Poll Diving – part 5

Before moving on, let's back up a bit to Wisconsin. First I neglected to include 2018 Senate race– Baldwin (inc D) 55.4% Vukmir 44.6%. Up for Baldwin in 2012 at 51.4%. In 2016 the incumbent GOP Senator Johnson won with only 50.2%.

Second is the information John Nichols presents in this interview:

Key Points:

1) WI is an evenly split state. It's deep pattern is to have very close presidential races.
2) WI is not a liberal state. It's a progressive populist [and republican] state that responds to economic issues.
3) COVID-19 is a very serious issue in the state from both a health and economic perspectives.
4) Rural support for Trump has softened. It's not about Biden, but that Trump has been cavalier about the economic consequences for rural WI in his trade wars and neglected the rural need for federal money to support hospitals and schools.
5) Down ballot races – young, progressive candidates. Building on the major turnarounds in 2018.
6) Foxconn – the mega-factory touted by Trump that would restore the economy in SE WI. The huge tax concessions granted to Foxconn, freeways rerouted, etc. has produced nothing. It's referred to as Fox-Conned in WI and was a major issue in the 2018 elections.

#1 was touched on in part 3. For #2 think Robert M. La Follette. Populist rhetoric resonates better in WI than other states. Even though it was bullshit, that's what gave Trump the advantage. #3 confirms that COVID-19 is an issue in WI. Trump has blown that and #4 and #6. Finally, there are WI activists and politicians seizing the opportunity that Trump opened.

The Wi polling (with the exception of the GOP Susquehana and Trafalgar) confirm Nichol's narrative. (And not even the GOP pollsters can get Trump close to 50%.) Rhetoric isn't going trump his performance in office.

Many more polls will be released in the next few days and while we wait for them to confirm or reject what could be seen over the past days, I want to consider other races.

The Senate

There are eighteen Senate elections – GOP 10 and Dem 8 – that are getting almost no coverage and therefore, no polling. There's not much noteworthy in those races except in NM and TN. NM US Representative Ben Lujan will become a Senator. Marquita Bradshaw, the TN Democratic nominee, won the primary without party backing or money. A real grassroots nominee with no chance in the general. These eighteen seats will absolutely not flip even if the incumbent party candidate dies before election day.

Races deemed undecided and not covered in parts 1 – 4 are: CO, KS, and KY. The last shall be first.

McConnell never begins his reelections in a strong position with his constituents. But he always wins. Nothing for Democrats, liberals, or leftists to get excited about this year. That's not a comment on his opponent or any of his past opponents; it's merely a recognition that KY voters aren't about to toss out a Senator that they don't much like when he's achieved power in DC. (Even if he wields that power to their detriment.) Same condition exists in the San Francisco House seat; although Pelosi is a better match with her constituents than McConnell is with his.

On Kansas, the good news is that after sixteen years in the House and twenty-four years in the Senate, the eighty-four year old Pat Roberts is retiring. The bad news is that the GOP nominee is as bad if not worse. It's a race between two physicians, an anti-science rightwing physician and a Republican physician who for the past two years has called herself a Democrat. Surprisingly, the Democrat has a large cash advantage over the Republican and currently it's a bit of a tie. Bollier appears to know what she needs to do, energize former Governor Sibelius voters and Sibelius has endorsed her. As Kansas voters have soured on the extreme rightwinger Kris Kobach, Bollier may have a chance. OTOH, Marshall looks like a Kansan and Trump's support hasn't softened here. This wouldn't be my wild-ass upset pick because the suggestions aren't strong enough.

Colorado. The shape of the 2014 mid-terms was advantage GOP. That and first term Senator Mark Udall's lackluster reelection campaign, he appeared tired and Gardner appeared youthful and energetic, is what made the difference. Not much polling here, but Hickenlooper's 2020 failed presidential campaign before jumping into the Senate race hasn't hurt him and the GOP has pulled back on funding Gardner. So, this seat flips.

The current Senate headcount is 53 Republicans, 45 Democrats, and 2 Independents that caucus with Democrats. Seats being defended are GOP 22, Dem 12, and I O. Only four of those seats are open and only one of them, KS, could possibly flip. Only one Dem seat, AL, is vulnerable and will flip.

All over except the counting of the ballots: AZ and CO.
Almost over: IA and ME.
Vulnerable in order of current vulnerability: NC, MT, and AK.
Not yet lost: KS. GA. and SC, but not a good idea to bet that they will.

Democrats need a net gain of four to retake the majority. The best chance for that fourth is NC or MT (and we might have something to talk about if MT ends up being the fourth).

13 users have voted.


a good idea? (You seem to think they are an outlier WRT Wisconsin)

They were, after all, about the only pollster to correctly call 2016.

Their October WI poll (conducted Oct 14-16) does have Biden leading - but by less that 1.5%, well within MOE.

In Michigan, (conducted Oct 11-14) they have Trump up, although by less than 1% - and Republican challenger John James up by slightly more over incumbent Peters link

Then there is Minnesota:

The race for U.S. Senate in Minnesota has narrowed substantially before the election, with the incumbent, Democratic Sen. Tina Smith, shedding her double-digit lead and finding herself in a dead heat with Republican Jason Lewis, according to a new poll.

A KTSP/SurveyUSA poll released on Wednesday showed Lewis and Smith in a virtual tie (43% to 42%). As recently as early October, Smith enjoyed a seven point lead over Lewis. That was even larger in mid-September when she led him 47% to 36%.

According to 5 Eyewitness News, suburban women -- a key and often-watched demographic -- have helped drive the shift by leaving Smith. Carleton College political analyst Steven Schier told the outlet that the race "could go either way and it's a bit of a surprise because Tina Smith has had a lot more resources than Jason Lewis."


Suburban women, huh... FWIW the above movement is happening despite Smith having out-raised Lewis by $14 million to $5 million.

Presumably the MN presidential contest is similarly tight...

That suggests the possibility of two senate seats flipping D to R plus Trump keeping WI and MI and flipping MN. Big enthusiasm gap overall in favor of R - looks like Trump is the one with momentum - his favorability according to Rasmussen is over 50% - with "strongly approve" from whites at 43%, blacks 27% and "other" 45%. That amount of black support cannot be comforting to Dems. source


This is going to continue to be interesting...

Interesting interview the other day with Robert Cahaly, Trafalgar Group' head pollster -
can't seem to link directly to video of the interview but you can from here

1 user has voted.

Is excepting Trafalgar a good idea?

I'm not exactly doing that. However, in every election cycle different pollsters get closer than others to getting it right and there's no magic cube to tell us which ones it will be. Plus, some get a better read on some state or states but can be way off on national polling. So, this is a bit of a guessing game for reading polls.

I don't know the basis for Trafalgar supposedly 'getting it right' in 2016. An example: say pollster A affiliated with party X comes up with 53% for X candidate and 47% for Y candidate in state 1. Pollster B unaffiliated gets X at 48% and Y at 49% in state 1. The actual results are X 49.5% and Y 49%. Which pollster was more correct?

In 2016 there was a A/USC national tracking poll that had Trump slightly leading. It was mostly dismissed by those like 270towin because it differed by too much from other national polls that had HRC leading. As it was a tracking poll, I gave it some weight, not nationally (which doesn't count anyway), but in states that were polling close to toss-ups, such as FL and NC. However, nationally that tracking poll was very wrong.

When the best an affiliated pollster (ie Trafalgar) can do is get results indicating its preferred candidate isn't losing by as much as what other pollsters are getting, I take that as confirmation that of a loss in that state.

In 2016 there were two biases among politicos: 1) Clinton can't lose and 2) Trump can't win. It was clear to me early on that Trump could win the GOP nomination and Sanders was strong enough that he could win the nomination. Both of those interpretations were correct (and if not for the DNC cheating, Sanders would have won). To repeat myself, for the general election, I only managed to see that Trump was only one mid-sized state short of a win with PA and NH extremely close. I paid no attention to MI and was blindsided by the Feingold polling in WI. I got PA wrong because I ended up deferring to its recent history -- with a D presidential win and R Senate win and not my projected trend in other states that there would not be ticket splitting.

At the moment I don't know enough about on-the-ground factors in MN and have acknowledged that the polls concern me -- but I'm possibly being overly conservative. Tend to let others parse out whatever white demo catches their fancy in any election cycle -- it was also suburban white women in PA in 2016. And MN was very close in the 2016 presidential election. So far I'm not seeing nearly enough down trending for Smith and increases for Lewis to put this one in the GOP column.

As for Trump holding MI and WI, none of the polls and none of the additional factors suggest any such thing.

2 users have voted.

@Marie @Marie

checking "My Comments" shows zero replies to my "Is excepting Trafalgar" comment...

Anyway, we'll see what happens. Pretty wildly conflicting polls out there - but it is no longer just Trafalgar that is showing tightening - Rasmussen Oct. 26 released Trump v. Biden General election has Trump up - by a single point, but I think that is a first since Biden's nomination.

And Trump doesn't need to worry about being down a couple points in the PV, anyway, as millions of those are going to be in California and basically irrelevant. OTOH is there any plausible path for Dems if they were to lose the popular vote?

A lot going on in the swing states - Republicans now ahead of Democrats in returned ballots in Wisconsin. And in FL a number of polls have Biden slightly up, but here comes Rasmussen again showing (Oct. 23) Trump up by four.

In the interview with Trafalgar's pollster he indicated a number of areas where he believed other pollsters were tending to undercount Trump support.

Among his points:

- The 'shy Trump voter' phenomena is a real and significant thing - that a significant number of Trump supporters were untrusting of polling and worried about possible negative repercussions if their views were expressed candidly.

- That (possibly related to the above) black and Hispanic support for Trump was being underrated - that Biden was falling well short of the black support he needs and that in Florida and Arizona Trump support among Latinos is over 50%.

- also what he termed "low propensity voters" who had voted only rarely in previous elections were being undercounted and that such voters were breaking heavily for Trump.

Note that the Rasmussen snapshot I posted above shows 28% of black and 45& of "other" as Strongly Approving of Trump. Trump managing to translate half of that strongly approving black support to votes it would constitute a 75% increase from 2016's 8% - and that was the best Republicans had done in decades.

Republicans may be helped in that regard by having a much-improved field of young, energetic black candidates down-ballot. This might not move the needle much in heavily blue states such as Maryland or California - but could be significant, even decisive in Michigan...

New Jersey CD 9 - (Incumbent D- Bill Pascrell)

Billy Prempeh

Californis CD 43 - (Incumbent D Maxine Waters)

Joe E. Collins III

Maryland CD 7 - (Incumbent D Kweisi Mfune = this is Elijah Cummings old district)

Kim Klasik

Michigan US Senate (Incumbent D Gary Peters)

John James

0 users have voted.

@Blue Republic There's no narrative to your belief in the GOP pollsters that would account for a significant shift to Trump in the past week. All you've got are "shy Trumpsters" that other pollsters can't find. "Shy Trumpsters" that the GOP pollsters couldn't find a couple of weeks ago either.

Check out the detail in this PA Trump +2 poll: Insider Advantage In 2016, Johnson/Weld took 2.38% in PA, but we're supposed to believe that Jorgenson will get 3.3%. The only thing of note and interest in this joke of a poll is the collapse of >64 years old vote for Trump. That's consistent with other polls. Nationwide in 2016 Trump got 52% of that vote. At 43%, he loses.

0 users have voted.

@Marie @

showed Biden ahead.

Things have moved in PA - Matt Towery, who oversaw this and previous Insider Advantage polling explains what he sees is going on there

The same poll had Biden in the lead less than two weeks ago. Why? InsiderAdvantage’s Matt Towery explained why things have moved in Trump’s favor:

“These results indicate a stark shift in the contest. Our last survey of Pennsylvania showed Joe Biden leading Trump by three points. But that survey was before the last debate. Since the debate Trump has picked up support from younger voters, who based on our prior survey strongly oppose future lockdowns over Covid-19 spikes. Trump has also bolstered his lead among male voters by some twelve points. Biden continues to hold a seven point advantage over Trump among female voters. It would be nothing more than mere conjecture to attempt to correlate Biden’s statements on energy and fracking in the last debate contest with the shift towards Trump in this survey. However, Trump saw gains even among senior voters which have not been his strong suit this election cycle. That suggests that some issue or set of events has caused a late shift in the contest.”

Towery adds “Trump also continues to hold about 14% of the African American vote in this survey. In twenty years of polling, and as one who has polled Pennsylvania many times, I have never seen a Republican candidate consistently hold these type of numbers among black voters this close to an election. And this appears to be a developing trend in numerous states.”

Matt Towery accurately predicted a national victory for Donald Trump for major owned and operated Fox affiliates on air the evening before the 2016 presidential election. He has polled presidential races across the nation for twenty years. Specifically he has polled Pennsylvania for numerous news organizations, including the 2008 contest between Barack Obama and John McCain for Politico where his surveys showed an Obama victory. He retired from his nationally syndicated column (Creators) in 2016 and as CEO of InsiderAdvantage but returned to provide polling of the Trump-Biden contest this year.



"...suggests that some issue or set of events has caused a late shift in the contest.”

Wonder what that/those could be? Would voters starting to pay attention be considered an 'event'?

"In the beginning was the thing. And one thing led to another."

- Tom Robbins

Yes, we will see.

0 users have voted.

@Blue Republic to pollsters, election state pros, and polling analysts pros tooting their own horn. It's part of how they make their money; just like the political operatives. Nate Silver did nothing exceptional in 2008. Or should say nothing that I didn't do on my own in 2004, 2006, and 2008 including the Senate seats. I called 2012 in January 2010.

As for 2016, I got a lot closer to being right than Silver did. At least I knew that Trump was only one state short.

I use some of my very rusty stat knowledge (the numbers and polling methodology, some history, and some psychology. And never forget that when it gets very close, a call may be no better than flipping a coin.

1 user has voted.

The left side (politically) in WI is progressive. One way to read that is "concerned with laborers/farmers instead of corporations".

It used to be much more progressive than it is now. Deep-red Republicans and "Tea Partiers" have really turned the suburbs of Milwaukee (and really around the Wakesha, Racine, Kenosha areas, just outside the suburbs of Milwaukee) really conservative. Same around Green Bay up in the northeast of the state, and Lake Winnebago area--generally very red.

Oddly, in the western rural areas of the state, progressivism is still largely dominant (La Crosse, Eau Claire). Maybe that's Minnesota influencing things, but it is also probably somewhat due to mostly Scandanavians populating the west side of the state more, and Germans the east (not to imply that either nationality has innate right or left leanings, but there is definitely a different *feel* of the culture and outlook on either side, and I think the Koch-brother efforts in the state over the past decades has resonated more with eastern Wisconsin for probably a myriad of reasons and happenstances).

Madison and Milwaukee (the city itself, not including the suburbs) remain very progressive.

Edit to add:

A typical year Wisconsin map in 2012--note blue is Republican and red Democratic in this map (reverse of usual). Generally, west is more Democratic, although it does look as though a bunch in the northern west is trending more Republican (which is a newer thing). But, you can see the Republican strength is in the eastern parts of the state:


However, in 2016, just nobody in the state liked Hillary. Progressive state, not liberal! I expect 2020 to be more like 2012 in the state.


4 users have voted.

@apenultimate Wonder if the absence of progressive populist rhetoric by Democrats in the eastern parts contributed to the Koch driven strong capture in the suburbs.

2 users have voted.


Keep in mind, Wisconsin turning Republican in 2016 was entirely because of Jill Stein! Pay no attention to the 22 counties that switched from Democratic to Republican from 2012 to 2016. That had absolutely nothing to do with the Democrats losing Wisconsin!

It was really all Jill Stein. (/snark)

3 users have voted.

@apenultimate only Wisconsin voters saw before they aired after the election.

I'm still trying to get used to how totally dumb DC Democratic politicians have become. "The commies done made me lose" used to be the domain of Republicans who believed that the people really wanted low paying jobs and employment insecurity.

1 user has voted.

@apenultimate the GOP collapse in 2018. Trump and Republicans have done nothing since to reverse that. Trump could exacerbate that in his next campaign rally by again touting Foxconn.

2 users have voted.

He really is thick.

2 users have voted.