The Socialist Bradley Effect

Something strange just happened in two Midwestern Democratic primaries and don’t just “blame it on the polling”. It is something else that, for whatever reason, people are scared to talk about.

In back-to-back Democratic primaries in Michigan and Wisconsin, candidates identified with the Democratic Socialists of America looked far stronger in surveys than they proved to be when voters actually cast ballots. Is that the pollster’s fault…or are voters just not comfortable saying what they REALLY think? Am I talking about race? Not really.

The Bradley Effect is a term in politics that came from the 1982 California Governor’s race. The Los Angeles Mayor, an African American, was the Democratic nominee for Governor. He was leading Republican George Deukmejian, an Armenian American, by almost 6-7% in all surveys. When election day finally did come Deukmejian bested Bradley by barely 1%. Most objective political observers were able to clearly see that what voters said in a survey was not exactly how they acted when filling out their ballots. It left many with the conclusion that voters did not want to sound racist during a survey…all the while still displaying a hesitancy about actually voting for an African American candidate.

So when we look at Abdul El-Sayed in Michigan and Francesca Hong, they had much closer races than expected. Therefore, it must be racism, right? Not exactly.

Michigan came first. Just days before the Democratic primary for U.S. Senate, an Emerson College poll showed progressive Abdul El-Sayed leading Congresswoman Haley Stevens 54 percent to 39 percent. When undecided voters were pushed to choose, his lead grew to 57–41.

El-Sayed did win. But instead of the double-digit landslide the polling suggested, he scraped by 48.5 percent to 47.5 percent - a one-point victory. The final polling average had him ahead by roughly ten.

One week later, Wisconsin produced an even more dramatic miss. A late-July survey showed Democratic socialist Francesca Hong at 44 percent in the race for governor, with Milwaukee County Executive David Crowley at just 15 percent. Even Crowley’s own later polling reportedly had him trailing by nearly 20 points.

On Election Day, Crowley won. With nearly all votes counted, he led Hong 39.8 percent to 39.4 percent.

To be precise, Hong is a member of the Democratic Socialists of America and was endorsed by Wisconsin DSA chapters. El-Sayed does not call himself a democratic socialist and was not endorsed by the DSA. But his campaign was embraced by many of the same activists, leaders and organizations. Both candidates occupied the same political lane: insurgent, anti-establishment and firmly on the party’s progressive left.

Put the two races side by side and it must be racism, right? Maybe Wisconsin, which is 85% white, has hesitation about voting for “Hong”. Maybe Michigan, which is 77% white, can’t get behind an “El-Sayed”.

Chalking these up to “racism” is just intellectually lazy. Is there racism in America still? Of course. But if you think it’s a “white people thing” you live in a sadly small and insulated world. America in 2026 is far more racially, religiously and ethnically complex than it was in 1982.

Obviously, race and socialism are not comparable things, and two primaries do not make a scientific law. The useful comparison is narrower: social-desirability bias.

Among Democratic primary voters, supporting the young, diverse, movement-backed progressive can feel like the approved answer. Saying you prefer the more conventional candidate may feel uninspiring, establishment-minded or even disloyal to the party’s activist base. A voter can admire the progressive candidate, tell a pollster that is where he or she is leaning, and then confront a different question on Election Day: Who is most likely to win in November? In 2004, this was summed up nicely with a bumper sticker popular with general election Democrats: “Dated Dean but married Kerry”. Electoral pragmatism.

That final calculation mattered in Wisconsin. Hong’s earlier calls to defund the police and abolish prisons, along with old social-media posts about “canceling Thanksgiving,” dominated the closing days. Crowley made electability the centerpiece of his campaign, and Democratic Governor Tony Evers reinforced that message with a late endorsement. Voters who liked Hong’s economic populism may still have decided that nominating her in one of the closest swing states in America was too great a risk.

Of course, there are more ordinary explanations for both misses. Both Michigan and Wisconsin allow broad participation in party primaries, making the electorate difficult to model. Younger voters and activists may be easier to reach through the text-to-web and online methods used in modern surveys but less reliable about actually voting. Wisconsin’s field changed dramatically after the late-July poll, giving anti-Hong voters an opportunity to consolidate behind Crowley. Late-breaking voters may simply have moved en masse.

That is why I would not claim voters were simply lying to pollsters. It could be a combination of social pressure, soft support, turnout error and late strategic voting. But from a campaign’s perspective, the practical warning is the same: publicly expressed enthusiasm is not always hardened electoral support.

The more immediate warning is for the progressive left: cultural energy is not the same thing as electoral depth. Socialists can dominate online conversation, generate enthusiastic volunteers and fill rallies. Those advantages are real. They can also create an echo chamber in which reporters, activists and even pollsters mistake intensity for breadth.

Nor should Republicans conclude that progressive candidates cannot win. El-Sayed did win Michigan, and other progressives have won convincingly elsewhere. The more useful lesson is narrower: in competitive Midwestern states, polling may be overstating the strength - and especially the margin - of candidates associated with the socialist left.

We need more elections before giving the phenomenon a permanent name. But after Michigan and Wisconsin, strategists should at least be watching for it.

Call it the Socialist Bradley Effect: the gap between the progressive candidate voters feel good telling a pollster they support and the candidate they ultimately trust with their vote.

Chris Faulkner, a United States Marine Corps veteran (1991–2001), serves as a Senior Advisor at Connector, Inc. where he leans on nearly three decades of winning campaigns to advise our clients on their political efforts and goals. He and his wife, Angela, live outside Knoxville, Tennessee with their poodle and pit bull, and are proud parents of three adult sons.