Three Polls Show Hochul Leading. They Cannot Agree on Much Else.

Politics

Oct 30

By Callahan Blackwood | October 30, 2026

Three October polls of the New York governor's race have been released in the past ten days. All three show Governor Kathy Hochul leading Nassau County Executive Bruce Blakeman by double digits. All three were conducted among likely voters, or something close to it. All three used professional shops with long New York track records. On those bare facts, the race looks settled.

Look closer, and the numbers start arguing with each other.

The disagreements are not rounding errors. They are methodological conflicts sharp enough that the two campaigns are treating them as entirely different elections. Understanding what the polls actually say, what they cannot say, and where they flatly contradict each other is the only honest way to read the closing stretch of this race.

What each poll found

Siena Research Institute released its results on October 28, after fielding from October 19 through 22 among 807 New York likely voters. The poll is directed by Donald Levy and was powered by ReconMR; the sample was drawn from the L2 Voter Database and weighted on age, region, race and ethnicity, education, party identification, vote likelihood, and gender. The margin of error is plus or minus 4.1 points, including design effects. Siena gives Hochul 54 percent to Blakeman's 40 percent, a 14-point lead. That is Siena's widest Hochul margin of the cycle, up from 50 to 41 in September and 49 to 39 in August.

Quinnipiac University posted its numbers a week earlier, on October 21, from a field period of October 15 through 19 among 1,167 likely voters. It used random-digit-dial with live interviewers on both landlines and cell phones; the margin of error is plus or minus 3.8 points, including the design effect. Quinnipiac has Hochul at 55 percent and Blakeman at 41 percent, again a 14-point lead. That is a narrowing from Quinnipiac's September poll, which had Hochul ahead 58 to 39, a 19-point spread. The race tightened in Quinnipiac's series even as it widened in Siena's.

Marist Poll, also released October 28, conducted its survey from October 22 through 25 among 1,448 New York registered voters using live phone interviewers, text, and online modes. Marist's own methodology states that 1,448 registered voters were surveyed; the 53-to-42 result it reports is drawn from the likely-voter subset within that sample, at a margin of error of plus or minus 3.9 points for likely voters. The polling aggregator 270toWin shorthandedly lists the Marist sample as 1,448 likely voters, which is the aggregator's convention, not Marist's characterization. Marist puts the race at Hochul 53, Blakeman 42, with 4 percent undecided. Among the broader registered-voter electorate, Marist reports Hochul's lead at 14 points, up from 11 points earlier in October.

Three polls, field dates spanning October 15 through October 25, producing Hochul leads of 11 to 14 points. On the topline numbers, the agreement is as solid as polling gets.

On the subgroups, it dissolves.

The independents problem

The most striking disagreement in the October polling involves independent voters, and the gap is not a small one.

Siena's own release states that Hochul leads Blakeman among independents by 11 points. Pollster Steven Greenberg laid out the structural argument in Siena's language: Republicans winning statewide in New York must run up big margins upstate and in the downstate suburbs and carry independents by a wide margin to offset Democrats' two-to-one enrollment advantage. Siena's October numbers show Blakeman running roughly even with Hochul upstate and in the suburbs. He is losing independents by 11. That combination, in Greenberg's framing, means the structural path to victory is not there.

Now look at what Marist found.

According to reporting by lohud and the New York Post from the Marist poll, Blakeman leads among independent and blank-enrollment likely voters by 10 points, 49 percent to 39 percent. That figure does not appear in Marist's own release summary; it comes from lohud's and the Post's reporting of the Marist data, not from Marist's own release. Both outlets report the figure consistently, and Blakeman's campaign pointed to it publicly.

Quinnipiac's own release has Blakeman ahead of Hochul among independents by 7 points, 51 percent to 44 percent.

Set those three readings beside each other. Siena has Hochul winning independents by 11. Marist, as reported by lohud and the Post, has Blakeman winning them by 10. Quinnipiac has Blakeman winning them by 7. The range from one end to the other is 21 percentage points, across polls conducted within ten days of each other, measuring the same category of voter in the same state.

That is not normal polling variation. It is a divergence large enough to reflect genuine methodological differences in how each pollster defines and screens for independence, how they weight their samples, and what their likely-voter model selects. Siena uses a dual-frame design drawing from the L2 Voter Database, weighting on party identification among other variables. Marist uses multi-mode outreach with live phone, text, and digital. Quinnipiac uses random-digit-dial with live interviewers. The independent electorates they are each reaching may not be the same group of people, even if all three call them independents.

None of the three public releases explains in enough detail exactly how their independent-voter categories are constructed to resolve this from the outside. What can be said is that Siena is the outlier of the three, and that both polls placing Blakeman ahead among independents still have him losing the race statewide by double digits. You can build a plausible argument for Blakeman from two of the three independent-voter readings. You cannot build a lead out of them.

Siena's figures also tell something about which parts of the coalition matter. Greenberg's release notes that Hochul leads women by 23 points and men by 6, leads Latinos by better than two to one, and leads Black voters 81 to 16. A 35 percent plurality of voters, including 45 percent of independents, say neither candidate would make life more affordable. On affordability specifically, 51 percent of Democrats say Hochul would help; 60 percent of Republicans say Blakeman would; independents again split the difference and then some.

Margins of error and the likely-voter screen

Each of the three polls states its margin of error precisely, and all three define and construct their likely-voter samples in different ways.

Siena's stated margin of error is plus or minus 4.1 points, including design effects, on a sample of 807 likely voters drawn from the L2 Voter Database and weighted on vote likelihood among other variables. Quinnipiac's margin is plus or minus 3.8 points including design effect, on 1,167 likely voters reached by random-digit-dial with live interviewers. Marist surveyed 1,448 registered voters across multiple modes; its likely-voter results are drawn from within that sample as a subset, at a stated margin of plus or minus 3.9 points. Those stated margins apply to the topline statewide numbers. The independent-voter crosstabs are drawn from smaller portions of each sample, and the three polls reached their independent-voter respondents through different sampling frames and different likely-voter methods.

The likely-voter screen compounds the difficulty, because all three polls use one and all three build it differently. Siena weights on vote likelihood as a variable. Marist reports likely-voter results as a subset of its full registered-voter sample at a separate margin of error. Quinnipiac applies its design effect to a random-digit-dial sample. These are not equivalent operations. In a race where the electorate's composition depends heavily on whether Republican enthusiasm can chip into a large Democratic registration advantage, the likely-voter model is doing work that the published methodology descriptions do not fully expose.

The independent-voter divergence across these three polls may partly reflect three different likely-voter screens constructing three different independent electorates rather than one electorate measured three ways. That is a possibility the public releases cannot rule out. None of this is a criticism aimed at any single pollster. It is a reason to treat subgroup findings from any single poll with appropriate caution when they conflict sharply with subgroup findings from contemporaneous polls by reputable shops.

Favorability: another conflict

The independent-voter disagreement has a direct parallel in the favorability numbers, and this one is even harder to explain away.

Siena, in its own release, reports Hochul's favorability at 51 percent favorable and 45 percent unfavorable among likely voters. Greenberg's statement in the Siena release calls this the first time in her five years as governor that a majority of New Yorkers view her favorably. Her September favorability in Siena was 48 favorable, 46 unfavorable; she had not broken 50 percent favorable in the Siena series before this poll. Her job approval rose alongside it, to 51 to 46 in October from 48 to 47 in September.

Marist, in the same week, reports Hochul's favorability at 41 percent favorable and 51 percent unfavorable among likely voters. Upside down by 10 points. Marist also notes that 61 percent of independents in its survey hold an unfavorable opinion of Hochul.

These two polls have overlapping field windows. Both measure likely voters, at least for their governor's-race results. They cannot both be accurate about Hochul's net standing in any simple sense. One describes a governor who has, for the first time in her tenure, crossed into majority favorable territory. The other describes a governor whose unfavorable rating leads her favorable by double digits. These are not different shadings of the same picture; they are different pictures.

Quinnipiac's October 21 release puts Hochul at 45 percent favorable and 46 percent unfavorable, with 6 percent who say they have not heard enough. It gives Blakeman 36 percent favorable and 40 percent unfavorable, with 23 percent not having heard enough about him. That last figure is Quinnipiac's headline point about Blakeman: nearly one in four voters were still insufficiently familiar with the Republican nominee to rate him, in a poll conducted ten days before Election Day.

Attribute each favorability reading to its pollster. They cannot be summarized as a single trend.

What the campaigns made of it

Blakeman did not accept the numbers.

On October 28, the day both Siena and Marist published their October results, Blakeman issued a statement reported by lohud, the New York Post, FOX 5, and CBS New York. "Today's polls are nothing more than voter suppression tactics designed to discourage Republicans and manipulate the outcome of this election," he said. He added that Hochul was "desperate to convince New Yorkers this race is over because she knows her failed leadership is on the ballot," and that the "only poll that matters" is on Election Day, and that "New Yorkers deserve a governor who fights for them, not one who relies on misleading polls to declare victory before a single vote is counted." The statement was reported as a campaign release on October 28; the sources opened for this article do not place it at a specific event or podium.

His campaign pollster, John McLaughlin, called Siena's result "a gift to Kathy Hochul at a time when she desperately needs one." Blakeman separately said his side's "polling and internal surveys across the state show a neck-and-neck race." No internal numbers were published in any of the reports opened for this pack.

Blakeman specifically pointed to the gap between Siena's independent-voter numbers and Marist's as evidence that at least one poll was distorted, per lohud's reporting. The gap is real. The distance between Siena's reading of Hochul leading independents by 11 and Marist's reading as reported by lohud and the Post of Blakeman leading independents by 10 is 21 points. Whether that gap reflects methodology, timing, sampling differences, or something else is a question the public releases cannot resolve. Calling it voter suppression is a different and considerably stronger claim, and nothing in the public record supports it.

Hochul called the characterization "outrageous." She told reporters the framing represented a novel approach to reading polls, noting that polling has existed in this country since around 1822, and that she had never heard it previously defined as a mechanism for suppressing the vote. Her campaign spokesperson, Ryan Radulovacki, said the campaign was "full steam ahead to the finish line" with "5 days until New Yorkers reject Blakeman's cost-raising, ICE-defending, Medicaid-cutting agenda."

The money

The October 23 campaign finance filings, as reported by lohud, show a race with no financial equivalence. The figures that follow are lohud's reported account of the filings.

Hochul's campaign raised more than $7.5 million in the October filing period, including $6.5 million from the New York State Democratic Committee. The campaign spent more than $14.2 million in the period. Hochul entered the final stretch with more than $12.3 million cash on hand.

Blakeman started October with less than $200,000. Between the end of September and October 23, his campaign brought in $5.4 million, including more than $2.4 million from the Nassau County Republican Committee. After spending more than $4.3 million, his campaign had $1.27 million cash on hand as of October 23.

The cash-on-hand ratio is roughly ten to one in Hochul's favor, per lohud's account.

Nicki Minaj, briefly

On October 28, rapper Nicki Minaj posted a series of attacks on Hochul on X, calling her "an evil woman" and saying the Democratic Party should forfeit her seat and that she "will not escape God's wrath" for failing to denounce evil when she had the chance. Blakeman responded the same day: "Nicki, I've got the U-Haul ready. It's time to fire the worst governor in America."

Asked about it after a get-out-the-vote rally she headlined in Albany that day, Hochul told reporters Minaj should get outside and be offline more. "Stop going to Trump rallies, it's kind of poisoning your mind," she said.

Lohud notes that Minaj had previously backed Hochul's ticket in 2018, when Hochul ran as Andrew Cuomo's lieutenant governor.

What seven in ten voters expect

Every pollster in October asked voters not whom they preferred but whom they expected to win. The answers cut in the same direction across all three surveys, and they complicate the closing weekend for Blakeman regardless of which independent-voter number you believe.

Siena found that by 68 to 23 percent, voters think Hochul will win, including 40 percent of Republicans. Only 49 percent of Republicans said they expect Blakeman to win. Marist found 71 percent of likely voters, regardless of their own candidate preference, expect Hochul to be re-elected.

Expectation questions measure something different from preference questions, and what they mean for turnout in a closing stretch is genuinely debated. The argument for their relevance: a voter who believes their candidate will lose may conclude their vote is less decisive and stay home. The argument against: partisans vote even in races they expect to lose, and the expectation gap may simply reflect Hochul's advantage in name recognition as an incumbent governor rather than any causal signal about behavior. The data in this pack does not resolve that argument.

The data does show that Blakeman enters the weekend carrying both the expectation gap and thin name recognition. Quinnipiac's October release found nearly one in four voters had not heard enough about him to rate him favorably or unfavorably. Quinnipiac's Assistant Director Mary Snow noted in the release that while Blakeman has improved his image among independents as he has become better known, with less than two weeks remaining he faced an uphill battle against a double-digit deficit.

Blakeman spent October 28 working the Southern Tier, with stops in Owego, Elmira, Hornell, and other communities, joined by Representative Nick Langworthy and several Southern Tier state lawmakers. The Elmira Elks Club stop drew a standing-room-only crowd, per the Elmira Star-Gazette. He called Hochul "the worst governor in America," pressed on energy, criminal justice, and immigration, and said he would sign an executive order on his first day declaring New York no longer a sanctuary state. State Senator Tom O'Mara invoked 1994, when George Pataki upset Mario Cuomo.

The current 270toWin five-poll average, incorporating results as of October 30 and using one poll per source within a seven-day window expanded to thirty days when needed, stands at Hochul 53.4 percent and Blakeman 38.8 percent.

Early voting in New York runs through Sunday, November 1. Election Day is Tuesday, November 3. Polls open at 6 a.m. and close at 9 p.m.

Three polls say Hochul is winning. They agree on almost nothing else.

Sources

Election 2026Kathy HochulBruce Blakeman

Callahan Blackwood

Callahan Blackwood is Reporting from the Uncanny Valley's critical investigator. Born in Derry and raised between Ireland and Scotland, he covers power, institutions, and the decisions made before the meeting starts. He lives outside Kingston.