Is Jon Tester Really the Democrats' Only Hope to Hold onto the Senate? (also: a follow-up to yesterday's post)
September 24, 2024
Source: Washington Post
[Montana Sen. Jon] Tester, known for his signature flat-top haircut and three missing fingers—he lost them in a meat grinder when he was 9 years old—is the most vulnerable senator up for re-election. If he goes down in this red-leaning state, so does the Democrats’ Senate majority. They currently have a 51-49 margin but are sure to lose a seat in deeper-red West Virginia, where Sen. Joe Manchin is retiring.
A loss by Tester would give Republicans a stronger hand in rewriting the tax code for millions of American households and businesses in 2025. The party that controls the Senate confirms Supreme Court appointments and can approve—or block—cabinet picks.
That is why this race is attracting tens of millions of dollars in outside spending and campaign ads. But for voters here, there is something more personal gnawing at them. The country is changing. Montana is changing. Has Tester changed too?
I find it annoying that so many in the media seem to be just repeating the same old conventional wisdom that the race comes down to Tester (and to a lesser extent Brown). Some of the races that journalists (and some polling analysts, to be fair) continue to consider toss-ups are in better shape for Democrats than Tester’s Republican opponent is in Montana. There are other states in which Republican incumbents are in real races. While the fundamentals might tell us those incumbents will eventually win, they are in close races.
As I have said many times before, the political conventional wisdom is always right - until it isn’t. In the 2022 midterms, almost all journalists and pundits thought Republicans were going to win big. The polling looked bad for Democrats, it was a mid-term election (so the president’s party had to lose seats, right?), and Republicans kept repeating that they were seeing a red wave coming. However, the polls were flooded by crappy Republican PR firms masquerading as polling outfits, first term presidents losing mid-term elections is based on a very small sample size of elections, and voter registration and early vote data was very good for Democrats. I was not the only person who saw this, but there were not a lot of us who at least admitted we saw it. I wish journalists would focus more on what we are actually seeing instead of what they expect we should be seeing.
Does this mean Democrats are on their way to big victories this fall in the Senate? No. But there is more going on in more races than you would guess from reading a few newspaper articles about it. To get a sense of how the reporting of the race for a Senate majority may be missing what’s really going on, we can compare the polling in the races the CW says are competitive and add a few more that they are missing or not taking too seriously.
Let’s look at the latest polling from 2 and 3 star pollsters for Senate candidates.1 The first group are the races the media continually refers to as the ones up for grabs. The second group are races that seem to be just as much up for grabs (except for Maryland, but Republicans seem to think so privately, so I included it).
Group 1 Races: Conventional wisdom says these are competitive
Arizona: Rep. Ruben Gallego (D) vs Kari Lake (R)
New York Time/Siena College: Gallego +9 (50-41) FW:2 Sept 17-21
Emerson College: Gallego +6 (48-42) FW: Sept 15-18
Morning Consult: Gallego +14 (53-39) FW: Sept 9-18
Michigan: Rep. Elissa Slotkin (D) vs Rep. Mike Rogers (R)
Suffolk University: Slotkin +2 (45-43) FW: Sept 16-18
Emerson College: Slotkin +5 (47-42) FW: Sept 15-18
Morning Consult: Slotkin +14 (51-37) FW: Sept 9-18
Marist College: Slotkin +7 (52-45) FW: Sept 12-17
Quinnipiac University: Slotkin +5 (51-46) FW: Sept 12-16
Montana: Sen Jon Tester (D) vs Tim Sheehy (R)
RMG Research: Sheehy +7 (50-43) FW: Sept 12-19
RMG Research: Tester +5 (49-44) FW: Aug 6-14
There have been no other polls in this state by 2 and 3-star pollsters. RMG’s polling represents a 12 point change in just over a month. Republicans have been reported to believe this race is over and Tester will lose. I do not know what anyone is seeing in the internal polling, but there is not enough publicly-available evidence to conclude Tester definitely will not win.
Nevada: Sen. Jackie Rosen (D) vs Sam Brown (R)
Emerson College: Rosen +7 (48-41) FW: Sept 15-18
Morning Consult: Rosen +13 (52-39) FW: Sept 9-18
Morning Consult: Rosen +10 (50-40) FW: Aug 30-Sept 8
There are a few other low-quality pollsters that also find Rosen up by eight points or more.
Ohio: Sen. Sherrod Brown vs Bernie Moreno (R)
Morning Consult: Brown +2 (46-44) FW: Sept 9-18
Morning Consult: Brown +3 (46-43) FW: Aug 30-Sept 8
Emerson College: Brown +1 (46-44) FW: Sept 3-5
Republicans are also talking as if Brown has already lost the state, but the public polling contradicts this conclusion. Suspect pollster Activote is the only one that has Moreno with a lead. Even a poll commission for Republicans found Brown leading earlier this month.
Pennsylvania: Sen. Bob Casey vs David McCormick
Emerson College: Casey +4 (47-42) FW: Sept 15-18
MassINC Polling Group: Casey +7 (49-42) FW: Sept 12-18
Morning Consult: Casey +8 (49-40) FW: Sept 9-18
Marist College: Casey +6 (52-46) FW: Sept 12-17
Washington Post: Even (48-48) FW: Sept 12-16
Quinnipiac University: Casey +9 (52-43) FW: 12-16
New York Times/Siena College: Casey +9 (49-40) FW: Sept 11-16
Insider Advantage: Casey +5 (49-44) FW: Sept 14-15
Franklin & Marshall College: Casey +8 (48-40) FW: Sept 4-15
Group 2 Races: Where conventional wisdom is not paying attention
Florida: Sen. Rick Scott (R) vs former Rep. Debbie Mucarsel-Powell (D)
Florida Atlantic University: Scott +4 (47-43) FW: Aug 10-11
Emerson College: Scott +1 (46-45) FW: Sept 3-5
Morning Consult: Scott +5 (47-42) FW: Aug 30-Sept 8
Morning Consult: Scott +4 (46-42) FW: Sept 9-18
Note that several 1-star and unrated pollsters have fielded polls with similar or, in one case (Activote, an unrated pollster, which recently said that Trump was beating Harris nationally 51-50; let that sink in for a moment), having Scott with an eight point lead.
Maryland: Prince George’s County Executive Angela Alsobrooks (D) vs former Governor Larry Hogan (R)
Morning Consult: Alsobrooks +12 (50-39) FW: Sept 9-18
Emerson College: Alsobrooks +7 (49-42) FW: Sept 12-13
Morning Consult: Alsobrooks +6 (48-43) FW: Aug 30-Sept 8
Republicans have recently announced a big spend in Maryland to try to pick up this seat. Since the primary in May, there has only been one poll that did not show Alsobrooks in the lead. The race was even and the pollster was Trump’s pollster (who is no dummy but is less than 2-star pollster and there may be a house effect in this poll).
Nebraska: Sen Deb Fischer (R) vs Labor union president Dan Osborn (I)
SurveyUSA: Fischer +1 (39-38) FW: Aug 23-27
YouGov: Fischer +2 (43-41) FW: July 31-Sept 12
There was another poll that had Fischer +1 at the same time as the SurveyUSA poll, but it is a less than 2-star pollster. I mention it because there is so little polling in this race.
Texas: Sen. Ted Cruz (R) vs Rep. Colin Allred (D)
Texas Hispanic Policy Foundation: Cruz +3 (48-45) FW: Sept 13-18
Morning Consult: Allred +1 (45-44) FW: Sept 9-18
Morning Consult: Cruz +5 (47-42) FW: Aug 30-Sept 8
Emerson College: Cruz +4 (48-44) FW: Sept 3-5
Note that Texas Hispanic Policy Foundation is an unrated pollster, but it is one of those regional pollsters that have a very good reputation within their own state, so I am including it here.
That’s a lot of polling data to digest, but you can see that there are three Republican incumbents who are in at least as much trouble as Gallego and and Slotkin are. In fact, the polling suggests that Democrats are in much better shape in these races than the media narratives out there would have you believe.
There is a very real chance that Scott will lose in Florida and a lesser but still real chance Cruz could lose in Texas. Also, there is also a real chance that Osborn will lose to the Independent in Nebraska. Democrats wisely chose to cede the race to Osborn, who has real appeal to many working-class Nebraskans and does have progressive positions on some issue. There are a lot of undecided voters in the Nebraska polling, so the race could break Fischer’s way on election day. In fact, I would expect that to happen if they are still undecided in November. But, at this moment, Fischer is in trouble. If Osborn can convince undecideds in the next few weeks, it will be a close race at the bell.
Instead of focusing on these three races, when thinking "out of the box," the media tends to look at Hogan's chances to win in Maryland. It seems almost certain that he will lose, but Democrats do have to contest that one - especially in light of Republican plans to heavily outspend Alsobrooks there over the next few weeks. What the media is missing, however, is that the polling suggests it is much likelier the Democrat wins in Florida than Hogan wins in Maryland.
While it may be true that Tester will lose, Brown is another matter. Yes, the Ohio race is close, and these are not comfortable numbers for an incumbent Senator to have at this point in the election.3 However, the same can be said of Cruz, Fischer, and Scott. If Sen. Elizabeth Warren (D-MA) or Adam Schiff (D-CA) were only ahead by single digits - let alone only by 3-5 points - it would be a huge story. So, why not focus more attention on the Texas, Nebraska, and Florida Senate races? Perhaps that will change soon. Or maybe, instead of continuing to lose ground like Scott and Cruz have, they will start to pull away in the polling. We will know soon.
Follow-up to yesterday’s post: Question 2 might not be in such a bad position after all
MassINC has a poll out that shows Question 2 (the MCAS question) winning 51-34 with 15% undecided. Undoubtedly a better poll for the MTA than the UNH poll. But it is not as good a poll for the measure as it looks. If you read yesterday’s post you will know why. Question 2 is a change in the status quo. As a result, we would expect most undecideds to break to the “no” position and for some of the “yes” votes to change their position when it comes time to vote. There is no guarantee this will happen, and campaigns do matter. If this result is more representative of how voters are thinking in Massachusetts than the UNH poll, then the MTA has a much better shot at getting to 50% on election day. However, it is important to remember that two years ago, the Fair Share Amendment (Question 1) was polling at 60% yes in late October. The measure passed with just 52% voting for it. The odds are stacked against Question 2, but this poll will give some hope to its proponents after the very bad UNH finding.
“Question 2 is ‘the one possible exception’ to the common slide from yes to no, [MassINC polling director Rich] Parr said. ‘Parents may hear messages in favor of that question from and around their children’s schools, and that could help to grow the ‘yes’ number between now and November.’”
The above-quote from Commonwealth Beacon’s report on the MassINC poll is a more nuanced way of saying campaigns matter, but there is another possible reason for the differing results from UNH and MassINC: question design. Generally-speaking, it is better to read the question (or a fair summary – they are often too wordy), make a statement that this measure will change the law, and state that a no vote will result in no change to the law. MassINC did the better job with question design – another reason for MTA to hope. Here are the questions asked as reported in their polling memos:
MassINC: “There will be a question on the November 2024 ballot in Massachusetts regarding standardized testing. A yes vote would eliminate the requirement that students pass the Massachusetts Comprehensive Assessment System (MCAS) in order to graduate high school but still require students to complete coursework that meets state standards. A no vote would make no change to existing law. If the election were held today, how would you vote on this proposal?”
UNH: “Question 2: [will] Repeal the requirement that students must achieve a certain competency level on the Massachusetts Comprehensive Assessment System (MCAS) exam to graduate high school. If the elecon were held today, how would you vote?”
I am surprised at the way UNH worded the question. It used framing consistent with the campaign against Question 2. In essence, that campaign says that removing the MCAS will undermine the competency of student achievement. MassINC did a better job by stating that the measure will require removing the requirement to “pass” the exam. While UNH’s language might be technically correct, it likely suggests something very different to the survey respondent. The term “competency” suggests that the removal of the exam will result in lower learning competency among students. MassINC’s wording is more descriptive: students won’t have to pass the exam but will still have to complete coursework that meets state standards. That is accurate and it does not inject interpretation of how this change might impact learning. (The real debate here among the proponents and opponents is over measures of accountability for teachers and school districts, not whether students themselves need this test.)
Here is what is actually on the ballot:
QUESTION 2:
Law Proposed by Initiative Petition
Elimination of MCAS as High School Graduation Requirement
Do you approve of a law summarized below, on which no vote was taken by the Senate or the House of Representatives before May 1, 2024?
SUMMARY
As required by law, summaries are written by the State Attorney General.
This proposed law would eliminate the requirement that a student pass the Massachusetts Comprehensive Assessment System (MCAS) tests (or other statewide or district-wide assessments) in mathematics, science and technology, and English in order to receive a high school diploma. Instead, in order for a student to receive a high school diploma, the proposed law would require the student to complete coursework certified by the student’s district as demonstrating mastery of the competencies contained in the state academic standards in mathematics, science and technology, and English, as well as any additional areas determined by the Board of Elementary and Secondary Education.
WHAT YOUR VOTE WILL DO
As required by law, the statements describing the effect of a “yes” or “no” vote are written jointly by the State Attorney General and the Secretary of the Commonwealth.
A YES VOTE would eliminate the requirement that students pass the Massachusetts Comprehensive Assessment System (MCAS) in order to graduate high school but still require students to complete coursework that meets state standards.
A NO VOTE would make no change in the law relative to the requirement that a student pass the MCAS in order to graduate high school.
As you can see, MassINC took its question design directly from the language that will be on the ballot. This is a much better practice than the design UNH used. If I were advising the MTA on these public polls alone (they should have internal polling), I would tell them they are in striking distance of winning. They should remain focused on framing the issue as the ballot has. But I would remind them of 2022, when their last ballot measure lost eight points in the week or two between the final public poll and the election.4 And that was not something that was a function of poor organizing or overconfidence. The MTA and the Fair Share Coalition ran a strong, well-funded campaign to win in 2022. Their margin for error this year remains razor thin, but after comparing these two polls I know longer think Question 2 is a near-certain loser.
538 has moved from a letter grade rating system to one in which pollsters are rated from 0.0 (worst) to 3.0 (best).
FW is shorthand for “field work.” This is the time period that the survey was “in the field,” or being administered to respondents.
It’s not a magic number, but in normal years it is usually not a good sign for an incumbent Senator to be polling below 50% in October.
There was internal polling that showed the race closer in the weeks leading up to the 2022 election, but it was not made available to the public. I have seen that internal polling.