NYT LIVE POLL THREAD:

[levgre]

Sienna College has an A rating over 60 polls on 538 so it seems they do a pretty good job.  Maybe they are weaker in some election types than others.

[Florida Man for Crime]

Likelihood of voting is based off of past voting as well as what the person says. If you are to say, "Ill definitely vote" and then respond that you havent voted recently, they wont mark you as a likely voter.

They do not ask anyone if they have voted recently. Vote history is hard data and is directly in the voter file. There is nobody responding that they "haven't voted recently" or that they have.

For an example of what data is in the voter file, you can look at the Ohio voter file here - https://www6.sos.state.oh.us/ords/f?p=111:1

However, this actually has substantially less data than the commercial voter file they are using.

As far as determining who is going to vote in their turnout model, if you are thinking that they are categorizing voters into strict "likely voter" and "not likely voter" categories like most pollsters do, that is probably not what they are doing.

When you are polling based off the voter file, there is no reason that it needs to be a binary choice for each individual voter. What campaigns do - and what I assume NYT/Siena is doing from their description - is to assign each individual voter a probability of voting based on their vote history and also their response to the "are you going to vote" question. And then they aggregate the results weighted by the probabilities for each individual voter.

That probability will be lower for x person who didn't vote in 2014 and says they will likely vote this time than for someone who did vote in 2014 and also says that they will likely vote this time, and that is entirely appropriate.

You must be logged in to read this quote.

LOL, I will have to remember to re-post this after they have completed some more polls of districts, including some more that are not rated tossups.

[Florida Man for Crime]

Their methodology is explained here.

https://www.nytimes.com/2018/09/06/upshot/live-poll-explainer.html

These are polls conducted based off the voter file, not using random digit dialing. This gives NYT/Siena access to accurate voter history and demographic data. Campaigns and the DCCC/NRCC use similar methodology internally, and this methodology is more expensive than most public polls.

[Zaybay]

You must be logged in to read this quote.

Ive already talked with another poster about this. They use both equally, and will devalue a caller if they havent voted recently. So the definitely category is both them saying definitely, and them voting recently. This was also done in Siena's other polls.

You must be logged in to read this quote.

If you really think every district that is rated tossup is going to be within 1% of each other, than I dont know what to say. Its rather statistically improbable, that all the districts rated tossup would be within 1%, and, TBH, it doesnt surprise me, considering this is Siena, who only switches to the same system everyone else uses in October. But if you want to follow a questionable pollster, which has over-sampled Rs in every poll they have done, and are using 2010 census data, then go right ahead.

[ON Progressive]

They aren’t using 2010 census’s data, they’re using the voter file.

[Sorenroy]

Why not, instead of bickering about what the NYT/Sienna might be doing, we instead actually link to their explainer on what they are doing and go from there.

https://www.nytimes.com/2018/09/06/upshot/live-poll-method.html

[Zaybay]

Why not, instead of bickering about what the NYT/Sienna might be doing, we instead actually link to their explainer on what they are doing and go from there.

https://www.nytimes.com/2018/09/06/upshot/live-poll-method.html

hmmm, its both worse and better than I thought.

Better: I was wrong about the census data, in fact, it doesnt even look like they use it period.

Worse: They are using 2014 as a baseline..........

[Florida Man for Crime]

If you really think every district that is rated tossup is going to be within 1% of each other, than I dont know what to say.

🙄

[Tintrlvr]

Interesting that the Roskam poll looks like it has zero responses from west of Palatine. What's that neighborhood like? It's hard to get a read on it from Google Maps.

It's a very low-density area compared to the rest of the district (especially adjacent areas to the north and east), and a significant portion is the Spring Lake Forest Reserve, so not surprising that they had few contacts there and no completed responses. It's very wealthy and very Republican. It's a small enough part of the district, population-wise, that I wouldn't think a lack of responses skews the result at all.

https://en.wikipedia.org/wiki/Barrington_Hills,_Illinois

[Xing]

Sure, there are issues with the methodology, but it's not like internal polls are going to be more accurate, especially since they're often released to send a specific message.

[john cage bubblegum]

Sure, there are issues with the methodology, but it's not like internal polls are going to be more accurate, especially since they're often released to send a specific message.

There's things we can nitpick at, but overall I'm quite happy with this project.  We get very few quality, non-internal House district polls.  This is giving us 1-2 district polls per day, with presidential approval and generic ballot numbers.  Not to mention they show us what the margin would be under different turnout scenarios and weighting schemes.

[Fmr. Gov. NickG]

If you are polling from an existing file of registered voters, how do you account for people who haven't registered yet, but will prior to the election day?  Voters in every state still have about a month left to register.  And I have to imagine those voters will heavily favor Democrats.

[Tintrlvr]

If you are polling from an existing file of registered voters, how do you account for people who haven't registered yet, but will prior to the election day?  Voters in every state still have about a month left to register.  And I have to imagine those voters will heavily favor Democrats.

You ignore those people. They're a small enough portion of the electorate that they wouldn't matter more than a few tenths of a percent at most in margin anyway. Which can certainly make the difference in a very close race, but polling is not precise enough to make distinctions at that level of detail.

[Predictor]

I'm gonna guess that VA-05 or VA-02 will be the next polled district.

[Xing]

I'm gonna guess that VA-05 or VA-02 will be the next polled district.

I'm going to guess OH-01

[Politician]

Pureval will win in OH-1, with T Ryan coattails, but R Cordray may lose to M Dewine

[Predictor]

And the next district is TX-23

[Sherrod Brown Shill]

Pureval will win in OH-1, with T Ryan coattails, but R Cordray may lose to M Dewine

Wow Ryan will win so big he'll have coatails on the other side of the state? /s

[Politician]

Pureval will win in OH-1, with T Ryan coattails, but R Cordray may lose to M Dewine

Wow Ryan will win so big he'll have coatails on the other side of the state? /s

It's mocking OC lol

[Sherrod Brown Shill]

Pureval will win in OH-1, with T Ryan coattails, but R Cordray may lose to M Dewine

Wow Ryan will win so big he'll have coatails on the other side of the state? /s

It's mocking OC lol

I actually just thought you mixed up Sherrod Brown with Tim Ryan, in which case your post wouldn't be crazy at all.

[Brittain33]

I'm gonna guess that VA-05 or VA-02 will be the next polled district.

Not with Florence coming.

[Zaybay]

Next poll is for Will Hurd's district. Will be interesting to see the results.

[IceSpear]

Sure, there are issues with the methodology, but it's not like internal polls are going to be more accurate, especially since they're often released to send a specific message.

Yeah, it's pretty amusing that people are taking issue with this considering how many users here gobble up any junky (and possibly fake) internal poll that pops out of the woodwork like Pacman.

[new_patomic]

Sure, there are issues with the methodology, but it's not like internal polls are going to be more accurate, especially since they're often released to send a specific message.

Yeah, it's pretty amusing that people are taking issue with this considering how many users here gobble up any junky (and possibly fake) internal poll that pops out of the woodwork like Pacman.

GA-06 is obviously a toss up now and all these polls need to be thrown in the trash for saying toss ups are toss ups.

[ltomlinson31]

https://twitter.com/Nate_Cohn/status/1039223425116585987?s=19