’re looking for certainty, you’ll have to look elsewhere.

Finally, even if I choose to accept your premise that polls are bad, the explanation of this being due to shy Trump voters who are lying when polled simply doesn’t add up. The problem has more to do with people not answering polls at all than it does with people agreeing to answer them, but lying to the pollster when they do.

This gets amplified by the difficulty in reaching a random sample in the first place due to the huge move away from landline phones and the deeper challenge of associating portable mobile phone numbers (think: area codes) with specific regions and locations. 
 

https://www.vox.com/policy-and-politics/2020/11/10/21551766/election-polls-results-wrong-david-shor

Quote

the kind of people who answer polls are systematically different from the kind of people who refuse to answer polls — and that this has recently begun biasing the polls in a systematic way. <...> Fundamentally, every “high-quality public pollster” does random digit dialing. They call a bunch of random numbers, roughly 1 percent of people pick up the phone, and then they ask stuff like education, and age, and race, and gender, sometimes household size. And then they weight it up to the census, because the census says how many adults do all of those things. That works if people who answer surveys are the same as people who don’t, once you control for age and race and gender and all this other stuff, but we can no longer accurately assume the people not answering are equivalent to those who do.

<...>

A lot of people think that the reason why polls were wrong was because of “shy Trump voters.” You talk to someone, they say they’re undecided, or they say they’re gonna vote for Biden, but it wasn’t real. Then, maybe if you had a focus group, they’d say, “I’m voting for Biden, but I don’t know.” And then your ethnographer could read the uncertainty and decide, “Okay, this isn’t really a firm Biden voter.” That kind of thing is very trendy as an explanation.

But it’s not why the polls were wrong. It just isn’t. People tell the truth when you ask them who they’re voting for. They really do, on average. The reason why the polls are wrong is because the people who were answering these surveys were the wrong people. 

 

Edited January 3, 2021 by iNow

[J.C.MacSwell]

57 minutes ago, iNow said:

 

You seem to be suggesting that one must physically attend a rally or else qualify for you’re “too shy to anonymously tell a pollster they’re voting for Trump” population. I’m sure we both agree that’s silly. 

 

Absolutely. 

57 minutes ago, iNow said:

 

You seem to be suggesting that one must physically attend a rally or else qualify for you’re “too shy to anonymously tell a pollster they’re voting for Trump” population. I’m sure we both agree that’s silly. 

 

That was what you seem to have suggested.

Quoting this:

19 hours ago, J.C.MacSwell said:

I suspect many, even in an anonymous poll, would be unwilling to admit they could vote for Trump<...> This might lead many to lie or be unwilling to participate in polling

 

You said this:

12 hours ago, iNow said:

Yes, definitely in my experience these last several years it’s become clear that most Trump supporters are a rather shy bunch of wallflowers who prefer keeping quiet and hiding their feelings about Trump from others 🙄

I recognized your silly point and chose to go along with it.

I apologize if you were actually serious. (I don't think you were but why believe I was?)

Back to my (and Science Nostalgia's) point, which I didn't think would be controversial:

(Didn't take much of a search)

‘People Are Going To Be Shocked’: Return of the ‘Shy’ Trump Voter?

https://www.politico.com/news/magazine/2020/10/29/2020-polls-trump-biden-prediction-accurate-2016-433619

“We live in a country where people will lie to their accountant, they’ll lie to their doctor, they’ll lie to their priest,” says Cahaly. “And we’re supposed to believe they shed all of that when they get on the telephone with a stranger?”

Edited January 3, 2021 by J.C.MacSwell

[iNow]

Hmm. I was using sarcasm. Maybe that’s the source of the misunderstanding. 

You said: “I suspect many, even in an anonymous poll, would be unwilling to admit they could vote for Trump.”

The meaning of my reply: “That doesn’t seem to align with the behaviors we actually tend to see from Trump voters. This doesn’t seem to be a valid explanation for polling misses.”

[J.C.MacSwell]

“We live in a country where people will lie to their accountant, they’ll lie to their doctor, they’ll lie to their priest,” says Cahaly. “And we’re supposed to believe they shed all of that when they get on the telephone with a stranger?”

50 minutes ago, iNow said:

Hmm. I was using sarcasm. Maybe that’s the source of the misunderstanding. 

You said: “I suspect many, even in an anonymous poll, would be unwilling to admit they could vote for Trump.”

The meaning of my reply: “That doesn’t seem to align with the behaviors we actually tend to see from Trump voters. This doesn’t seem to be a valid explanation for polling misses.”

Clearly then, not the homogenous group you seem to be implying the Trump voters are. Though surely you would have known any "shy" Trump voters would be underrepresented in any photos of mobs yielding Trump signs. 

The polls vs actual results have some reason or reasons. That's my opinion (as well as others as per the link)

The further point I was making in addition to SNostalgia's is that they, the "shy" Trump voting group, don't even have to lie...they'll skew the results simply by refusing to be polled due to their dislike of Trump (or simple embarrassment to admit they actually like him)

Let me know what you think of the article, and whether you agree or not with the gist of it. It was written late October before the election.

Edited January 3, 2021 by J.C.MacSwell

[J.C.MacSwell]

One more day and Trump will no longer be a useful idiot of the GOP, as far as current elections go, when the balloting closes in Georgia.

Current polling puts both Democrat candidates leading marginally.

The results could reflect on the discussion, but I'm not sure in what way.

[iNow]

On 1/3/2021 at 9:36 AM, J.C.MacSwell said:

Clearly then, not the homogenous group you seem to be implying the Trump voters are.

 

On 1/3/2021 at 8:39 AM, iNow said:

I’m sure some fraction of people polled chose to hide their support for Trump when asked. Nobody disputes that. However, you’re suggesting that enough people did this in the last 2 elections to totally skew the results of polls across the board conducted by different organizations. To me, that’s self-evidently absurd and I’ve already offered a more parsimonious explanation for the failure in polls above. The people who lie are marginal at best, and their responses tend to average out as noise that has minimal to no impact on the sample.

 

 

On 1/3/2021 at 9:36 AM, J.C.MacSwell said:

Let me know what you think of the article, and whether you agree or not with the gist of it. It was written late October before the election.

I think the narrative of a shy Trump voter explaining misses in polls is itself misguided, even though it's become so commonly shared. Instead, getting people to answer their phones in the first place is the bigger deal. Those who do answer their phones are decreasingly likely to be representative of those who don't. 

This is not party-specific nor in any way isolated to Trump.

4 hours ago, J.C.MacSwell said:

Current polling puts both Democrat candidates leading marginally.

It's still within the margin of error, and given what we've seen at the presidential level, it won't surprise me at all if we don't have to wait a few more weeks watching false astroturfed claims of fraud getting made and empty court cases get filed before we can say who won.

Edited January 4, 2021 by iNow

[J.C.MacSwell]

On 1/3/2021 at 10:39 AM, iNow said:

 

I’m sure some fraction of people polled chose to hide their support for Trump when asked. Nobody disputes that. However, you’re suggesting that enough people did this in the last 2 elections to totally skew the results of polls across the board conducted by different organizations. To me, that’s self-evidently absurd and I’ve already offered a more parsimonious explanation for the failure in polls above. The people who lie are marginal at best, and their responses tend to average out as noise that has minimal to no impact on the sample.

 

 

As I've suggested, it's not just people who lie. You have to include the one's who are simply unwilling to participate in the polls. They too skew the results.

I don't know what explanation you have offered. Is there more than just this?:

20 minutes ago, iNow said:

 

 

I think the narrative of a shy Trump voter explaining misses in polls is itself misguided, even though it's become so commonly shared. Instead, getting people to answer their phones in the first place is the bigger deal. Those who do answer their phones are decreasingly likely to be representative of those who don't. 

 

 

Why would you think that would lead to a consistent Democrat bias in polling results? (not to be interpreted as intentional bias)

[iNow]

13 minutes ago, J.C.MacSwell said:

Why would you think that would lead to a consistent Democrat bias in polling results? (

As a general rule, democratic voters tend to be more focused on community and social issues whereas republican voters tend to be more focused on the self and personal liberty issues. This psychologically extends to a basic willingness to offer their unpaid time helping with a survey (volunteering to help) versus being untrustworthy of the pollster (more isolated and protectionist in thinking style). 

[J.C.MacSwell]

2 minutes ago, iNow said:

As a general rule, democratic voters tend to be more focused on community and social issues whereas republican voters tend to be more focused on the self and personal liberty issues. This psychologically extends to a basic willingness to offer their unpaid time helping with a survey (volunteering to help) versus being untrustworthy of the pollster (more isolated and protectionist in thinking style). 

It's a small percentage effect, so even if the tendency is a slight one this can play a part. 

[CharonY]

I have not read enough about that, but it is possible that different methodologies only reduce the error to ~5% (give or take). For a variety of types of polls this is actually fine. However, to the inflexibility of the US system, where huge chunk of the electorate are more or less fixed, there is only a narrow margin of votes going to determine outcomes. In turn, this might mean that polls as a whole are not terribly predictive for many US elections, unless there are big margins in swing states.

[iNow]

Again though, the polls have been largely accurate. The narrative that they’re not isn’t very well supported (I posted about this above). 

They give likely tendencies, not definitive predictions. 

Edited January 4, 2021 by iNow

[J.C.MacSwell]

1 hour ago, iNow said:

Again though, the polls have been largely accurate. The narrative that they’re not isn’t very well supported (I posted about this above). 

They give likely tendencies, not definitive predictions. 

As per CY's post above, the question is how accurate they are at predicting outcomes. They can get that wrong even when being generally or largely accurate.

The fact remains that they tend to slightly underestimate GOP support, for some reason/s.

[iNow]

I feel I’d be repeating myself if I respond here to those points 

[beecee]

Can I ask a question here? Why the bloody hell, does the US leave a defeated President in office for 2 months, before installing the bloke [Biden] that defeated him? Is this something to do with the US constitution? I understand that vote counting can take some days or even weeks in certain circumstances, if counting is close...that happens everywhere. I just cannot understand for the life of me, why Trump after losing all his so called challenges, can still linger in that position, and potentially create and pass new laws and legislation that can stifle the incoming President. Can someone please educate me as to this apparent anomaly in US politics?

[Janus]

13 minutes ago, beecee said:

Can I ask a question here? Why the bloody hell, does the US leave a defeated President in office for 2 months, before installing the bloke [Biden] that defeated him? Is this something to do with the US constitution? I understand that vote counting can take some days or even weeks in certain circumstances, if counting is close...that happens everywhere. I just cannot understand for the life of me, why Trump after losing all his so called challenges, can still linger in that position, and potentially create and pass new laws and legislation that can stifle the incoming President. Can someone please educate me as to this apparent anomaly in US politics?

Yes, it is in the constitution. And remember when the Constitution was written.  Vote results had to travel by land to a centralized point in each state, then counted. Then electors selected, gathered together, and cast their votes.  These votes had to be delivered from each state to congress without any quick means to do so.  Once Congress got and counted the elector votes, they could announce the winner, who, if he wasn't already in the capital, would have to travel there to be inaugurated.    The whole process just took time, and the constitution was written to account for this.

To change it would require amending the constitution.

 

 

[beecee]

24 minutes ago, Janus said:

Yes, it is in the constitution. And remember when the Constitution was written.  Vote results had to travel by land to a centralized point in each state, then counted. Then electors selected, gathered together, and cast their votes.  These votes had to be delivered from each state to congress without any quick means to do so.  Once Congress got and counted the elector votes, they could announce the winner, who, if he wasn't already in the capital, would have to travel there to be inaugurated.    The whole process just took time, and the constitution was written to account for this.

To change it would require amending the constitution.

 

 

Thanks for that...makes sense. But then again, why cannot the constitution be changed to suit the times? [Probably also  many other areas of it could see the USA benefit by changes] 

[StringJunky]

24 minutes ago, beecee said:

Thanks for that...makes sense. But then again, why cannot the constitution be changed to suit the times? [Probably also  many other areas of it could see the USA benefit by changes] 

There are named amendments, so it has been in the past. I think the stars haven't been aligned in terms of having the necessary House and Senate votes to have allowed that to happen in recent history, AFAIK

Edited January 6, 2021 by StringJunky

[Janus]

49 minutes ago, beecee said:

Thanks for that...makes sense. But then again, why cannot the constitution be changed to suit the times? [Probably also  many other areas of it could see the USA benefit by changes] 

It can, But the amendment to do so would first have either get a 2/3 majority support in both Houses (House of Rep and Senate) or be applied for by the state legislatures in 2/3(34) of the states. In which case Congress must call a convention.

An amendment proposed by either of these means would then have to get ratification from the legislatures  of 3/4(38) of the states.