Spring forward.
To delete this message, click the X at top right.

Is Political Polling Meaningful Anymore?

Trying to make sense of a crazy world, with limited success mostly
Slarti the White
Posts: 442
Joined: Mon Feb 22, 2021 1:23 pm
Location: Michigan
Verified: Badges... we don't need no stinkin' badges

Is Political Polling Meaningful Anymore?

#1

Post by Slarti the White »

In 2008 Nate Silver and 53 took poll aggregation mainstream with their highly accurate predictions of the presidential election. It seemed that we had entered the age where polling was becoming rigorous science. Now we have articles like these from The Hill and Wonkette:

Are we facing another big polling miss? Probably.

'Red Wave' Narrative May Be Built On Crap Polling? Color Us Shocked.

How did this incipient 'Golden Age' turn to ashes so quickly and why?

I think, as with the rise of Trump, it was mainly caused by denialism regarding major changes in our political environment which left the mainstream of experts and pundits with now power to combat the norm-breaking and false narratives pushed by the Trumpists. Everyone knew about the oncoming issues with the basic assumptions in polling, even in 2008. Landlines were becoming less numerous and the demographics of the remaining people with landlines becoming more and more divorced from the demographics of the nation as a whole meant that samples were becoming less representative. Furthermore, changing demographics and changes within demographics (think of Trump's increase in support amongst African-Americans and Hispanic voters from 2016 to 2020) meant that the demographic models used to correct for unrepresentative samples were also becoming less accurate.

Into this environment of decaying polling reliability came the tempest of the politics of hate and divisiveness turning all of the basics assumptions on their head -- many people were no longer voting for the candidate they liked the best, but were rather voting against the candidate that they hated more. If we had examined the basic statistical assumptions behind all of the polling and aggregating methodology we were using we might have been able to foresee the problem, but we likely still would have been blindsided by the rise of the authoritarian right and the pernicious and corrosive effect of election denialism on our democracy.

At this point, I think that any value in standard polling has been lost because the polling narrative has been entirely coopted by the authoritarian right. As Wonkette says, Republican leaning polling houses have flooded the zone in recent weeks. After the election it's easy to see what's coming: If there is a "Red Wave", every poll that didn't predict it will be held up as evidence of how the Democrats were trying to rig the elections (but were stopped by the valiant efforts of groups like 'True the Vote'). If the Red Wave fails to materialize, all of those polls will "prove" that efforts to rig the elections were successful and we'll have the election denialist crowd ramping up their efforts to incite violence and tear down the system and any faith people have left in it.

In any case, I think it behooves us to start having an honest conversation about the state of political polling today, how the authoritarian right might use this opportunity to further erode our institutions and the people's faith in them, and what alternatives might help restore the value and credibility of the polling industry. Otherwise, just like the media who flocked around Trump for all of the ratings and page clicks he brought, once we realize the problem it will be too (not also) late to do anything about it.

Just my $0.01999999999...
:towel:
User avatar
Slim Cognito
Posts: 6552
Joined: Mon Feb 22, 2021 9:15 am
Location: Too close to trump
Occupation: Hats. I do hats.
Verified:

Is Political Polling Meaningful Anymore?

#2

Post by Slim Cognito »

Please, oh please, oh please...
Pup Dennis in training to be a guide dog & given to a deserving vet. Thx! ImageImageImage x4
User avatar
Foggy
Dick Tater
Posts: 9554
Joined: Mon Feb 22, 2021 8:45 am
Location: Fogbow HQ
Occupation: Dick Tater/Space Cadet
Verified: as seen on qvc zombie apocalypse

Is Political Polling Meaningful Anymore?

#3

Post by Foggy »

This thread needs a poll, what is even wrong with you people?

A meaningful political poll. Or else! Or else something!

:oldman:

:hoover:

:groupdance:

:mbounce:
Out from under. :thumbsup:
User avatar
Suranis
Posts: 5828
Joined: Mon Feb 22, 2021 5:25 pm

Is Political Polling Meaningful Anymore?

#4

Post by Suranis »

Honestly, I've been expecting the polls to be swamped at some point. Up till now the GOD have been content to let a small group like Rasmussen pull the averages a bit their way, but the problem with that is that People like Nate Silver had attached ratings to pollsters, so Rasmussen is no linger effective as Silver's model attaches ratings to pollsters so he corrects their affect on the Polling average. Now a whole load of new polls are very effective as there's no data on them, so if SIlver or anyone else includes them they will have the full affect.

Poll swamping is an obvious tactic.

The Wonkette article is extremely good, people should read it as I would just be repeating what it says
Hic sunt dracones
User avatar
AndyinPA
Posts: 9853
Joined: Mon Feb 22, 2021 10:42 am
Location: Pittsburgh
Verified:

Is Political Polling Meaningful Anymore?

#5

Post by AndyinPA »

I haven't had much faith in the polls since 2000. I think that's the year exit polling began to deviate from "polling."
"Choose your leaders with wisdom and forethought. To be led by a coward is to be controlled by all that the coward fears… To be led by a liar is to ask to be told lies." -Octavia E. Butler
Danraft
Posts: 492
Joined: Mon Feb 22, 2021 1:43 pm
Location: Michigan
Contact:

Is Political Polling Meaningful Anymore?

#6

Post by Danraft »

Slarti the White
Posts: 442
Joined: Mon Feb 22, 2021 1:23 pm
Location: Michigan
Verified: Badges... we don't need no stinkin' badges

Is Political Polling Meaningful Anymore?

#7

Post by Slarti the White »

Danraft,

While that's obviously making it clear that a "red wave" hasn't happened yet and promising for the Democrats, it still doesn't really allow us to make accurate predictions. This is the first post-COVID midterms so the early vs. election-day split is really just a matter of guesswork. Too, also, regardless of the outcome this info is going to be used by the Republicans to drive their election-denial agenda. I don't know that there is any way to fight that within the context of polling as we know it. Right now it seems that we are one "Red Wave" away from authoritarianism -- unless we break out of the paradigm which has been thoroughly gamed by the Republicans I don't see how that threat can be defused.
User avatar
June bug
Posts: 721
Joined: Tue Feb 23, 2021 7:34 am

Is Political Polling Meaningful Anymore?

#8

Post by June bug »

Republicans have been encouraged and/or propagandized and/or instructed not to use mail-in ballots or vote early. Election Day in-person voting is their mantra, so I don’t think pre-election party voter differentials are trustworthy predictors of overall turnout.
User avatar
Slim Cognito
Posts: 6552
Joined: Mon Feb 22, 2021 9:15 am
Location: Too close to trump
Occupation: Hats. I do hats.
Verified:

Is Political Polling Meaningful Anymore?

#9

Post by Slim Cognito »

I was on Marketwatch yesterday and pretty much every opinion was the Rs are going to sweep everything bigly. Is there another site that keeps it nonpartisan?
Pup Dennis in training to be a guide dog & given to a deserving vet. Thx! ImageImageImage x4
Slarti the White
Posts: 442
Joined: Mon Feb 22, 2021 1:23 pm
Location: Michigan
Verified: Badges... we don't need no stinkin' badges

Is Political Polling Meaningful Anymore?

#10

Post by Slarti the White »

June bug wrote: Sat Nov 05, 2022 10:44 am Republicans have been encouraged and/or propagandized and/or instructed not to use mail-in ballots or vote early. Election Day in-person voting is their mantra, so I don’t think pre-election party voter differentials are trustworthy predictors of overall turnout.
Exactly. We can collect the data and mine it for information, but we cannot develop reliable intelligence from it. It isn't that the information is bad, it's just that there are too (not also) many unknowns for us to say with certainty what it means.

Slim Cognito wrote: Sat Nov 05, 2022 11:45 am I was on Marketwatch yesterday and pretty much every opinion was the Rs are going to sweep everything bigly. Is there another site that keeps it nonpartisan?
It is easier to feed your confirmation bias if you want to predict a red wave and that may turn out to be accurate, but I don't believe that their certainty is justified. 538's analysis is non-partisan, but it doesn't matter how strong or objective your methodology is when the signal is swamped by this much noise (see what I did there? :towel: ). My point in this thread is that no one knows what will happen -- everyone is flailing around in a 100-dimensional parameter space (to quote a colleague of mine). Because there is no accountability for pundits and data wonks after the fact (everything slides down the memory hole) and the incentive is to get page clicks leading up to the election there is no feedback mechanism for scientifically determining the credibility of the any of the myriad voices out there claiming to be able to accurately predict what is going on. Right now I think the best course of action is to steal Fox Mulder's mantra: TRUST NO ONE and do whatever you can to move the needle toward the outcome you prefer. In any case, this cake is already pretty much baked.

But, to me, the question is not how to fix polling so it can accurately predict elections again, but rather what are we going to do -- starting Wednesday morning -- to create a new polling paradigm which will help shape election outcomes to better address the issues and principles that the electorate cares about. I'll share my ideas on that Wednesday morning... [moar will be revealed]
User avatar
SuzieC
Posts: 925
Joined: Mon Feb 22, 2021 10:25 am
Location: Blue oasis in red state
Occupation: retired lawyer; yoga enthusiast
Verified:

Is Political Polling Meaningful Anymore?

#11

Post by SuzieC »

I don't trust 538. I don't think they're non-partisan. They include Republican targeted polls in their aggregates which feeds the RED WAVE narrative. They were epically wrong on the Kansas vote, by 22 points.
User avatar
Suranis
Posts: 5828
Joined: Mon Feb 22, 2021 5:25 pm

Is Political Polling Meaningful Anymore?

#12

Post by Suranis »

SuzieC wrote: Sat Nov 05, 2022 3:57 pm I don't trust 538. I don't think they're non-partisan. They include Republican targeted polls in their aggregates which feeds the RED WAVE narrative. They were epically wrong on the Kansas vote, by 22 points.
I think Silver tried in the beginning, but he's lazy now.

He believes in bringing in every poll as a data point and adding a handicap to each poll to get it, but all these new polls have no data on accuracy so they have no rating in his model. So, ya his model is completely useless now that they are put in with no handicap.

Not that his handicaps were worth shit anyway. He had Rasmussen as +2 or +3 R back when I paid attention to him, which was just ludicrous
Hic sunt dracones
Slarti the White
Posts: 442
Joined: Mon Feb 22, 2021 1:23 pm
Location: Michigan
Verified: Badges... we don't need no stinkin' badges

Is Political Polling Meaningful Anymore?

#13

Post by Slarti the White »

SuzieC wrote: Sat Nov 05, 2022 3:57 pm I don't trust 538. I don't think they're non-partisan. They include Republican targeted polls in their aggregates which feeds the RED WAVE narrative. They were epically wrong on the Kansas vote, by 22 points.
While it is certainly up to you how much or how little credibility you put in 538, I don't think that there is any reason to believe that 538 has a partisan agenda or is in any way skewing their results with partisan bias. First off, Nate Silver is left-leaning (as is, I believe, most of his staff), but he is, first and foremost, a data wonk. If he is ever caught putting his finger on the scale, his professional credibility is destroyed. I do think 538 is doing their best to make accurate predictions using established methodologies. The problem is that their methodology is breaking down, not their integrity. They rank pollsters and account for "house effects" and then weight the adjusted polls accordingly. They are using a process that has been steadily (and transparently) evolving since they first came on the scene in 2008. If they started eliminating polls for ad hoc reasons at this point it would be a partisan act and an expression of lack of confidence in their methodology -- essentially making everything they say worthless.

As for Kansas, the question is "why were they epically wrong?". The obvious answer is that nothing in their methodology enabled them to account for the effect of the Dobbs decision -- again, more evidence that methodology is breaking down and no evidence at all of any partisan bias on the part of 538. There are structural problems in the polling industry and, so long as we don't acknowledge them, they will continue to be used by the authoritarian right to further erode confidence in our elections and our institutions. Accusing organizations like 538 of having a partisan agenda only strengthens this narrative in my opinion. The problem isn't that 538 is doing anything wrong, it is that the entire (political) polling industry is suffering from a GIGO (garbage in, garbage out) issue that cannot be addressed within the current paradigm. At least that's what I think.
User avatar
keith
Posts: 3705
Joined: Mon Feb 22, 2021 10:23 pm
Location: The Swamp in Victorian Oz
Occupation: Retired Computer Systems Analyst Project Manager Super Coder
Verified: ✅lunatic

Is Political Polling Meaningful Anymore?

#14

Post by keith »

A big part of 538 methodology is to allow for house bias based on historical results and methodology.

Excluding polls with no historical data points is not an ad hoc reason, it is self protecting their methodology.

And I dont see that happening. As you said GIGO.
Has everybody heard about the bird?
Slarti the White
Posts: 442
Joined: Mon Feb 22, 2021 1:23 pm
Location: Michigan
Verified: Badges... we don't need no stinkin' badges

Is Political Polling Meaningful Anymore?

#15

Post by Slarti the White »

Suranis wrote: Sat Nov 05, 2022 4:06 pm
SuzieC wrote: Sat Nov 05, 2022 3:57 pm I don't trust 538. I don't think they're non-partisan. They include Republican targeted polls in their aggregates which feeds the RED WAVE narrative. They were epically wrong on the Kansas vote, by 22 points.
I think Silver tried in the beginning, but he's lazy now.

I can see why you would think that, but I'd say that he "earned his bones" by doing something that no one had done before (and doing it very well). Since then, he's become the head of a very successful organization which is the leader in the field -- which is to say that he's gone from being the scrappy insurgent trying to prove himself to being the industry leader trying to maintain its position. It's no surprise that he's gotten much more conservative in his methodology.

He believes in bringing in every poll as a data point and adding a handicap to each poll to get it, but all these new polls have no data on accuracy so they have no rating in his model. So, ya his model is completely useless now that they are put in with no handicap.

I don't know how many of the polls we're referring to are from new pollsters, but presumably 538 already had a methodology for dealing with new pollsters and are applying it. You are right that he obviously can't adjust polls from a pollster with no history for "house effect", but the bigger problem is that even for known pollsters the quality of their sampling methodology and demographic models are breaking down. There is no possible way that can be accounted for.

Not that his handicaps were worth shit anyway. He had Rasmussen as +2 or +3 R back when I paid attention to him, which was just ludicrous

I do not at all doubt that result is what you get from the data and 538's methodology -- you can argue that the result is wrong, but, assuming 538 didn't alter the data or how the house effect is calculated for Rasmussen (or anyone else), they are acting appropriately and professionally. More transparency and accountability for how well they did after the election would increase their credibility -- or destroy it. It would be very interesting to see 538's "post-game" analysis of their house effect calculations since 2008. I'm guessing it probably wouldn't inspire confidence, though...
keith wrote: Sat Nov 05, 2022 7:06 pm A big part of 538 methodology is to allow for house bias based on historical results and methodology.

My point is that this hasn't changed, but that other structural changes are inducing errors that are not addressed (or addressable) by this methodology.

Excluding polls with no historical data points is not an ad hoc reason, it is self protecting their methodology.

So long as they are treating pollsters without prior history in the same manner they have always done, they are protecting their methodology. Given the details of what they are doing, you can argue about the quality of that methodology, but so long as they are sticking to the processes that they believe (in their professional opinion) are best practices, you can't question their integrity, only their competence/ability and structural factors that are not controlled for by their methodology (which is what I'm doing).

And I dont see that happening. As you said GIGO.

My personal belief is that they have a reasonable methodology and are applying it without bias, it's just that the assumptions underlying their entire process are no longer valid and therefore the result they produce are not reliable. In other words, I don't think 538 is doing it wrong (thus implying someone else could theoretically do it right), I think that what 538 (and others) are trying to do simply doesn't work anymore and no aggregation methodology or post-processing can fix it.
We need to take a close look at the dynamics of the polling leading up to the election and the results after Tuesday and ask some very hard questions about how we should use polls and how much credibility we should give pollsters (and aggregators) going forward. From the results we can determine what the best strategies would have been and from the polls we can determine the strategies they suggested would be best. If there is no correlation between these two sets of strategies then political polls are not useful beyond giving pundits (and candidates) a vehicle to drive narratives that generate page clicks and donations. If that is so, maybe it's time we start looking for something better.

yes/no?
User avatar
Kriselda Gray
Posts: 3125
Joined: Sun Nov 28, 2021 10:48 pm
Location: Asgard
Occupation: Aspiring Novelist
Verified:
Contact:

Is Political Polling Meaningful Anymore?

#16

Post by Kriselda Gray »

Slarti the White wrote: Mon Nov 07, 2022 1:44 am We need to take a close look at the dynamics of the polling leading up to the election and the results after Tuesday and ask some very hard questions about how we should If there is no correlation between these two sets of strategies then political polls are not useful beyond giving pundits (and candidates) a vehicle to drive narratives that generate page clicks and donations. If that is so, maybe it's time we start looking for something better.

yes/no?
Yes
User avatar
Foggy
Dick Tater
Posts: 9554
Joined: Mon Feb 22, 2021 8:45 am
Location: Fogbow HQ
Occupation: Dick Tater/Space Cadet
Verified: as seen on qvc zombie apocalypse

Is Political Polling Meaningful Anymore?

#17

Post by Foggy »

I don't think the problem with polling is methodology or statistical analysis.

I think the problem is, people are sick of polls and won't participate, especially younger people, who have other things to do.

Have you ever actually taken a telephone poll before? They're highly annoying. They take forever. They push an agenda, most of them.

And people are fucking with the pollsters, the ones who even play along. They're lying so the poll will be wrong.

You want an honest poll, waterboard the motherfuckers and then see how long they lie about who they're voting for! :boxing:
Out from under. :thumbsup:
User avatar
Liz
Posts: 96
Joined: Tue Feb 23, 2021 5:52 pm

Is Political Polling Meaningful Anymore?

#18

Post by Liz »

What Froggy said.. :yeahthat:
The American going right way or wrong way poll is at 72% wrong way, 28% right way. It's reported as a negative for Biden and Dem's.
I'm Democratic straight ticket but if I was asked I would answer wrong way.... because of what has become of the republicans.
So, it's a stupid question.
User avatar
Suranis
Posts: 5828
Joined: Mon Feb 22, 2021 5:25 pm

Is Political Polling Meaningful Anymore?

#19

Post by Suranis »

Another problem is that Polls rather than being regarded as a tool, have become a news story used by 24 hour news orgs desperate for new exciting content. So its polls with 2 hours of discussion again and again.

You may not know this, but here in Ireland after a string of newspapers publishing poll numbers as headlines election morning, everyone was completely sick of it ad people rightly saw newspapers publishing these polls on the day as influencing the election, intentional or not. So they basically banned coverage of the election the day before the election.
Hic sunt dracones
User avatar
Ben-Prime
Posts: 2599
Joined: Mon Feb 22, 2021 11:29 pm
Location: Worldwide Availability
Occupation: Managing People Who Manage Machines
Verified: ✅MamaSaysI'mBonaFide

Is Political Polling Meaningful Anymore?

#20

Post by Ben-Prime »

But the sunshine aye shall light the sky,
As round and round we run;
And the truth shall ever come uppermost,
And justice shall be done.

- Charles Mackay, "Eternal Justice"
User avatar
northland10
Posts: 5596
Joined: Mon Feb 22, 2021 6:47 pm
Location: Northeast Illinois
Occupation: Organist/Choir Director/Fundraising Data Analyst
Verified: ✅ I'm me.

Is Political Polling Meaningful Anymore?

#21

Post by northland10 »

Foggy wrote: Mon Nov 07, 2022 6:58 am I don't think the problem with polling is methodology or statistical analysis.

I think the problem is, people are sick of polls and won't participate, especially younger people, who have other things to do.

Have you ever actually taken a telephone poll before? They're highly annoying. They take forever. They push an agenda, most of them.
I was thinking the same right before I read your post. I don't even pick up a phone number I don't recognize. It has been a long time since I talked to a pollster, and all I remember was that it was unpleasant.

I avoid constant surveys as well, except the ones where the day jobs give us a personal day if we hit 95% participation. I can be bribed, or at least, don't want to be blamed for ruining it for everyone else.
101010 :towel:
User avatar
p0rtia
Posts: 4915
Joined: Mon Feb 22, 2021 9:55 am

Is Political Polling Meaningful Anymore?

#22

Post by p0rtia »

I responded to one poll* this cycle: Got a text naming the pollster and asking about my House race. I clicked and got to a simle age asking who I was planning to vote for in NY21, Castelli or Stefanik. I selected Stefanik.

A few second later I got another text, asking me if it would change my mind if I learned that Castelli had been dishonorably discharged from the CIA because of a sexual harassment case.

I informed the Castelli campaign. That is the last unsolicited poll* I will ever respond to. So:

1. We need to define poll* better. I know that push polls like the above are not what y'all are talking about when you talk about polls*, but that's all the more reason to define polls*.

2. Additional/new regulations/standards? Yeah, we fucking need that in spades.

3. Polls* have become the cocaine of MSM. They are a source of cheap, ignorant time-filling and self-fulfilling prophesy. As Katy Tur just said on that supposedly left-leaning TV channel "Tomorrow is the first election since the country has been hit with devastating inflation."

She (her infantile, corrupt producers) was comfy walking right past "...since the overturn of Roe v Wade." and "...since democracy was attacked and undermined by election deniers" and many others But no. All because the polls* told her that inflation was "the" issue of the campaign, because it was drawing 17 percent of the responses from respondents. With crime ranking second. No comment about how right wing media started pushing those two stories a couple of months ago with fear-mongering and lies.

Polls. We are not in a place where they are any use at all. They are now a tool of the right.
humblescribe
Posts: 1091
Joined: Mon Feb 22, 2021 3:42 pm
Occupation: Dude
Verified:

Is Political Polling Meaningful Anymore?

#23

Post by humblescribe »

A skeptical accountant is what I am:

Someone pays for these polls. Someone may be straight arrow, or someone may have an agenda.

In days of yore, polling was conducted at vetted locations and voters were interviewed after they cast their ballots. This information was then used to deduce the results.

Nowadays, as pointed out so eloquently above, polls are recurring. To my way of thinking, there is inherent bias in conducting a poll months before an election. How often does early polling influence later polling? ("Gee, it looks like Candidate X is trailing. Imma gonna say I'm voting for X, even though I was not going to vote or was still undecided.")

I've taken a few of these over the phone in the past when we had a land line. It seemed to me that these pollsters are sneaky. They start off by asking very generic questions about my life. Am I happy? Do I have a job? After these preliminary phatic questions, they slip in a question or two about a candidate or a philosophy. Then it is back to the softball questions about agreement (strongly, sorta, no opinion, disagreement). After four or five of those, it is back to another candidate.

They lull you into some sort of serenity and then hit you with the real purpose of their questions while you are still pondering your answers to previous red herring questions. Or maybe they aren't red herrings. Maybe they frame our state of mind and are used to promote or discount our answers to the other survey questions.
"Some cause happiness wherever they go; others whenever they go." O. Wilde
Slarti the White
Posts: 442
Joined: Mon Feb 22, 2021 1:23 pm
Location: Michigan
Verified: Badges... we don't need no stinkin' badges

Is Political Polling Meaningful Anymore?

#24

Post by Slarti the White »

Foggy wrote: Mon Nov 07, 2022 6:58 am I don't think the problem with polling is methodology or statistical analysis.

I don't think the problem is with methodology or statistical analysis either, I think the problem is with the assumptions on which the methodology and statistical analysis are based.

I think the problem is, people are sick of polls and won't participate, especially younger people, who have other things to do.

Yes. A "good" response rate is 1% -- when 99% of your sample opts out, it is nigh-impossible that significant selection bias isn't introduced.

Have you ever actually taken a telephone poll before? They're highly annoying. They take forever. They push an agenda, most of them.

No -- all of them have an agenda, even if it is only taking your time to collect data that they believe they can extract valuable information from.

And people are fucking with the pollsters, the ones who even play along. They're lying so the poll will be wrong.

Another issue which is basically impossible to control for.

You want an honest poll, waterboard the motherfuckers and then see how long they lie about who they're voting for! :boxing:

I believe that there is another way to get an honest poll: help a person collect data for their own use and buy it from them if they are willing to sell it. Besides, waterboarding gets everything all wet... and violates international law too. also.
Liz wrote: Mon Nov 07, 2022 10:03 am What Froggy said.. :yeahthat:
The American going right way or wrong way poll is at 72% wrong way, 28% right way. It's reported as a negative for Biden and Dem's.
I'm Democratic straight ticket but if I was asked I would answer wrong way.... because of what has become of the republicans.
So, it's a stupid question.

Another very serious problem with standard polling -- it has no way of capturing nuance or why someone is answering in the way they did. Approval rating polls suffer from this as well.
Suranis wrote: Mon Nov 07, 2022 10:59 am Another problem is that Polls rather than being regarded as a tool, have become a news story used by 24 hour news orgs desperate for new exciting content. So its polls with 2 hours of discussion again and again.

They are regarded as a tool by the media -- a tool to get page clicks and viewers. Unless there is some kind of accountability, or at least comparing what the polls said to the results afterwards then the value of polling as a constructive tool (to, for instance, guide where resources should go) is basically nil.

You may not know this, but here in Ireland after a string of newspapers publishing poll numbers as headlines election morning, everyone was completely sick of it ad people rightly saw newspapers publishing these polls on the day as influencing the election, intentional or not. So they basically banned coverage of the election the day before the election.

That might be a small step, but at least it's in the right direction. Personally, I'd like to ban all political polls in the media starting when the first candidate announces (looking at you Mr. Trump).
Ben-Prime wrote: Mon Nov 07, 2022 1:11 pm The Daily Beast has weighed in on this very question. Their answer: :shrug:

Good article, even if it is long on describing the problem and short on suggesting any solution. I found Nate Silver's comments interesting -- he recognizes the issues but believes that polling, though less accurate, is still useful. Which, of course, is highly motivated reasoning on his part -- if polling isn't useful any more he might as well close up shop at 538 and focus on sports statistics. I'd love to ask him how the evidence of the problems with polling effects his prior confidence regarding its usefulness. I think he needs a good Bayesian update...
northland10 wrote: Mon Nov 07, 2022 2:28 pm
Foggy wrote: Mon Nov 07, 2022 6:58 am I don't think the problem with polling is methodology or statistical analysis.

I think the problem is, people are sick of polls and won't participate, especially younger people, who have other things to do.

Have you ever actually taken a telephone poll before? They're highly annoying. They take forever. They push an agenda, most of them.
I was thinking the same right before I read your post. I don't even pick up a phone number I don't recognize. It has been a long time since I talked to a pollster, and all I remember was that it was unpleasant.

Because they are making you give up something of value to you (your time) so that they can get something of value to them (your opinion) for free. I would think that theft is rarely pleasant.

I avoid constant surveys as well, except the ones where the day jobs give us a personal day if we hit 95% participation. I can be bribed, or at least, don't want to be blamed for ruining it for everyone else.

In other words, so long as you are getting something that you value more than avoiding the unpleasantness, you are willing to participate. Any type of polling that ignores this principle is going to have problems.
p0rtia wrote: Mon Nov 07, 2022 3:17 pm I responded to one poll* this cycle: Got a text naming the pollster and asking about my House race. I clicked and got to a simle age asking who I was planning to vote for in NY21, Castelli or Stefanik. I selected Stefanik.

Um... I think you meant to say Castelli given the context and things you've posted elsewhere. Unless you were deliberately fucking with the poll.

A few second later I got another text, asking me if it would change my mind if I learned that Castelli had been dishonorably discharged from the CIA because of a sexual harassment case.

Wow. That is a new level of sliminess.

I informed the Castelli campaign.

Good for you :thumbsup: although there probably isn't anything they can do about it.

That is the last unsolicited poll* I will ever respond to.

And no pollster can control for that sort of selection bias.

So:

1. We need to define poll* better. I know that push polls like the above are not what y'all are talking about when you talk about polls*, but that's all the more reason to define polls*.

On the contrary, while your experience is extreme, I believe that even when pollsters have the best of intentions the context, order, and wording of their questions induces bias of some sort. In other words, I think all standard polls are push polls to one degree or another. And I completely agree that we need better and more nuanced definitions to describe polling.

2. Additional/new regulations/standards? Yeah, we fucking need that in spades.

Yes, although the question I have is if sufficient "guard rails" were introduced, would there be anything left of standard polling that was allowed?

3. Polls* have become the cocaine of MSM. They are a source of cheap, ignorant time-filling and self-fulfilling prophesy. As Katy Tur just said on that supposedly left-leaning TV channel "Tomorrow is the first election since the country has been hit with devastating inflation."

Not cocaine, crack. Polls are not just a drug, but a drug with an instant, powerful effect -- the easier to addict their viewers with.

She (her infantile, corrupt producers) was comfy walking right past "...since the overturn of Roe v Wade." and "...since democracy was attacked and undermined by election deniers" and many others But no. All because the polls* told her that inflation was "the" issue of the campaign, because it was drawing 17 percent of the responses from respondents. With crime ranking second. No comment about how right wing media started pushing those two stories a couple of months ago with fear-mongering and lies.

Yup.

Polls. We are not in a place where they are any use at all. They are now a tool of the right.

Which, in my mind, is a good reason to design a new tool that can't be co-opted by the authoritarian right in this manner.
humblescribe wrote: Mon Nov 07, 2022 3:47 pm A skeptical accountant is what I am:

Someone pays for these polls. Someone may be straight arrow, or someone may have an agenda.

Ever pollster has an agenda, even if it is only to make money by sampling public opinion as honestly as they are able.

In days of yore, polling was conducted at vetted locations and voters were interviewed after they cast their ballots. This information was then used to deduce the results.

Exit polling is still done and is a beast of a different color. There aren't as many problems with exit polling as there are with pre-election polling, but they don't include absentee voters (or early voters usually).

Nowadays, as pointed out so eloquently above, polls are recurring. To my way of thinking, there is inherent bias in conducting a poll months before an election. How often does early polling influence later polling? ("Gee, it looks like Candidate X is trailing. Imma gonna say I'm voting for X, even though I was not going to vote or was still undecided.")

Very good questions that pollsters would probably prefer you didn't ask.

I've taken a few of these over the phone in the past when we had a land line. It seemed to me that these pollsters are sneaky. They start off by asking very generic questions about my life. Am I happy? Do I have a job? After these preliminary phatic questions, they slip in a question or two about a candidate or a philosophy. Then it is back to the softball questions about agreement (strongly, sorta, no opinion, disagreement). After four or five of those, it is back to another candidate.

Some of that is to try to determine whether or not you will actually vote, but regardless of their intentions, they are biasing their results in some manner.

They lull you into some sort of serenity and then hit you with the real purpose of their questions while you are still pondering your answers to previous red herring questions. Or maybe they aren't red herrings. Maybe they frame our state of mind and are used to promote or discount our answers to the other survey questions.

Depends on how much integrity they have -- no way for you to tell. In any case, it doesn't provide you with more incentive to spend your time answering polls.
It seems like we're all pretty much on the same page as to the question of is political polling meaningful anymore, which begs the question of what would a viable alternative look like? I have an idea about that -- I call it "pull polling" because the person being polled gets to pull the poll towards whatever they care about rather than being pushed by the pollster and their agenda. I'll start a new thread about it after the election.

Everyone get out and vote!
Danraft
Posts: 492
Joined: Mon Feb 22, 2021 1:43 pm
Location: Michigan
Contact:

Is Political Polling Meaningful Anymore?

#25

Post by Danraft »

Ok, the skinny on polling from an evolutionary view is that we actually don’t know how we are going to vote or why (in most cases).

There are many motivations. Whichever motivation ends up holding the reins of action at that point is the one getting the power which is explained by rationalization or narrative.

The limbic system, the more primitive segments, causes substantial changes just based on weather, smells, and other mood influencing environmental factors. That’s just a start.

We shall see what comes.
Post Reply

Return to “Current Politics”