MT-AL: Rob Quist (D) vs. Greg Gianforte (R) vs. Mark Wicks (L), May 25

[cinyc]

The weighting stuff is irrelevant because a huge share of people voted Unknown in age bracket, no1 knows what their age is & makes the whole exercise meaningless. The number of people who selected male selected is higher than female, so don't know how weighting hurts Quist (but anyways problem of Unknowns here as well).

You could say this is a toss-up election with a fair share of undecided/3rd party voters !

We've experimented with different ways of weighting the unknowns.  Totally ignoring them in the 2016 Presidential GE polling generally lead to worse results than including them.  Since then, following the lead of RRH for the Google component of their polls, I've generally weighted the unknowns at 1.  Totally ignoring the unknowns gives Quist a 4 point lead.

There may have been more males overall in the poll, but there were fewer Male 65+s sampled compared to their 2014 counterparts or 65+ women.  That means that they got a weight of about 2.26, compared to 1.8 for Female 65+s or .31 for Male 18-24s.  And Males 65+ were Gianforte's best group.

The general problem with Google Surveys is that they tend to skew younger than the electorate.  Thus, not weighting a Google Survey poll overemphasizes those who are less likely to vote, generally (but not always) skewing it toward minor party candidates, like Libertarians, and (usually, but not always) Democrats.  IIRC, weighting my initial poll of this race last month actually worked in Quist's favor.  IVR phone polls tend to skew old, which is why RRH Elections has been running composite phone/Google Survey polls recently.

By the way, technically, nobody selected male or female.  Google infers sex and age data.

The thing I'm not convinced of is that these Google Surveys are worth spending money on.  The track record of Atlas' 2016 GE polling was terrible.

[cinyc]

RRH is trying to raise money to poll this race.  Due to the robocall ban, they need about $2000.  Yikes.  If I had only known yesterday, before I decided to put my crappy Google Survey into the field...

[cinyc]

An interesting thing to watch in this race will be the East-West divide. Basically all GCS polls so far have shown the gap to be unbelievably high (sometimes even more than 40 points), with Quist leading in the Western part of the state and Gianforte in the East. I doubt it will be that dramatic on election day, but we'll see.

Another thing to keep in mind is that the early returns are skewed toward Democrats. So please... don't freak out when Quist leads 55-40 with 5% in. Oh wait, this is Atlas, why am I even saying this.

Yeah, every single Google Survey poll I've had access to (including the one currently in the field) had a big East/West Gap.  The weighted and raw numbers:

3/12 to 3/14   Overall: Q+17 (Q+14 Raw); West: Q+33 (Q+29); East: G+9 (G+10) Gap: 42
3/14? to 3/16   Overall: Tie (Tie); West:Q+15 (Q+11); East: G+20   (Q+18) Gap: 35
3/18 to 3/20   Overall: Q+8 (Q+8); West: Q+14 (Q+13); East: G+6 (G+5) Gap: 20
4/6 to 4/8         Overall: G+1 (Q+2); West: Q+13 (Q+20); East: G+22 (G+26) Gap: 35
4/19 to 4/?       TBD

A 20-point gap is possible in Montana.  The 35/42 gaps are probably overselling the east-west divide.

[cinyc]

My Google Survey is now complete.  I put it into the field on April 19. It completed today, April 21. The question asked was:

You must be logged in to read this quote.

The choices were: (randomized) Democrat Rob Quist, Republican Greg Gianforte, Libertarian Mark Wicks and "I am not likely to vote in this election" (always last). As expected about 33% of the 533 respondents chose the "not likely to vote" option. Among the 356 respondents to answer with one of the candidates, the weighted results were as follows:

Gianforte 51%
Quist 41%
Wicks 8%

These results were weighted for sex and age to the percentage of those subgroups who reported voting in the 2014 November CPS survey. The raw results were Gianforte 49%, Quist 42%, Wicks 9%.  Google weighted to the Internet Audience, it is Gianforte 48%, Quist 42%, Wicks 10%.

As I've seen in the other recent Google Surveys of the race (more on this above), there is a huge divide between Eastern and Western Montana: Quist leads by 2 points (weighted)/9 points (unweighted) in Western Montana (n=188); Gianforte leads by 24 points (weighted)/23 points (unweighted) in Eastern Montana (n=160).

Eastern Montana is slightly overrepresented in the poll results. It makes up about 41% of the electorate in your typical election and 46% in the poll. I estimate that controlling for this would cause the poll to move about 2 points toward Quist.

There was no large gender gap in the raw results. Men were about 3 points less likely to choose Quist, but about 4 points more likely to choose Wicks. Suburbanites were much more likely to vote for Quist Q+6 (raw) than Rural residents G+30 (raw). Montana has very few urbanites, according to the poll. 87% of respondents earned $25,000-$49,999 per year, making discerning an income gap difficult.

[cinyc]

Thanks for putting a poll in the field cinyc!

If we take the trend in the Google surveys seriously, then it seems like the attacks on Quist have made an impact. Unfortunate, but not surprising.

Or Google Surveys are trash.  Consider that Castro ran a Google Survey immediately after my Quist +17 likely outlier, and found the race tied.

[cinyc]

Thanks for putting a poll in the field cinyc!

If we take the trend in the Google surveys seriously, then it seems like the attacks on Quist have made an impact. Unfortunate, but not surprising.

Or Google Surveys are trash.  Consider that Castro ran a Google Survey immediately after my Quist +17 likely outlier, and found the race tied.

Yeah, this is likely the case. The "if" in my statement was doing a lot of work.

This poll at least lines up with the recent Gravis poll, if you trust Gravis.  Again, a big if there. 

We need better pollsters to poll this race.

[cinyc]

  Lets assume the Emerson poll is somewhat off and Gianforte is in fact ahead in the high single digits. Shouldn't the race be closer than this based on what we've seen in Georgia and Kansas?

MT-AL is more rural than GA-6 or (probably) even KS-4.  Rural voters swung Trump in 2016, and didn't abandon the Republican candidate in KS-4 by nearly a much as Wichita urbanites/suburbanites.  Plus, Gianforte doesn't have the baggage of being tied to an unpopular Republican governor.

[cinyc]

Emerson is a weird poll, like really weird, but it's not good.  The good news I guess is that it seems dems have been over performing their polling, but probably not enough to overcome high single digit leads.  

Yeah.  Emerson provided the raw data, along with the cross-tabs.  Assuming Age Group 1 are the 18-34s, they polled a whopping 18 LVs out of the 648 or so respondents (after weighting).  I'm not sure exactly how they weighted, but given that 18-34s represent 123/648 in the presumably weighted cross-tabs, the 18-34s must have a weight of about 7.

So yes, Emerson's poll skewed old.  I think it also skewed heavily Western Montana in the formulation we've been using.  If I've done the county translations right, Western Montana is actually more pro-Gianforte than Eastern Montana in this poll, at least in the raw data.

Best I can make heads or tails of the Emerson numbers, the raw data was only Gianforte +4 (for the highly likely to votes).  They must have really weighted the data to get to Gianforte+15 out of that.

[cinyc]

What were the raw totals for the 18-34 age group? Considering they only polled 18 voters they probably had an unrepresentative sample to begin with.

9-6 Gianforte, with 3 Undecideds.

To be fair, even my recent Google Survey with more younger voters showed Gianforte up among at least the 18-24s (Quist only had a slight lead among the 18-34s, combined).  And I don't like unskewing polls.  Emerson's polls almost always skew old because I think they just poll landlines.  Yet they had a decent track record in 2016, despite their methodological shortcomings.

[cinyc]

By the way, even though the poll skewed heavily toward Western Montana, Emerson seems to have controlled for that when weighting the results.

[cinyc]

Montana Democrat Rob Quist Is Regular Performer at Nudist Resort
Free Beacon/Brent Scher

You must be logged in to read this quote.

More at the link, including a link to some pictures that I would not dare post here.

I'm not sure how much of a scandal this would be in a state like Montana.  Utah, for sure.  But Montana?

[cinyc]

Hunting records of U.S. House candidates released
TOM LUTEY/Billings Gazette via The Missoulian
Apr 26, 2017 Updated 36 min ago

You must be logged in to read this quote.

The oppo dumps on Quist continue, it appears.  I found this article in my check to see if the Nudist Resort story had any legs in the Montana media.  I don't see anything there about it yet.

[cinyc]

Watched the debate & Gianforte seemed like a hard right, uber religious, "muh defense spending increase & national security", strong supporter of Trump, cut benefits of Healthcare coverage, Climate change denying "Humans have some impact in Climate Change ala Scott Pruitt" guy. He has also got a big likeability & honesty problem. You may not agree with McCain's or Collins' policies, but they are honest & likeable. Gianforte seemed like a "Slimy Politician" 101 type person! Doesn't look like a candidate, a left leaning MT Treasurer type people will vote for !

He is obviously trying to move to to center just for this campaign but he seemed like a small version of Ted Cruz. Wouldn't be surprised if he joined the Freedom Caucus & Tea party lunatics!

Gianforte went to Fox & called Quist a Pelosi with a hat who wants a 90% Income Tax & said the Trump train is going forward well in Montana, praised Trump & is campaigning hard with Donald Trump Jr! That 90% Income tax is garbage but he can speak any lie & get away just like his hero Trump !

Trump's favorables are +13 in Montana, according to the Emerson poll.  So campaigning with Donald Trump Jr. and praising President Trump is simply smart politics.

[cinyc]

Does the DCCC not understand how to respond to attack ads, or do they just not care?

The DCCC (smartly, in my opinion - the less actual ties to Pelosi, the better for Quist) thus far gave  money to the state party to spend on the race.  I don't think they are directly involved in advertising.  It's up to Quist and the state party to respond to attack ads.

[cinyc]

"We don't wanna give him money because it'll nationalize the race" is such a cop out.

Except they are giving him money?

Yes.  The DCCC gave an undisclosed six-figure amount to the Montana Democratic Party for the MT-AL race.  That the DCCC is not advertising directly probably makes tactical sense.  Let the state party do it.

[cinyc]

roughly 50k voters are being diverted from their customary polling places due to the unusual date and manner of the election

Why did Governor Steve Bullock schedule this election for the Thursday before Memorial Day weekend, anyway?  He couldn't find a Tuesday that worked better?  Were his hands tied on the timing?

[cinyc]

Local papers are now reporting Quist under-reported his income.

Stick a fork in him, boys, he's done.

Ugh. That seems hard to recover from.

Quist's argument is that the original disclosure form was due before he did his 2016 taxes, so he didn't have all the numbers together when he did the original disclosure.  He characterizes the change as an amendment to the original forum.

It doesn't look good that this amendment only happened after the press started asking questions, though.

[cinyc]

I don't want to be a stereotypical lame lib, but until Trump releases his tax returns I really don't give a hoot what happens with Quists taxes and I assume many Montanans feel he's just an honest guy being attacked by Gianforte henchmen.

It's not Quist's tax returns which is at issue, but his financial disclosure form - which Trump filled out.  And Gianforte's henchmen didn't have anything to do with it.  The AP was the one asking questions.

But never let spin get in the way of the facts.

[cinyc]

Why is Gravis polling the Green Party candidate, Breck?  I thought he was thrown off the ballot.  Did that change?

[cinyc]

Why is Gravis polling the Green Party candidate, Breck?  I thought he was thrown off the ballot.  Did that change?

No, it didn't. Lolgravis.

Not only is Breck not on the ballot (he appealed all the way up to the Supreme Court to no avail), but according to his own website, he missed the deadline to accept write-ins, and endorsed Doug Campbell's write-in campaign.

So, LolGravis, indeed.

[cinyc]

Why is Gravis polling the Green Party candidate, Breck?  I thought he was thrown off the ballot.  Did that change?

https://uselectionatlas.org/FORUM/index.php?action=post;msg=5639799;topic=254607.775;sesc=4f121f212327b1da7976227f8fbc5200

No, it didn't. Lolgravis.

Not only is Breck not on the ballot (he appealed all the way up to the Supreme Court to no avail), but according to his own website, he missed the deadline to accept write-ins, and endorsed Doug Campbell's write-in campaign.

So, LolGravis, indeed.

I mean... even if he were on the ballot, there's no way all Third party candidates combined would get 8% of the vote. I'm also not sure why Undecideds went from 2% in the first survey (conducted April 27th) to 10% in the second one (conducted from May 2nd to the 4th) when the debate was held on April 29 and the race hasn't been getting any less attention since then.

Are you planning on doing another GCS Survey before the election, cinyc?

Gravis' first first poll was actually on or around April 6.  Uncertains there were 7%.  So the undecideds actually bounced from 7 to 2 to 10.  

They also polled Breck in their real first release, who got 2%.  But it at least made some sense to poll Breck at that time, as he was suing or appealing to get back on the ballot.  Now, it makes no sense.

I also don't understand why Gravis released two polls on the same day, instead of releasing the April 27 numbers last week.  lolGravis, I suppose.

I am planning on doing another GCS survey - but much closer to the election.  I'll probably start it around the 20th or 21st.

[cinyc]

When is my man Bernie getting up there to campaign with Quist?

I don't think they've set a date yet.  The Missoulian says sometime later this month, in their article noting that Vice President Pence will campaign and fundraise for Gianforte in Billings this Friday.

[cinyc]

What's the track record on these Google polls?

Not so good.  They've also been all over the place in this race, from Quist +17 (weighted) to Gianforte +10 (weighted).  Nevertheless, I'm probably going to do a final poll in the week before the election day.

Castro's poll is Quist +4 when weighted by just age and sex, like we've been doing.  I have been experimenting with also weighting by region, which ups it to Quist +5 because Eastern Montana was slightly overrepresented in the sample.  

As Castro said, the east-west divide is there, but it's almost so large as to be unbeliveable.  There also was a significant gender gap (women more pro-Quist than men) and age gap (65+s were pro-Gianforte, with many younger groups pro-Quist).

You can see this all at the linked spreadsheet.  Weighted Crosstabs are available in the 3 Factor Iterative Weight tab.  Unweighted crosstabs are in the Complete Responses tab.

[cinyc]

JW why does everyone keep calling him "Pianoforte"

Pianoforte is a real word.  It is another word for piano.  Someone must have originally posted Pianoforte as a result of an auto-correct of Gianforte's last name.  And since then, it's caught on as a nickname for Gianforte.

[cinyc]

So in reddit, 1 of the Bernie folks conducted another google survey.

Raw results -

Quist - 48.5
Gianforte - 41.9
Wicks - 9.6

Age -

18-24
Quist - 45.9, Gianforte - 32.8, Wicks - 26.2

25-34
Quist - 56.3, Gianforte - 24.6, Wicks - 10.6

35-44
Quist - 44.3, Gianforte - 28.7, Wicks - 7.9

45-54
Quist - 50.9, Gianforte - 56.8, Wicks - 6.1

55-64
Quist - 55.8, Gianforte - 39.2, Wicks - 6.2

65+
Quist - 43.0, Gianforte - 70.3, Wicks - 7.9

Male -
Quist - 41.5, Gianforte - 51.1, Wicks - 11.5

Female -
Quist - 56.8, Gianforte - 31.8, Wicks - 9.6
Obviously not a great poll, but if Wicks is doing this good among young voters, I think Quist can make up this ground.

Also interesting that Quist is winning in most Google Surveys (which tend to be not so good) ! If Quist has this a lead with women or young voters, then he can definitely win !

Did they provide a link to the survey?