Pollsters staying away from Georgia races

[LateToCult 41,830]

After a disastrous November election for the polling industry, when the polls again underestimated President Donald Trump (who lost regardless) as well as GOP candidates down the ballot, pollsters are mostly sidelined in the run-up to the Jan. 5 Georgia elections, which most observers regard as toss-ups.

The public polls that drove so much of the news coverage ahead of November — and generated tremendous distrust afterward — have all but disappeared in Georgia, and they are set to stay that way: Some of the most prolific, best-regarded media and academic pollsters told POLITICO they have no plans to conduct pre-election surveys in Georgia.

And while the campaigns and outside groups are still using survey data to make critical decisions, in many cases, polling has taken a back seat. Strategists are leaning more heavily on absentee and early voting stats, along with the detailed results of the election held just last month.

...

Part of the reason public pollsters are staying away from Georgia is the awkward timing of the races. With the elections being held on Jan. 5, the final two weeks of the race are coinciding with the Christmas and New Year’s holidays — typically a time when pollsters refrain from calling Americans on the phone. The voters who would answer a telephone poll or participate in an internet survey over the holidays might be meaningfully different from those who wouldn’t, which would skew the results.

https://www.google.com/amp/s/www.politico.com/amp/news/2020/12/21/pollsters-disappear-georgia-senate-448854

[BUtterfield 8 41,816]

So do we have a chance or what 

[MJHolland 12,725]

Well yeah they would likely stay away from polling for this kind of election anyway because it's very very tricky to do it accurately 

Also... will forever cringe at the media blaming the polling industry for "disastrous" inaccuracies in 2020 (or even 2016). Polls, on average, were just about as right in 2020 and 2016 as they've been since the 1970s. 

It's the media's/public's fault for not interpreting them correctly — i.e. taking them as fact, obsessing over incremental changes from week-to-week and even day-to-day, etc. 

[LateToCult 41,830]

5 minutes ago, BUtterfield 8 said:

So do we have a chance or what 

Nobody really knows. 

 

[LateToCult 41,830]

4 minutes ago, MJHolland said:

Well yeah they would likely stay away from polling for this kind of election anyway because it's very very tricky to do it accurately 

Also... will forever cringe at the media blaming the polling industry for "disastrous" inaccuracies in 2020 (or even 2016). Polls, on average, were just about as right in 2020 and 2016 as they've been since the 1970s. 

It's the media's/public's fault for not interpreting them correctly — i.e. taking them as fact, obsessing over incremental changes from week-to-week and even day-to-day, etc. 

I would put more of the blame on people like Nate Silver who took all this polling data and made it seem like the inevitable outcome was a huge win for the Dems in every branch. They presented it in such a way that regular people took it as fact too.

The pollsters also screwed up. People were straight up lying to them and bragging about it but pollsters dug their head in the sand and said it wasn’t true. 

[MJHolland 12,725]

41 minutes ago, LateToCult said:

I would put more of the blame on people like Nate Silver who took all this polling data and made it seem like the inevitable outcome was a huge win for the Dems in every branch. They presented it in such a way that regular people took it as fact too.

The pollsters also screwed up. People were straight up lying to them and bragging about it but pollsters dug their head in the sand and said it wasn’t true. 

I cannot agree. 

I think Silver and 538 are actually a beacon of reason when it comes to the industry. One of the biggest changes 538 made after 2016 was the never talk about polling results like "Biden 7 points ahead," because they realized normal people took that as "race over, Biden will win." Instead, ever since 2017, they've remained committed to probabilistic language. 

Since 2017, 538 has always communicated polling averages/forecasts in probabilities. I.e. "Biden has an 8-in-10 chance of winning, and Trump has a 2-in-10 chance of winning." 

If you listen to their podcast, they discuss week after week how they can do a better job at making sure the media/normal people have the tools/mindset to interpret polling and forecasts correctly.

I've also seen no evidence or reports of large #s of people lying to pollsters.  

[LateToCult 41,830]

55 minutes ago, MJHolland said:

I cannot agree. 

I think Silver and 538 are actually a beacon of reason when it comes to the industry. One of the biggest changes 538 made after 2016 was the never talk about polling results like "Biden 7 points ahead," because they realized normal people took that as "race over, Biden will win." Instead, ever since 2017, they've remained committed to probabilistic language. 

Since 2017, 538 has always communicated polling averages/forecasts in probabilities. I.e. "Biden has an 8-in-10 chance of winning, and Trump has a 2-in-10 chance of winning." 

If you listen to their podcast, they discuss week after week how they can do a better job at making sure the media/normal people have the tools/mindset to interpret polling and forecasts correctly.

I've also seen no evidence or reports of large #s of people lying to pollsters.  

(I’m so sorry this is so long.  ) 

 

I would have to disagree. In my opinion, Nate could be reasonable but I don’t think this was reflected in 538 as well as it should have. His online meetings with staff and the written pieces on 538 were grounded in reality and really focused on the numbers but the graphics on their site weren’t doing a good enough job at reflecting the data and I don’t think the podcast talks really did anything when it came to tackling issues with the site. As a result you had everyday people who maybe weren’t too knowledgeable on probability seeing all these simple and sleek visuals that looked like they overwhelmingly placed dems as the winners. I think this was a little irresponsible on 538’s part.
 

They used their visuals to generate traffic to their site but didn’t address how effective those visuals were at delivering the correct data. Were people leaving the site understanding the probabilities tied with what they saw? No. They were leaving with an expectation of how the election would turn out and they trusted that what they saw was a play by play of what would happen because that’s how it was presented to them. I think Nate knew this. In the hours before the election it honestly looked like he was trying to cover his ass on social media by making distinctions between the probabilities and what people had come to expect because of how they interpreted what 538 was saying. 
 

I mean, sure you can make the argument that it’s the general public’s fault for not knowing what they were looking at but when you have this huge platform thats job is to share data in a way that’s digestible to the public, then for me it’s hard not to put most of the blame on 538. 
 

Oof, now to the second part. 
 

Fair enough, I didn’t present anything to support the theory that people definitely lied. But, to me at least, this seems like it could be a real reason why some polls were so wrong. There was a red wave that was completely missed by pollsters. Something obviously went very wrong with how they collected data from the public. This makes it hard for me to believe the certainty that pollsters have that lying wasn’t a thing. At the very least it has to be something they treat seriously moving forward.

[MJHolland 12,725]

3 minutes ago, LateToCult said:

(I’m so sorry this is so long.  ) 

 

I would have to disagree. In my opinion, Nate could be reasonable but I don’t think this was reflected in 538 as well as it should have. His online meetings with staff and the written pieces on 538 were grounded in reality and really focused on the numbers but the graphics on their site weren’t doing a good enough job at reflecting the data and I don’t think the podcast talks really did anything when it came to tackling issues with the site. As a result you had everyday people who maybe weren’t too knowledgeable on probability seeing all these simple and sleek visuals that looked like they overwhelmingly placed dems as the winners. I think this was a little irresponsible on 538’s part.
 

They used their visuals to generate traffic to their site but didn’t address how effective those visuals were at delivering the correct data. Were people leaving the site understanding the probabilities tied with what they saw? No. They were leaving with an expectation of how the election would turn out and they trusted that what they saw was a play by play of what would happen because that’s how it was presented to them. I think Nate knew this. In the hours before the election it honestly looked like he was trying to cover his ass on social media by making distinctions between the probabilities and what people had come to expect because of how they interpreted what 538 was saying. 
 

I mean, sure you can make the argument that it’s the general public’s fault for not knowing what they were looking at but when you have this huge platform thats job is to share data in a way that’s digestible to the public, then for me it’s hard not to put most of the blame on 538. 
 

Oof, now to the second part. 
 

Fair enough, I didn’t present anything to support the theory that people definitely lied. But, to me at least, this seems like it could be a real reason why some polls were so wrong. There was a red wave that was completely missed by pollsters. Something obviously went very wrong with how they collected data from the public. This makes it hard for me to believe the certainty that pollsters have that lying wasn’t a thing. At the very least it has to be something they treat seriously moving forward.

Can you share one of their data visualizations that you think was irresponsible? The one I linked in my previous post was the forecast's home page — and as you can see it said Trump still had a 20% (or 1-in-5 chance) (until it updated to 10% post-election) chance of winning, so it's definitely a far cry from "Biden has this in the bag," like you alluded. 

If you think it's bad that pollsters (which 538 is not) paint their reports as "X candidate is definitely going to win," but you also think probabilities (which is what polls actually like really truly represent) don't cut it either, I'm curious as to what you want them to do? 

I guess I'm not a good sample for the general public because I went to college and studied applied political/economic statistic methods (polling), but their data visualization game, to me, is very very clear. Like best in the industry in my opinion. I applied to be a data scientist with them but never got an interview :') 

Also... Nate had been saying since mid-Summer when their forecast went live that Trump could very reasonably win. Like, he said every podcast that he thought even his forecast was too bullish on Biden. And he said almost everyday that he expected the race to dramatically tighten in the last week/few days. 

[LateToCult 41,830]

1 hour ago, MJHolland said:

Can you share one of their data visualizations that you think was irresponsible? The one I linked in my previous post was the forecast's home page — and as you can see it said Trump still had a 20% (or 1-in-5 chance) (until it updated to 10% post-election) chance of winning, so it's definitely a far cry from "Biden has this in the bag," like you alluded. 

If you think it's bad that pollsters (which 538 is not) paint their reports as "X candidate is definitely going to win," but you also think probabilities (which is what polls actually like really truly represent) don't cut it either, I'm curious as to what you want them to do? 

I guess I'm not a good sample for the general public because I went to college and studied applied political/economic statistic methods (polling), but their data visualization game, to me, is very very clear. Like best in the industry in my opinion. I applied to be a data scientist with them but never got an interview :') 

Also... Nate had been saying since mid-Summer when their forecast went live that Trump could very reasonably win. Like, he said every podcast that he thought even his forecast was too bullish on Biden. And he said almost everyday that he expected the race to dramatically tighten in the last week/few days. 

Sure, the first one that comes to mind is their “path to victory” (or at least that’s what I’ll call it). It’s like a literal path with segments of different sizes representing the states and the number of their electoral votes. I think you know which one I’m talking about. 
 

And yeah, I know 538 isn’t a pollster. The thing with pollsters that I mentioned was separate. And I don’t think I said that probabilities don’t cut it. I’m saying I don’t think 538 does a good enough job at translating probability to a visual because it’s easy to take that visual as a definitive statement. Frankly, it probably can’t be improved much so maybe they should be asking themselves if things like that victory path should be included at all. They’re great to look at and I’m sure are very click baity but I think they’re part of the reason why 538 has gotten trashed so hard after general electionscompared to a site like Real Clear, which basically just gives you a table and a graph as far as visuals go.