August 6, 2026
Polls in the Michigan Senate race between Abdul El-Sayed and Haley Stevens were wrong by 10 to 15 points. Not all of them, and not during the entire race. But in the final weeks leading up to the August 4 primary, they were off by 10% or more on average. They showed Abdul El-Sayed winning by a substantial margin. Late polling suggested a rout in favor of El-Sayed—a prediction that might have affected turnout. The prediction markets (Polymarket and Kalshi) gave El-Sayed a 98% chance of winning. But instead of the predicted route, El-Sayed won by 14,893 votes in a race with 1.53 million votes cast by Democrats, an outcome that the press has dubbed as a “nail-biter.” See details of the final polling in Newsweek, Abdul El-Sayed Chances Against Haley Stevens: Final Michigan Senate Primary Polls.
The chart below presents polling data from the Newsweek article linked above. (Note: I used Claude AI to prepare this chart from the Newsweek data.) The polling “miss” is in the far-right column.
The why of the polling miss is less important than the fact that it happened. It’s a reminder that polls are not elections; they are imperfect attempts to measure voter sentiment at a point in time. Polls do not determine outcomes; voters do. Or rather, voters who show up to vote determine outcomes.
Today, I have heard three commentators in the progressive media say, “There is no path to victory in the Senate for Democrats if they do not win Michigan.” That is an unmitigated, unadulterated, steaming-hot pile of nonsense. (And you know what I mean.) Six months ago, those same commentators were saying, “There is no path to victory in the Senate for Democrats. Period. Full stop.”
As a regular public service announcement, I remind readers that the length of time between now and the 2026 midterms, measured in Trump years, is equal to infinity.
Making predictions based on prior elections is hazardous where the relevant comparisons do not include a president who (1) started an enormously unpopular war eight months before midterms, a war that jacked up gas prices (2) imposed tariffs on consumer goods while cutting subsidies for Medicare prescription drugs, (3) demolished one-third of the White House for a gigantic ballroom no one wanted, and (4) reported billions in crypto revenue during his first year in office, largely from the sale of his self-branded cryptocurrency that made him rich at the expense of nearly every investor except his close golfing buddies. Modeling that scenario is rough!
The point is not to beat up pollsters; they do a fine job of that themselves. The point is to say that we should neither over-rely nor under-rely on polls. They are tools for professionals to use in guiding campaign strategy, not for calibrating our level of effort or expectations. We are in the fight of our lives, and polls can mislead us by providing a false sense of security or predicting disaster before it happens.
What should shake our confidence in polls is that there were obvious signs that the number, nature, and quality of the polls may have been overstating El-Sayed’s support. Poll aggregators presumably know how to correct for the influence of poorly designed or poorly executed polls. In the Michigan Senate race, those adjustments were not made, for reasons that are not clear.
In defense of pollsters, they do not “predict” outcomes, and their results are often mischaracterized as single-point results. That’s not the case. Polls have a margin of error, which describes the range of likely outcomes at a specific confidence level (95%). But polls have “tails” that extend beyond the margin of error, and those tails admit the possibility that highly unlikely outcomes might occur. Although such outliers are rare, they are inherent in the statistical models that underlie opinion polls. Pollsters therefore design their polls to minimize the likelihood of outliers.
Which brings us back to Michigan. The pollsters “missed” something in their models that is roiling the electorate. I don’t know what caused the “miss,” but Simon Rosenberg of Hopium Chronicles and Josh Marshall of Talking Points Memo discuss the polling miss in more detail. But (as usual), it is Josh Marshall who hits the nail on the head:
It’s a good reminder that whether you’re a volunteer, a voter, an activist, or a campaign worker, you play these things out till the very end.
Coda: While many readers reacted positively to yesterday’s newsletter rebutting the notion that “Democrats are in disarray,” I did receive some sharp criticism from readers who said I glossed over the issue of Israel and Gaza, which played an important role in motivating some voters to vote for either Abdul El-Sayed or Haley Stevens.
Jennifer Rubin addressed the issue directly in her Contrarian column on Wednesday, Michigan Nail-Biter. I quote three paragraphs below, but I recommend her entire column to your attention. It is an insightful, measured analysis. She writes, in part:
Though legacy media and the MAGA-friendly punditocracy will insist the outcome of the race is evidence of — you guessed it! — “Democrats in disarray,” this is how primaries work. They test the relative strengths of candidates in a tough battle and begin the process of forging consensus for the general election. The only issue for Democrats now is to hold the seat. [¶] While pundits and partisan axe-grinders told us for weeks the race turned on an ideological schism in the party, ordinary voters, even in primaries that attract the most partisan segment of the electorate, choose between real candidates, not partisan manifestos. And what motivates them is not necessarily where they precisely stand on the left-to-right spectrum. Savvy political analyst Karen Finney observed: "National coverage of Michigan has turned this primary into a proxy war over Israel and Gaza while voters are drowning in medical debt, premiums, and everyday costs. Ignoring the full dynamics erases the stakes for voters and obscures what people across MI & the country are saying about their anger over costs and daily struggle."
While we must await the post-election analysis to determine which issues were responsible for Abdul El-Sayed’s narrow victory and Haley Stevens strong showing, both candidates urged party unity in their statements after the election was called in favor of El-Sayed. That is the first step in the process of creating unity in defeating Trump’s hand-picked candidate, GOP Rep. Mike Rogers
Update on Todd Blanche’s nomination as Attorney General.
As I noted yesterday, given Susan Collins’s statement that she will vote against confirmation of Todd Blanche, he can afford to lose only one more Republican vote in the Senate and still be confirmed. If two more Republican Senators join Collins in voting against him (for a total of 3 Republian “no” votes), he will not be confirmed.
On Wednesday, Blanche met with the two most likely no-vote candidates: Lisa Murkowski (R-AK) and Bill Cassidy (R-LA). Trump opposed Murkowski’s 2022 Senate re-election bid and supported Bill Cassidy’s GOP primary opponent, who is now the GOP nominee to take Cassidy’s seat in November. The meeting between Blanche, Murkowski, and Cassidy was fittingly held in “a basement hideaway room” in the Capitol. See NBC News, Sens. Murkowski and Cassidy meet with Blanche as attorney general vote remains uncertain
We should never count on Republicans to do the right thing, but each has strong reasons to vote against Blanche. Senator Cassidy, a physician, has been highly critical of Robert Kennedy’s conduct after Cassidy put him over the top. Lisa Murkowski opposes Trump’s policies on a number of issues, including women’s reproductive rights. Blanche was quoted yesterday as promising Christian groups that he would endeavor to ensure the “Dobbs decision becomes permanent in every single state.”
And, in case you are wondering, Senator Fetterman will oppose Blanche’s confirmation. See The Hill, John Fetterman will vote against Todd Blanche for Attorney General
GOP representative drops out of reelection bid after censure by House.
Rep. Chuck Edwards says he will end his reelection bid after censure recommendation by the House Ethics Committee. The censure recommendation comes after the committee found that Edwards engaged in “inappropriate conduct” with two female staffers. See Politico, Chuck Edwards says he will end reelection bid after censure recommendation.
Republicans in North Carolina have until Monday to replace Edwards on the ballot for the November election. The Democratic candidate is Jamie Ager, a fifth-generation North Carolinian farmer. Check out Jamie’s video message here: Jamie Ager for US Congress.
Iran and Oman are close to a deal to divide control of the Strait of Hormuz.
It is exceedingly odd that the conclusion of the US war against Iran is being hammered out between Iran and Oman. Those countries are close to an agreement to divide control of the Strait of Hormuz between them, something that would have been unthinkable before Trump started his ill-conceived war against Iran. See NYTimes, Iran Says It Is Nearing Final Agreement With Oman on Strait of Hormuz. (Gift link.)
Trump has moved on from the war and has begun the “who do I blame for my failure” after-action review. See WaPo, Trump, Hegseth clashed at Camp David over Iran missile depletion concerns. (Gift link.)
Per WaPo,
President Donald Trump’s frustration over the Iran war boiled over at Camp David last week, where he demanded answers from Defense Secretary Pete Hegseth on why he had apparently been misled on extreme munitions shortages that now threaten to limit military options with Iran, two people familiar with the exchange told The Washington Post.
Per the Post, Hegseth blamed Deputy Secretary of Defense Stephen Feinberg for the shortages and for failing to inform the president about the diminished stockpiles. Hegseth’s deflection of the blame is pathetic. It was Hegseth who urged Trump to initiate the attacks that depleted the stockpiles. As for informing Trump, the rapid depletion has been front page news for six months.
In a suprising reversal, FDA approves an mRNA flu vaccine
It is part of Robert Kennedy’s conspiracy theories about Covid that mRNA vaccines were ineffective and /or dangerous. Unsprisingly, the FDA initially balked when Moderna sought approval for an mRNA flu vaccine. But the public backlash against the FDA’s initial stance was so strong that the agency reversed itself and approved the vaccine. See NYTimes, F.D.A. Approves Moderna’s mRNA Flu Vaccine. (Gift link.)
Per the Times,
Initially the F.D.A. had refused to even review Moderna’s application for approval earlier this year, but quickly reversed course after a public outcry. The flu kills more than 300,000 people worldwide each year and was particularly harsh in the 2024-25 season, with an estimated 45,000 deaths in the United States . . . .
Let’s hear it for “public backlash.'“ Your voice matters!
Concluding Thoughts
The Michigan race for the U.S. Senate is dominating political news on Wednesday, so I will take the opportunity to give all of us a break from the firehose of commentary. A few other stories are worth noting, but they can keep for another day.
There are many ways to interpret the spate of incumbent House Democrats losing to more liberal primary challengers. From where I stand, it appears that the Democratic electorate wants representatives who will be more aggressive in taking on Trump on everything—the economy, the war, ICE, corruption and self-dealing, the climate crisis, and protection of individual rights against government overreach.
Oh, and there is a definite sense that the old guard in the Democratic Party should focus on elevating the next generation of leaders—instead of hanging on to power by their fingernails.
Generational change is painful. But it is also a healthy signal that the next generation is impatient to begin positioning the Democratic Party for the future. Truth be told, it is always so. If it were otherwise, the future of the Democratic Party would be bleak. So, instead of being anxious that a half-dozen Democratic incumbents have lost their primaries, let’s ask ourselves what message the Democratic electorate is sending to its party leaders
