How did the pollsters get it wrong AGAIN?

[hal2kilo]

Those are the outliers. Most Trump supporters I know tell me they voted for him in secret.

Who knew that non cucks could be such snowflakes.

 

[dank69]

Bingo. Get back to equality for all. In general, human beings understand and love this concept, regardless of race or gender. The current democratic party appears to be a twisted mess of what it once was.

Thankfully the Republican party is still grounded.

 

[UNCjigga]

In the B2B marketing world, you've had a steady shift over the past decade from email marketing and generic display advertising to more organic channels like content, PPC/SEO and social media marketing. The basic idea is that B2B sales cycles are longer, and you want to nurture buyers/decisionmakers along their journey and measure at each step to inform your next engagement.

I think polling needs to adjust to the 21st century and adopt some sort of a content engagement model. Responses to a static poll question at a moment in time are now less relevant than how often a meme or an ad gets shared on Facebook/Twitter or mentioned online in blogs/news. I don't have the answer on how to do this--maybe independent pollsters will need to surreptitiously "seed" different memes, thought pieces etc. based on candidates or issues and then measure engagement scores. To be honest, that seems to be what Russia's Internet Research Agency was doing (was that engagement data shared with the Trump campaign?) Might be something campaigns would be better at managing themselves rather than relying on independent pollsters, but they'd need to develop content for both sides and measure how both performs.

Not sure if I'm making sense or if my point is getting across (how many marketing professionals here?)

 

[Zorba]

Actually, as states flip as urban areas and mail-in ballots are counted, it looks like the polls weren't THAT far off given their margin of error.

This thread brought to you by Red Mirage!

You shouldn't always be on the same extreme of your uncertainty bands.

 

[Amused]

You shouldn't always be on the same extreme of your uncertainty bands.

Huh?

 

[Zorba]

I think the biggest problem in polling is trying to measure who is a "likely voter". Our country seems so evenly split that it kinda comes down to who has the more motivated voting base.

There is also something to be said about being exited enough about your candidate to evangelize them. If Bernie was the D candidate, I know lots of people who would be absolutely driven to not only vote for him but to bring along as many people as possible. For BIden, these same people did vote for him, but weren't super exited about him. They were more scared/angry/against Trump.

There was a massive volunteer effort for phone and text banking. Sanders would've been killed in FL.

 

[Zorba]

Huh?

If you are always on the same extreme end of your uncertainty band, you have a bias in the measurement.

 

[Staples]

These polls are asked of random people who may or may not actually go through the effort of voting. I am glad they got a bigger turnout than usual but still tons of people didn't vote (probably many of whom were surveyed).

On another thought. I think there were better candidates than Biden. He is an old white guy who has half the energy as Trump. He does not appeal to anyone really. With that said, he is fine and perfectly competent to be president but that does not fix the image situation.

 

[Zorba]

These polls are asked of random people who may or may not actually go through the effort of voting. I am glad they got a bigger turnout than usual but still tons of people didn't vote (probably many of whom were surveyed).

This probably the biggest issue. I have a feeling they had a deflated likely voter percentage for Trump.

Miami-Dade was just way of though.

 

[Chryso]

Apparently they are scared of rioting and looting in the cities that they don’t live in/eye roll


Actually we know what “reprisals” they really don’t want...it’s to have their opinions questioned and countered. They want to speak their mind without people telling/showing them how fucking stupid they sound.

You demonstrate the reprisals nicely.

 

[brycejones]

You demonstrate the reprisals nicely.

Do pollsters do any of those things?

 

[Lezunto]

All the votes must be counted. No excuses to refuse to do this.

No excuses.

 

[Pipeline 1010]

There was a massive volunteer effort for phone and text banking. Sanders would've been killed in FL.

My point was that I think the polls might have been more accurate had there been a Sanders vs Trump election. I wasn't stating that he would do any better or worse than Biden. My premise/theory is that the polls were inaccurate in part due to the difficulty in determining what a "likely voter" actually is.

 

[darkswordsman17]

Honestly, why do people care so much what the polls show? I've never understood this. I get why the number people do, but the ones using it for making broad statements is just stupid and makes you look dumb for doing it.

This probably the biggest issue. I have a feeling they had a deflated likely voter percentage for Trump.

Miami-Dade was just way of though.

I think the issue is that there's still a lot of variables that need to be accounted for. This modern way of modeling (that 538 has been doing for instance) is a work in progress and there's clearly things it doesn't account for properly. For instance, when they target certain groups with disinformation on social media, that's not always widely known by polling organizations. From what I gathered there was quite a lot of that happening specifically towards Latinos in Southern Florida just here in like the last month. And while it was reported on, I have a hunch that type of stuff is difficult to quantize in current models. I believe 538 mentioned it several times as being a likely cause of where their model struggled.

 

[darkswordsman17]

These polls are asked of random people who may or may not actually go through the effort of voting. I am glad they got a bigger turnout than usual but still tons of people didn't vote (probably many of whom were surveyed).

On another thought. I think there were better candidates than Biden. He is an old white guy who has half the energy as Trump. He does not appeal to anyone really. With that said, he is fine and perfectly competent to be president but that does not fix the image situation.

The image problem is that that is the problem. People want fucking clowns, not politicians, and then act surprised when they get pies in the face.

 

[dank69]

My point was that I think the polls might have been more accurate had there been a Sanders vs Trump election. I wasn't stating that he would do any better or worse than Biden. My premise/theory is that the polls were inaccurate in part due to the difficulty in determining what a "likely voter" actually is.

More likely that voter suppression is mostly to blame for the consistent discrepancy in Trump's favor. It's one thing to be super gung-ho about voting and another to see a multiple hour line when you go to do it.

 

[K1052]

Reading political twitter today and I believe I have found the one thing that both Ds and Rs unanimously agree on: They're angry at their pollsters.

Ds pissed about phantom leads and Rs pissed about spending huge sums in races that aren't close. They also both spend lots of money for these products.

 

[TheVrolok]

I'm going to enjoy this Trump win more so than the last. I can't wait to see the riots from the snowflakes and the celebrities crying on TV. Its going to be fucking glorious.

Yikes .. it looks like this is already not aging well..

 

[UNCjigga]

Reading political twitter today and I believe I have found the one thing that both Ds and Rs unanimously agree on: They're angry at their pollsters.

Ds pissed about phantom leads and Rs pissed about spending huge sums in races that aren't close. They also both spend lots of money for these products.

What’s weird is I thought Mitch McConnell would be gloating today about his 20 point win (and Graham, Ernst, Collins etc.) but even he seems a bit measured—I think even he is seeing more cracks in the foundation they’ve built for themselves in the South since the Civil Rights Act. Cobb in GA flipping blue is a massive warning sign and he seems to get that if they lose those suburbs and women the party is finished.

 

[Sunburn74]

Well I mean they are only one or two cycles away from losing texas and they may lose georgia already.

AZ used to be solidly republican. Now its democratic. The reality is, there is no state other than maybe PA where you can confidently say the GOP is getting stronger.

I will however say this election has showed me that the rumors of the death of the GOP is way overblown. They aren't going anywhere anytime soon, particularly in the senate.

 

[vi edit]

I think AZ is the new IA. Will be interesting to see what 2024 holds there. I wouldn't bet on it being blue reliably.

 

[Ken g6]

I think AZ is the new IA. Will be interesting to see what 2024 holds there. I wouldn't bet on it being blue reliably.

It really depends on whether the Democrat population is growing faster than the retirement population. If it is, AZ could be the new CO instead.

 

[hal2kilo]

All the votes must be counted. No excuses to refuse to do this.

No excuses.

Depends on which state has Biden ahead of Trump

What’s weird is I thought Mitch McConnell would be gloating today about his 20 point win (and Graham, Ernst, Collins etc.) but even he seems a bit measured—I think even he is seeing more cracks in the foundation they’ve built for themselves in the South since the Civil Rights Act. Cobb in GA flipping blue is a massive warning sign and he seems to get that if they lose those suburbs and women the party is finished.

Newt's old stronghold.

 

[woolfe9998]

Silver seems to think Biden will win the PV by 5% when all votes are eventually counted. Which would still be a 3.5% polling error on the PV. While the swing state error ranges from 1 to about 7.

The stupid thing is people will not hate on the polls as much this time if Biden wins, even though they have been less accurate than the 2016 polling.

 

[Torn Mind]

From my experience is just learning textbook statistics, people still can't grasp that sampling is not a perfect process and it is intellectually bankrupt and indicative of a lack of understanding to assume "the sample was sufficiently representative' in the actual sense, and not merely the mathematically correct sense(where n is greater than some ratio involving measures involving variance and other numbers).

plus the pollster oftentimes have incentives to influence......although going pro- one side might actually backfire because it makes the public complacent seeing someone having a lead.