polling stuff

[PoliSciPulse]

What I want to know is what happens if you remove the Trafalgar Group polls, which seem to go way right . I remember once one group had like +1 Harris and Trafalgar had like +6 Trump on the same day (I'm probably exaggerating it a bit, but the point is, Trafalgar Group stands out). Don't believe I've seen a single Trafalgar Group poll that ever had Harris up, and this includes right after the Convention, when Harris got her bounce.

 

[PoliSciPulse]

Yes, it's an awesome comment by dobieg, and I thank you for your kind words in turn!

After the election, I know you'll be reaching-out to the other party in the event of a Trump win, and rest assured I'll be doing my best should Harris win.

I can't tell you how disgusted I am that as Americans, we are at each-other's throats.

Remember,

"Democracy is a messy thing!"

This is an awesome quote and sums up how I try to approach election cycles on the whole. Politics should not be tearing families and stuff apart. And DP is one big family

 

[maxparrish]

A friend of mine, in 2015, called him "Trumpolini".

 

[maxparrish]

I'm a believer in "reversion to the mean". Essentially, that races vacillate to-and-fro over time.

An interesting application of that principle. However, the probability of a reversion to the actual voting mean (as opposed to polling mean) for each side isn't 50-50. Unknown is actual turnout to vote, which in my view will determine this outcome.

And like you, my (and Nate Silver's gut) is that Trump will snatch a victory.

MSN

Now the question is, if Trump wins, what will the Judge in Trumps NY conviction do? Will he have the hutzpah to give an actual jail sentence to an incoming President, to be served when?

 

[maxparrish]

I think if it looks like a Trump landslide from early vote numbers that will prevent the cheaters from trying as hard as they did in 2020 because there will be consequences this time with a likely Trump victory!

No matter who wins there will be lawsuits, accusations of cheating, controversy over voting machines, etc. The only difference this time is that if Trump loses, the capital will be surrounded with water cannons, machine guns, tanks, and armed military as a warning to the nutjobs, attempt to breach the steps and you will likely die.

Sadly, in our country, this will become SOP for all elections.

 

[Chomsky]

Hear, hear! Unfortunately I like some many others have the ability to get very ugly and vile. I don't want to be and I don't like it but it's one of my weaknesses.

Sometimes some of the things we say on here is as embarrassing in hindsight as the day after a bender in our younger days. You just close your eyes real tight and cringe. It's awful and you feel terrible but an hour later you're doing the same damn thing.

Don't hesitate to put a post aside for a few minutes, part of the day, or even overnight, before finally hitting "Post Reply". That saved my arse, often!

Also, never put anything in print you don't want the whole world to see. Assume everyone in the world, your wife & kids, friends, employer, Priest, neighbors, everyone - can see it - because someday they just may!

 

[Chomsky]

This is an awesome quote and sums up how I try to approach election cycles on the whole. Politics should not be tearing families and stuff apart. And DP is one big family

Thank you!

 

[Chomsky]

An interesting application of that principle. However, the probability of a reversion to the actual voting mean (as opposed to polling mean) for each side isn't 50-50. Unknown is actual turnout to vote, which in my view will determine this outcome.

And like you, my (and Nate Silver's gut) is that Trump will snatch a victory.

MSN

Now the question is, if Trump wins, what will the Judge in Trumps NY conviction do? Will he have the hutzpah to give an actual jail sentence to an incoming President, to be served when?

Thanks for your insight and opinion. And yes - by reversion I was speaking to the polling mean.

Yes. Turnout is the wildcard. It pretty much takes the "science" out of it!

As to the NY case, I think the judge is obligated to sentence - unless the SCOTUS decision precludes it. I would hope the judge sentences without political consideration, letting the cards fall where they may.

I'm also wondering about the GA case.

 

[maxparrish]

Thanks for your insight and opinion. And yes - by reversion I was speaking to the polling mean.

Yes. Turnout is the wildcard. It pretty much takes the "science" out of it!

As to the NY case, I think the judge is obligated to sentence - unless the SCOTUS decision precludes it. I would hope the judge sentences without political consideration, letting the cards fall where they may.

I'm also wondering about the GA case.

On one hand we have this shocking news:

Republicans' early voting lead in Nevada "unheard of"

There is "no good news" for Kamala Harris in early voting numbers from Nevada, according to newly released data.

www.newsweek.com

Voters Across America Line Up To Cast Early Ballots
By James Bickerton
US News Reporter

Nearly 20,000 more registered Republicans have voted in Nevada thus far in the presidential election than registered Democrats, according to data released on Wednesday, which one local politics expert described as "unheard of at this point in any other presidential cycle."

On the other hand we have this:

Kamala Harris as popular among Democrats as Obama in 2008: poll

A new survey has found a 72 percent approval rating for Kamala Harris among Democratic voters — one of the highest of any nominee going back 70 years.

www.newsweek.com

 

[Chomsky]

On one hand we have this shocking news:

Republicans' early voting lead in Nevada "unheard of"

There is "no good news" for Kamala Harris in early voting numbers from Nevada, according to newly released data.

www.newsweek.com

Voters Across America Line Up To Cast Early Ballots
By James Bickerton
US News Reporter

Nearly 20,000 more registered Republicans have voted in Nevada thus far in the presidential election than registered Democrats, according to data released on Wednesday, which one local politics expert described as "unheard of at this point in any other presidential cycle."

On the other hand we have this:

Kamala Harris as popular among Democrats as Obama in 2008: poll

A new survey has found a 72 percent approval rating for Kamala Harris among Democratic voters — one of the highest of any nominee going back 70 years.

www.newsweek.com

The Republican early voting lead in swing states like NV, NC, and elsewhere, figures into my calculus for a Trump win.

Also, in key swing states like PA - there have been more Republican voter registrations than Democratic registrations, and conversely there have been more interparty defections from Democrats than Republicans.

Add in things like Trump gaining support to 40% of Hispanics, along with 25% of Black men, and I see a lot of things giving me pause.

 

[maxparrish]

The Republican early voting lead in swing states like NV, NC, and elsewhere, figures into my calculus for a Trump win.

Also, in key swing states like PA - there have been more Republican voter registrations than Democratic registrations, and conversely there have been more interparty defections from Democrats than Republicans.

Add in things like Trump gaining support to 40% of Hispanics, along with 25% of Black men, and I see a lot of things giving me pause.

Yes, all that plays into my own apprehension, but also making me think the huge gender gap and women's vote might tip it to Harris (unless this is an illusion).

And for what its worth, a 2005 Commission headed by Carter and Baker (a dem and pub) stated that early voting does not affect total turnout:

https://www.eac.gov/sites/default/files/eac_assets/1/6/Exhibit%20M.PDF

 

[Chomsky]

Yes, all that plays into my own apprehension,

but also making me think the huge gender gap and women's vote might tip it to Harris (unless this is an illusion).

And for what its worth, a 2005 Commission headed by Carter and Baker (a dem and pub) stated that early voting does not affect total turnout:

View attachment 67539357

https://www.eac.gov/sites/default/files/eac_assets/1/6/Exhibit%20M.PDF

BTW - The bolded is the perfect term for it - "apprehension"

 

[PoliSciPulse]

The Republican early voting lead in swing states like NV, NC, and elsewhere, figures into my calculus for a Trump win.

Also, in key swing states like PA - there have been more Republican voter registrations than Democratic registrations, and conversely there have been more interparty defections from Democrats than Republicans.

Add in things like Trump gaining support to 40% of Hispanics, along with 25% of Black men, and I see a lot of things giving me pause.

This is probably a dumb and uninformed question, but: How do they know Republicans are leading in early voting if the votes aren't counted until Election Day?

 

[Peter]

How about all the votes are personally hand counted by Trump himself so he can be sure there isn't fraud?

It's the only way to be sure.

 

[gbg3]

What I want to know is what happens if you remove the Trafalgar Group polls, which seem to go way right . I remember once one group had like +1 Harris and Trafalgar had like +6 Trump on the same day (I'm probably exaggerating it a bit, but the point is, Trafalgar Group stands out). Don't believe I've seen a single Trafalgar Group poll that ever had Harris up, and this includes right after the Convention, when Harris got her bounce.

It sure wouldn't make sense to remove Trafalgar. Look at where they came out in terms of "best" multiple state pollsters in 2016 and in 2020. It would be illogical to remove a pollster who was right at the top of the best list in the last two presidential elections. While their small average errors often favor Republicans, just look at the list of all the pollsters whose much bigger errors often or always favor Dems.

So, you are right about their errors often going right, but not at all "way right". Their errors are among the very smallest errors among the pollsters. As you can see here, Trafalgar is very good at what they do.

 

[gbg3]

This is probably a dumb and uninformed question, but: How do they know Republicans are leading in early voting if the votes aren't counted until Election Day?

The states don't know who the voters voted for, but they do know how many registered Republicans and how many registered Democrats have voted - unless they are a state like GA, which doesn't register voters by party affiliation.

 

[PoliSciPulse]

It sure wouldn't make sense to remove Trafalgar. Look at where they came out in terms of "best" multiple state pollsters in 2016 and in 2020. It would be illogical to remove a pollster who was right at the top of the best list in the last two presidential elections. While their small average errors often favor Republicans, just look at the list of all the pollsters whose much bigger errors often or always favor Dems.

So, you are right about their errors often going right, but not at all "way right". Their errors are among the very smallest errors among the pollsters. As you can see here, Trafalgar is very good at what they do.



View attachment 67539369

View attachment 67539368

I guess my larger point here is that I've noticed Trafalgar is an outlier. I'm sure there are others. There was one the other day that had Harris up by like double digits in a swing state poll not conducted by Trafalgar. That would also be an outlier in the opposite direction. My point still stands though: is there any reliable way to remove polls you consider to be outliers and then see what the resulting polls would be without finding every last one of them and removing them manually?

And maybe not "way right" per se, but compared to where the rest of the polls are, for most of the election cycle, they have been an outlier.

 

[CimpleMan]

It sure wouldn't make sense to remove Trafalgar. Look at where they came out in terms of "best" multiple state pollsters in 2016 and in 2020. It would be illogical to remove a pollster who was right at the top of the best list in the last two presidential elections. While their small average errors often favor Republicans, just look at the list of all the pollsters whose much bigger errors often or always favor Dems.

So, you are right about their errors often going right, but not at all "way right". Their errors are among the very smallest errors among the pollsters. As you can see here, Trafalgar is very good at what they do.



View attachment 67539369

View attachment 67539368

https://www.realclearpolitics.com/rcp-pollster-scorecard/

I know 2022 wasn't a Presidential election year, but Trafalgar didn't do as well.

 

[gbg3]

I guess my larger point here is that I've noticed Trafalgar is an outlier. I'm sure there are others. There was one the other day that had Harris up by like double digits in a swing state poll not conducted by Trafalgar. That would also be an outlier in the opposite direction. My point still stands though: is there any reliable way to remove polls you consider to be outliers and then see what the resulting polls would be without finding every last one of them and removing them manually?

And maybe not "way right" per se, but compared to where the rest of the polls are, for most of the election cycle, they have been an outlier.

But they are among the best at getting closest to the actual outcome - so with that in mind, I'd strongly argue being a correct outlier is far superior to being in a big pack of incorrect pollsters who are far off from the actual result. What pollsters are trying to do is get close to the outcome, not match the bulk of other polls.

Therefore, the last thing you want to do is remove the best polling, just because it's not in the pack. If you did that, the poll average would be further off instead of more accurate.

Being in a clump or pack doesn't indicate that clump or pack is necessarily doing a good job. Being closest to the outcome and consistently among the closest to the outcomes - does indicate doing a good job.

 

[gbg3]

https://www.realclearpolitics.com/rcp-pollster-scorecard/

I know 2022 wasn't Presidential election year, Trafalgar didn't do as well.

True, their success has specifically been at polling swing states and determining how those states will go in terms of the presidential race in those individual states.

And Trafalgar was very good at that for the last two presidential elections. We'll soon see how they do this year and if they'll make it 3 presidential elections in a row.

 

[CimpleMan]

True, their success has specifically been at polling swing states and determining how those states will go in terms of the presidential race in those individual states.

Will be interesting to see if the trend continues.

 

[gbg3]

Will be interesting to see if the trend continues.

Yeah, I just had that exact same additional thought and added it to the end of my comment.

 

[PoliSciPulse]

But they are among the best at getting closest to the actual outcome - so with that in mind, I'd strongly argue being a correct outlier is far superior to being in a big pack of incorrect pollsters who are far off from the actual result. What pollsters are trying to do is get close to the outcome, not match the bulk of other polls.

Therefore, the last thing you want to do is remove the best polling, just because it's not in the pack. If you did that, the poll average would be further off instead of more accurate.

Being in a clump or pack doesn't indicate that clump or pack is necessarily doing a good job. Being closest to the outcome and consistently among the closest to the outcomes - does indicate doing a good job.

Man, I've been clumsy today. Problems communicating. Grumble.

Let's try this: regardless of what I feel specifically about the Trafalgar Group polls, is there an easy way for me to remove them from so I can neatly compare what the Trafalgar Group polls say compared to the rest of the crowd without the Trafalgar Group polls? If so, how?

 

[Chomsky]

This is probably a dumb and uninformed question, but: How do they know Republicans are leading in early voting if the votes aren't counted until Election Day?

In many places voters are registered by party.

 

[CimpleMan]

The Polls Were Historically Accurate In 2022

Let’s give a big round of applause to the pollsters. Measuring public opinion is, in many ways, harder than ever — and yet, the polling industry just had one of…

fivethirtyeight.com

Here's a breakdown following 2022 Midterms. Polls still had a bias favoring Dems but down to .8.