Baala Shakya, Photography Editor

When New York state Assemblymember Zohran Mamdani stunned New York City’s political establishment by emerging as the likely first-round winner of the city’s Democratic mayoral primary last week, his apparent victory defied a Yale poll result that had former governor Andrew Cuomo comfortably ahead.

The poll, conducted from June 17 to 22 in partnership with data analytics firm YouGov, projected Cuomo leading Mamdani 57 to 43 percent in final-round ranked-choice tabulations. But on primary night, the democratic socialist assemblymember claimed 44 percent of the vote in the first round, while Cuomo trailed at 36 percent and conceded the race that night.

The outcome has prompted a process of self-reflection for the Yale Youth Poll, a student-run initiative housed in Yale’s Institution for Social and Policy Studies that operates the Yale Youth Poll, separate from its New York mayoral primary survey. The group’s leaders told the News after this article’s original publication that the organization is in the process of being renamed Yale Polling.

“Obviously, our poll was off,” Milan Singh ’26, the director of the Yale Youth Poll and the News’ former opinions editor, told the News on Thursday. “Primary polling can be tricky, and we’re working on putting together a postmortem report that will discuss why our poll was off for this race and how we can avoid those mistakes in the future.”

The prediction that missed

The poll studied the preferences of 460 self-identified registered Democrats across New York City using an online survey. Of those, 371 reported having already voted or intending to vote in the Democratic primary. 

The poll simulated how candidates would fare through ten rounds of counting in New York’s ranked choice voting system, eliminating the lowest-ranked contenders in successive rounds until one claimed a majority of ballots. YouGov, a data analytics company, tabulated the ranked-choice results for the Yale Youth Poll.

It was one of the 32 polls tracked by the New York Times, only one of which — by Public Policy Polling — predicted that Mamdani would be leading after the first round. A poll from Emerson College also had Mamdani prevailing, but only after ranked choice ballots were redistributed.

Singh noted that the Yale Youth Poll accurately predicted Cuomo’s base of support. The poll forecasted 38 percent for him in the first round, just two points off the final tally. Ranked-choice tallies for the final round of votes, which determines the official winner of the mayoral Democratic primary, are expected on Tuesday, July 1.

But Mamdani’s rise caught the Yale pollsters off guard. The poll had him at 28 percent in the first round, 16 points below his actual showing.

The Yale poll overestimated seven other candidates. It projected New York City Council Speaker Adrienne Adams at 12 percent and former comptroller Scott Stringer at 6 percent. They received 4 percent and 2 percent of the reported first-round vote, respectively.

Youth Poll addresses error 

Shortly after the poll’s release on the evening of Monday, June 23 — the night before the primary election — YouGov clarified that it had made an error in the ranked choice voting simulation, mistakenly eliminating one candidate, Stringer, based on unweighted results instead of the weighted representative sample.

“While weighting adjustments are small, in this case, it was sufficient to erroneously drop the wrong candidate between rounds 7 and 8,” YouGov wrote in a statement to the News. “The first and final rankings in our RCV tally were not changed. We are investigating our remaining data processing closely, but are confident no other results in this survey, nor the underlying data, were impacted.”

Jack Dozier ’27, deputy director of the Yale Youth Poll, Zachary Donnini ’25 and Singh confirmed that the mistake made by YouGov did not affect the outcome, which still showed Cuomo ahead of Mamdani in the final round.

Dozier noted that YouGov uses “the most modern technology in surveying voters” and the YouGov sample remains “fairly representative of the New York Democrats’ population.”

Still, the miscalculation pointed to how delicate ranked-choice simulations can be in close, multi-candidate contests, especially when modeling a volatile and shifting electorate.

“Every piece of evidence, every finding, every top line and every conclusion that comes out of any poll is an opportunity to learn,” Dozier said.

Cuomo’s confidence, miscalibrated 

In the final days of the campaign, Cuomo’s team leaned heavily on polling that showed him in command. The Yale Youth Poll’s projections reinforced the impression that the former governor remained a dominant force in city politics, capable of winning in a fractured, ideologically diverse field.

On the campaign trail, Cuomo dismissed suggestions that his lead might be in jeopardy. In a video posted on X by a New York Daily News reporter, Cuomo showed confidence in polls favoring his victory, downplaying the Emerson College poll, which predicted Mamdani would prevail, as an “outlier.” He mentioned seeing the latest poll added to the New York Times’ list: “a poll from something called YouGov from Yale University.”

When the reporter noted that the Yale poll “apparently had errors in it,” Cuomo brushed it off as a campaign staffer confirmed the outcome remained “exactly the same.”

But already, in a Substack analysis accompanying the poll’s release, the pollsters had warned that their projections might not hold.

They wrote that “the demographic makeup of the actual electorate can differ significantly from polls due to differential turnout rates.”

Data from the poll showed Cuomo leading among older voters, non-college-educated voters, and voters of color. In an interview before the primary election, Donnini noted that “early voting has thus far reflected a disproportionately white and younger electorate compared to historical norms in New York City” — demographics likely to favor Mamdani.

Donnini added that the majority of votes were still expected on Election Day.

“If turnout among Cuomo’s base does not pick up on Tuesday,” the pollsters wrote in their Substack post, “he could massively underperform our poll.”

Singh echoed that in an interview after the election, noting that “some of the voters Cuomo was counting on are no longer in the Democratic coalition.” 

Singh said that some New Yorkers who backed Joe Biden in 2020 but Donald Trump in 2024 may have drifted away from Democratic primaries altogether.

“Democrats have been getting steadily more liberal for decades now,” Singh added. “We will need more time and data to confirm or refute this hypothesis, but it’s one possible explanation.”

The 2025 New York City mayoral general election is on Nov. 4.

Correction, June 29: An earlier version of this article misstated the number of registered Democrats who were surveyed in the Yale poll; it was 460, not 647. The article also previously misstated the number of candidates whom the poll overestimated; it was seven, not two. The article has been updated to clarify that Yale Youth Poll is the publicly identified name not only of the undergraduate-run initiative that managed the New York mayoral primary poll, but also of a separate poll managed by the group, and that the group is being renamed Yale Polling.