NYT LIVE POLL THREAD:

[Florida Man for Crime]

OK, I will start posting here. I think this should become a megathread.



Of the various weighting options, it seems like across all the polls the way NYT/Siena is weighting is the most GOP-friendly of all the other options they list.

Dems generally do better when you don't weight by primary vote, use census data instead of voting records, and don't weight by education.

However, I think the NYT/Upshot methodology on weighting is the most methodologically sound of the options, and likely the most accurate as compared to the others. That means that some other public polls may be overstating the Dems, perhaps.



As opposed to weighting methodology, it is a little bit harder to tell a systematic difference between different turnout models. So far, in some districts the broader turnout models seem better for the GOP, and in others better for Dems.

[Florida Man for Crime]

Where are we getting 5 a day from? They only plan to do 100 races with 61 days until election day.

Well, they are running 5 polls simultaneously now. I am just assuming they will keep doing the same amount as they are starting off with (though you are right, that is just an assumption, so maybe not).

However, all the polls are conducted over multiple days, and it seems like they are staggering when they start the polls. So my guess is that perhaps at any one time there will be about 5 polls running, but that is entirely consistent with having only 1 or 2 NEW polls start (or finish) each day. The other polls going at any given time will be ones that already started, but need another day or two to get more responses.

[Florida Man for Crime]

Anyway, all these numbers seem believable. I think it's more likely than not to get worse for Roskam (based on what my best friend has alluded to who works for him) and Rohrabacher. Possibly Bost too.

I hope they poll some of ME-2, CA-25, CA-39, CA-10, and VA-2 next.

What makes you think it would get worse for Rohrabacher and Roskam? In both cases, the undecideds seem to lean Republican.

IL-06: "About 11 percent of voters said that they were undecided or refused to tell us whom they would vote for. On questions about issues, these voters most closely resembled Republicans."

It's currently 45 GOP - 44 DEM -11 UNDECIDED

CA-48: "About 10 percent of voters said that they were undecided or refused to tell us whom they would vote for. On questions about issues, these voters most closely resembled Republicans."

It's currently 45 GOP - 45 DEM - 10 UNDECIDED 

It would be pretty surprising if the undecideds didn't more closely resemble Republicans than Democrats. These are generally Republican-leaning districts.

There are more Republican-leaning voters than Democratic-leaning voters in the districts Democrats are trying to pick up in general, so it stands to reason that more of the undecideds would be Republican-leaning.

[Florida Man for Crime]

MN-03 poll starting tonight. I will guess that Paulsen is down by 4 at the end of the night (with an incomplete sample).

https://twitter.com/Nate_Cohn/status/1038110434350575623

[Florida Man for Crime]

KY-06 and MN-08 are running again

KY-06:
Andy Barr - 46%
Amy McGrath - 44%

MN-08:

Joe Radinovich - 45%
Pete Stauber - 38%

Oh no, Stauber caught up to Radinovich in MN-08 now that we are up to 78 responses...

42-42 now. Red wave incoming!


--- edit ---

79th respondent went Dem. Radinovich pulls ahead 43-41!

BLUE WAVE INCOMING!

[Florida Man for Crime]

Regarding KY-06, yesterday they way undersampled 18-29 year olds. They only got 2 responses out of 371 18-29s called yesterday.

So hopefully they get some more responses from the youngs, and perhaps that may be one thing that helps push McGrath back into the lead. But we will see...

[Florida Man for Crime]

Regarding KY-06, yesterday they way undersampled 18-29 year olds. They only got 2 responses out of 371 18-29s called yesterday.

So hopefully they get some more responses from the youngs, and perhaps that may be one thing that helps push McGrath back into the lead. But we will see...

That's weird. One response caused the margin to go from 46-44 Barr to 47-44 McGrath. Was it a glitch or something? Or were they not calculating the percentages properly before then?

See my earlier comment about 18-29s above.

They are now up to 4 responses from 18-29s (and McGrath has 100% of them so far). Because they weight by age, the few responses they get from 18-29s count more.

So what you saw was may have been the 18-29s coming in to the sample.

[Florida Man for Crime]

It is interesting just how different MN-08 and KY-06 are in terms of name recognition. Both candidates in KY-06 have only 10% who don't have a fav/unfav opinion. But in MN-08, 61% have never heard of Radinovich and 54% have never heard of Stauber.

[Florida Man for Crime]

Yeah... nyt should either try to get the sample up to 800 or weight the results less

I'd rather have polls of 8 house districts with a 500 sample size each than 5 disstricts with an 800 sample size each. Since there are so many districts competitively within the reach, that is the only way to get polls for a good number of them, and 500 is a pretty reasonable sample at least for the topline.

[Florida Man for Crime]

The MN-03 poll has begun. 4 calls, 0 responses so far

[Florida Man for Crime]

The first respondent in MN-03 is voting Republican.

Titanium Safe R MN-03, you heard it here first.

Paulsen is literally up 100% to 0%. No way such a large lead can be made up.

[Florida Man for Crime]

65+ white female on a cell phone non-college, HS grad or less, republican, almost certain to vote,prefers Republicans to retain the house, doesn't know if she supports assault weapon ban, disagrees that she feels like a stranger in her own country, disagrees that discrimination against whites has gotten as bad as against blacks, economic situation is better than parents, doesn't bother her to hear immigrants speak a foreign language in a public place, and doesn't know someone who has struggled with addiction to opiods.

[Florida Man for Crime]

NYT call center for MN 3rd live look

Pretty much. It seems like they only have a single person calling, and it is not with computer-assisted dialing (I guess?)

[Florida Man for Crime]

65+ white female on a cell phone non-college, HS grad or less, republican, almost certain to vote,prefers Republicans to retain the house, doesn't know if she supports assault weapon ban, disagrees that she feels like a stranger in her own country, disagrees that discrimination against whites has gotten as bad as against blacks, economic situation is better than parents, doesn't bother her to hear immigrants speak a foreign language in a public place, and doesn't know someone who has struggled with addiction to opiods.

Doesn't sound like a typical MN-03 resident.

Also she lives in pretty much the North-east corner of the district, in the Coon Rapids area.

[Florida Man for Crime]

It is a non-college educated white male Dem. I thought they were supposed to all be Republicans now!

He disagrees with the Republican lady on everything except agreeing with her that discrimination against whites is not as much of a problem as against blacks and also not being bothered by hearing immigrants speak a foreign language in public.

And he lives on the other side of the district in Bloomington (south-east).

[Florida Man for Crime]

MN-03 is starting to confirm the stereotype a bit more.

Phillips is really killing it with white college grads, post-grads, and independents. 100% of all of those.

However, he is up 80-20 with males. Paulsen really underperforming for a Republican with men. Maybe he needs to do some sort of ad to make himself appear more masculine somehow.

[Florida Man for Crime]

Reading into crosstabs unironically in a poll with two responses so far is peak Atlas.

I stand corrected.

LOL Paulsen is doing better (tied) with non-whites than with whites (whites going Dem 59-34). I guess he must have really great minority outreach, and Trump is gaining ground with non-whites.

[Florida Man for Crime]

#analysis in this thread is godawful even by Atlas standards.

I'm pretty sure everyone is joking about the MN-03 stuff.

[Florida Man for Crime]

Paulsen has pulled into the lead, with the help of the one and only 18-29 year old respondent, who is voting Republican. #GENZISGENGOP

[Florida Man for Crime]

Interesting that they stopped polling KY-06 so close to getting 500 responses. Since they've targeted the Midwest quite a bit so far, I wonder if the next district that they're going to poll will be somewhere in the South, or maybe a Northeastern district like NY-19 or NY-22.

I'd really rather see some KS-02/KS-03 polls to be honest. Much more interesting races.

Also with 93 respondents as of right now, Paulsen is +9.

I'd like to see races that are currently "Likely R" polled.

We know that the races currently rated tossup are generally competitive, more or less. I don't really care that much if the Dem is up 3 or down 3 in those, we know that they will generally be in that sort of area already.

On the other hand, we could see something surprising in those Likely R races, and that would give us an idea of to what degree they are within reach (or not).

Also, I would really like to see the handful of races where ratings are mostly based upon private un-released polling, such as TX-23 and CA-21. Do polls really show the R incumbents safe in those districts? Would be nice to get confirmation (or disconfirmation) in a public poll.

[Florida Man for Crime]

They are resuming polling again. Looks like they are polling during the day on the weekend.

[Florida Man for Crime]

Dems are gaining at the moment in MN-03. Paulsen's lead is down to 49-46, up to 116 respondents.

[Florida Man for Crime]

Phillips up 49-44 over Paulsen in MN-03 now (378 respondents). Looking a bit more like a burbstomping now. Will the momentary fluctuation of Phillips' lead hold up?

[Florida Man for Crime]

As the three white males of WV-03 go, so goes the nation.

[Florida Man for Crime]

WHITE FEMALE COLLEGE GRAD FOR OJEDA TO SAVE THE DAY!!!