The media and the DNC will not allow Biden to lose on ED. This will get stretched out for weeks, and we will be seeing massive protests/riots/ballet harvesting.
Big winners? The MSM(think of the advertising dollars at stake here) and people who like to set fire to public buildings in cities.
The media and the DNC will not allow Biden to lose on ED. This will get stretched out for weeks, and we will be seeing massive protests/riots/ballet harvesting.
Big winners? The MSM(think of the advertising dollars at stake here) and people who like to set fire to public buildings in cities.
I am going to wait until all the ballot gatherers are finished finding the box of ballots that they left in the drivers seat of their car to decide what to believe
The media and the DNC will not allow Biden to lose on ED.
That’s their final strategy. Right now the idea is to suppress the GOP Election Day vote by making it seem an overwhelming Biden advantage when the opposite is true. After that fails, they go to this. If/when places like NH and MN flip, they can’t play this game anymore either.
The Tug is quick to dismiss the polls, and given their 2016 performance, that's not an unfair reflex. But the fact remains that the stats nerds have been working tirelessly to get their guys in improve the methodology that led to 2016's massive polling failure. Namely, the improper weighting given to non-college whites, who were the biggest factor in driving that error.
All that's to say that if Trump does win, it's not going to be because of the same polling mistakes that were made four years ago. It's going to be because of an unforeseen miss in the polls that is new to this election. A "silent majority" that was silent for a reason that the analytics guys are still missing.
All the stats in the world can't make up for the fact no one answers their fucking phone in 2020
The Tug is quick to dismiss the polls, and given their 2016 performance, that's not an unfair reflex. But the fact remains that the stats nerds have been working tirelessly to get their guys in improve the methodology that led to 2016's massive polling failure. Namely, the improper weighting given to non-college whites, who were the biggest factor in driving that error.
All that's to say that if Trump does win, it's not going to be because of the same polling mistakes that were made four years ago. It's going to be because of an unforeseen miss in the polls that is new to this election. A "silent majority" that was silent for a reason that the analytics guys are still missing.
I actually disagree with this big time
I think a lot of the pollsters view 2016 as a bit of a mirage and while you're right that they've attempted to change their methodology they are falling into a trap that anybody that works with data knows is the biggest issue of all ... terrible data quality
In the current environment of "cancel culture" and an inability to agree to disagree, there's very little incentive out there for Trump voters to tell anybody that they are a Trump voter. Yes, some will fly flags and be in your face about it. But most will subscribe to anonymity. So unless you materially change how you go about identifying a "likely Trump voter" versus a "likey Biden voter," you'll continue to miss the mark.
I suspect that what a lot of pollsters are doing is trying to quantify the "error factor" that they have for underestimating Trump and applying that to their numbers. In the end, it's a flat out guess.
I feel like old school polling is gradually becoming an outdated method and there are likely far better leading indicators out there.
The Tug is quick to dismiss the polls, and given their 2016 performance, that's not an unfair reflex. But the fact remains that the stats nerds have been working tirelessly to get their guys in improve the methodology that led to 2016's massive polling failure. Namely, the improper weighting given to non-college whites, who were the biggest factor in driving that error.
All that's to say that if Trump does win, it's not going to be because of the same polling mistakes that were made four years ago. It's going to be because of an unforeseen miss in the polls that is new to this election. A "silent majority" that was silent for a reason that the analytics guys are still missing.
I think that is a generous assumption. You dont get graphs like this unless there is something fucky going on with your pole...
They aren't (publicly) changing their methodology mid election season and the electorate isn't wildly swinging its opinion in 3-5 days. The only way you get this type of data generated is if you are trying to create a certain perception and still get close to the correct answer at the only time accuracy matters (Nov. 3rd).
It allows polestars to say dumb shit like "Texas is in play" or "Biden up 17 points over Trump in Wisconsin" for months up to the election (hype up the wave!) and then still get the poles into the margin of error on election night and claim the polls were "technically correct" on election night.
Bingo
It's also leading to create a false narrative for Democratic leadership in thinking that they are safer than they actually are as well as allowing them to focus resources into areas that they think they have a shot but don't have a shot.
Any conversations that I've heard about Texas going blue I just laugh at ... I'd be SHOCKED if Biden won it in this election.
The Tug is quick to dismiss the polls, and given their 2016 performance, that's not an unfair reflex. But the fact remains that the stats nerds have been working tirelessly to get their guys in improve the methodology that led to 2016's massive polling failure. Namely, the improper weighting given to non-college whites, who were the biggest factor in driving that error.
All that's to say that if Trump does win, it's not going to be because of the same polling mistakes that were made four years ago. It's going to be because of an unforeseen miss in the polls that is new to this election. A "silent majority" that was silent for a reason that the analytics guys are still missing.
I actually disagree with this big time
I think a lot of the pollsters view 2016 as a bit of a mirage and while you're right that they've attempted to change their methodology they are falling into a trap that anybody that works with data knows is the biggest issue of all ... terrible data quality
In the current environment of "cancel culture" and an inability to agree to disagree, there's very little incentive out there for Trump voters to tell anybody that they are a Trump voter. Yes, some will fly flags and be in your face about it. But most will subscribe to anonymity. So unless you materially change how you go about identifying a "likely Trump voter" versus a "likey Biden voter," you'll continue to miss the mark.
I suspect that what a lot of pollsters are doing is trying to quantify the "error factor" that they have for underestimating Trump and applying that to their numbers. In the end, it's a flat out guess.
I feel like old school polling is gradually becoming an outdated method and there are likely far better leading indicators out there.
The Tug is quick to dismiss the polls, and given their 2016 performance, that's not an unfair reflex. But the fact remains that the stats nerds have been working tirelessly to get their guys in improve the methodology that led to 2016's massive polling failure. Namely, the improper weighting given to non-college whites, who were the biggest factor in driving that error.
All that's to say that if Trump does win, it's not going to be because of the same polling mistakes that were made four years ago. It's going to be because of an unforeseen miss in the polls that is new to this election. A "silent majority" that was silent for a reason that the analytics guys are still missing.
It's pretty fucking simple. Phone rings - number you don't recognize but answer anyway. They ask who you're voting for.
Given the bullshit that Antifa and BLM has caused this summer, are you even going to give them any HINT that you might be voting for Trump? If you have my phone number and how I'm voting - you pretty well have all you need to show up on my doorstep, ruin my business, destroy my property.
Fuck, no I"m not answering the question. I'm probably not even answering the phone. So you're naturally going to get more Biden people picking up and blabbing how they plan to vote.
To the extent that this dynamic is true, it's exactly what the pollsters are trying to build in to their sample weighting. That and 1000 others factors. Naturally, you have to LIPO for the accuracy of that sample weighting.
Can we? all admit the polling was fucking garbage solely meant to create a narrative.
Even Fat Frank Luntz was apologizing today for it.
The Tug is quick to dismiss the polls, and given their 2016 performance, that's not an unfair reflex. But the fact remains that the stats nerds have been working tirelessly to get their guys in improve the methodology that led to 2016's massive polling failure. Namely, the improper weighting given to non-college whites, who were the biggest factor in driving that error.
All that's to say that if Trump does win, it's not going to be because of the same polling mistakes that were made four years ago. It's going to be because of an unforeseen miss in the polls that is new to this election. A "silent majority" that was silent for a reason that the analytics guys are still missing.
It's pretty fucking simple. Phone rings - number you don't recognize but answer anyway. They ask who you're voting for.
Given the bullshit that Antifa and BLM has caused this summer, are you even going to give them any HINT that you might be voting for Trump? If you have my phone number and how I'm voting - you pretty well have all you need to show up on my doorstep, ruin my business, destroy my property.
Fuck, no I"m not answering the question. I'm probably not even answering the phone. So you're naturally going to get more Biden people picking up and blabbing how they plan to vote.
To the extent that this dynamic is true, it's exactly what the pollsters are trying to build in to their sample weighting. That and 1000 others factors. Naturally, you have to LIPO for the accuracy of that sample weighting.
Can we? all admit the polling was fucking garbage solely meant to create a narrative.
Even Fat Frank Luntz was apologizing today for it.
Create a narrative and influence campaign donations. How many races that ended up being won by a the GOP or were way closer than the polling were impacted by the polling that said they were down double digits? If you think a candidate is really down 10 points are you going to be so quick to open up the check book?
Comments
Big winners? The MSM(think of the advertising dollars at stake here) and people who like to set fire to public buildings in cities.
#molotov
I think a lot of the pollsters view 2016 as a bit of a mirage and while you're right that they've attempted to change their methodology they are falling into a trap that anybody that works with data knows is the biggest issue of all ... terrible data quality
In the current environment of "cancel culture" and an inability to agree to disagree, there's very little incentive out there for Trump voters to tell anybody that they are a Trump voter. Yes, some will fly flags and be in your face about it. But most will subscribe to anonymity. So unless you materially change how you go about identifying a "likely Trump voter" versus a "likey Biden voter," you'll continue to miss the mark.
I suspect that what a lot of pollsters are doing is trying to quantify the "error factor" that they have for underestimating Trump and applying that to their numbers. In the end, it's a flat out guess.
I feel like old school polling is gradually becoming an outdated method and there are likely far better leading indicators out there.
It's also leading to create a false narrative for Democratic leadership in thinking that they are safer than they actually are as well as allowing them to focus resources into areas that they think they have a shot but don't have a shot.
Any conversations that I've heard about Texas going blue I just laugh at ... I'd be SHOCKED if Biden won it in this election.
Even Fat Frank Luntz was apologizing today for it.