NYT/CBS/YouGov crosstabs for every Senate and Governors race.

[Senator Cris]

https://twitter.com/Nate_Cohn/status/493447432395825153
It is an excel file.

[free my dawg]

This poll has Cory Booker and Tom Udall up by only 7. Throw this entire thing in the trash.

[Adam Griffin]

[IceSpear]

This poll has Cory Booker and Tom Udall up by only 7. Throw this entire thing in the trash.

Cory Booker up 7?

These polls look even worse than Sabato's ratings.

[ElectionsGuy]

While there are some good numbers, it looks kind of lazy. They didn't do any specific house races, just republican vs. democrat (which doesn't really help determine anything). There are also some states like New Jersey where all the numbers are completely off and you can just tell by looking, and there are states that completely contradict the polls that we've analyzed. Overall its wishy washy, and I don't think its a good source.

[Flake]

Should I enter these polls into the database? It seems like they've polled a lot of states that haven't been polled yet/as much as they should have.

Nevermind, they have Scott up 48-43 and Brownback up 52-40.

[Person Man]

At least one of my guys is at 50%.

[King]

...

Udall gets 50% of whites and 72% of hispanics and only leads 51-44? Did they mix up the demographics for New Mexico with Jacksonville FL?

EDIT: 80% of the respondents for this poll were white. This makes no sense for a poll out of this state unless I'm missing something completely.

[Small Business Owner of Any Repute]

...

Udall gets 50% of whites and 72% of hispanics and only leads 51-44? Did they mix up the demographics for New Mexico with Jacksonville FL?

EDIT: 80% of the respondents for this poll were white. This makes no sense for a poll out of this state unless I'm missing something completely.

The "poll" is self-selecting, that's what you're missing. It's 80% white, because 80% of the people who decided to go through the trouble of signing up to give opinions via the YouGov website are white.

[Tender Branson]

From what I remember, YouGov wasn't really that bad in both 2008 and 2012 with their state polls:

http://cdn.yougov.com/r/1/2012%20Election%20results%20table_Actual_YouGovLV.pdf

But some of these polls here make not much sense, for example Maine-Gov. (which has Cutler in single-digits, or SD - which doesn't even include Pressler who has 15%+ in other polls).

Warner only up 10, the same as Shaheen ? Unlikely.

YouGov is polling 3 more waves until November, so things might change a bit until then.

[Tender Branson]

Here's an overview of their Senate results:

[Senator Cris]

All the YouGov polls will be entered in the database?

[Tender Branson]

All the YouGov polls will be entered in the database?

I'm too lazy to enter 80 polls ...

If someone else wants to enter it though, do it.

[tmthforu94]

I have several disagreements, but considering that it is only July, for future reference, I think it would be beneficial to put them in.

I'll start working on it tomorrow if someone doesn't do it first.

[Senator Cris]

I think that it is better take the data from the excel crosstabs.
An example. In the NYT graphics, Risch leads Mitchell in ID 65-31, but in the crosstabs, Risch leads 63-30-1. In the graphics, Ernst leads 48-47, but in the crosstabs Ernst leads 46-45-2. And these are only two examples, we have many other cases ...

[Tender Branson]

I think that it is better take the data from the excel crosstabs.
An example. In the NYT graphics, Risch leads Mitchell in ID 65-31, but in the crosstabs, Risch leads 63-30-1. In the graphics, Ernst leads 48-47, but in the crosstabs Ernst leads 46-45-2. And these are only two examples, we have many other cases ...

Yes, rounding problems.

I used the "sum"-tool in Excel for the exact numbers here:

https://uselectionatlas.org/FORUM/index.php?topic=196305.msg4247062#msg4247062

[Flake]

These results do not seem accurate at all. I wouldn't really put them into the database, Tmth.

[Senator Cris]

In the crosstabs:
For the state of Wyoming, there is only the Enzi vs. Hardy race, not the Enzi vs. Wilde race.
For the Oklahoma special, there is only the Lankford vs. Johnson race, not the Lankford vs. Rogers race.
For Tennessee, there is only the Alexander vs. Ball race, not the Alexanders vs. Adams race.
For Alaska, there is only the Begich vs. Sullivan race, not the Begich vs. Treadwell race.
For Hawaii, there is only the Schatz vs. Cavasso race, not the Hanabusa vs. Cavasso race.
For Massachusetts, there is only the Markey vs. Addivionola race, not the Markey vs. Herr race.

[Tender Branson]

The polls were all done July 5-24.

This would mean Walsh was trailing by 16 even before his scandal became known, not really in line with other recent polls that showed the race tightening.

[Phony Moderate]

Take a look at the gender gap for the MA Gubernatorial race...jeez.

[Antonio the Sixth]

lolyougov

[Lief 🗽]

YouGov is a real pollster and these polls are non-partisan. They should be entered into the database.

[NHLiberal]

YouGov is a real pollster and these polls are non-partisan. They should be entered into the database.

They aren't weighted at all though; these are literally just the results of an online survey. Non-partisan doesn't mean credible.

[Senator Cris]

I think that there are reasonable numbers and not, but we must insert the polls in the database, as has been done with other polls.

[Tender Branson]

YouGov is a real pollster and these polls are non-partisan. They should be entered into the database.

They aren't weighted at all though; these are literally just the results of an online survey. Non-partisan doesn't mean credible.

The topline figures are based on weighted numbers, only the crosstabs in the Excel file are not.

You must be logged in to read this quote.