I have often used my column to take advantage of a teachable moment. After all, I spent years in graduate school and taught political science for three years at Bucknell University.

So when I recently received a press release from New York Republican Rep. Elise Stefanik, who has been preparing to become a candidate for governor next year, I immediately went into teaching mode.

Stefanik was once regarded as a pragmatist who reflected the views of most Republicans from the Northeast. But she has instead become a follower of President Donald Trump and fills her statements with exaggeration and hyperbole.

The Stefanik press release cited an article in NewsMax, a conservative, pro-Trump media company, that allegedly showed the results of an internal poll by Stefanik’s campaign. That was a red flag right from the beginning.

The Sept. 20-26 survey of 1,250 “likely voters” conducted for Stefanik’s political committee E-PAC by Grayhouse Strategies found her trailing incumbent Democratic Gov. Kathy Hochul by a mere 5 points, 48 percent to 43 percent.

But when “respondents were given information about both candidates’ positions, including Hochul’s endorsement of New York City mayoral candidate Zohran Mamdani, Stefanik edged ahead by less than a single percentage point, 46.4%-45.9%.”

If Stefanik’s lead seems odd to you, you aren’t alone.

Reputable pollsters generally don’t report results to a tenth-of-a-percentage point, since that would create the impression that polling results are more accurate than they really are.

Everywhere else in the press release, percentages are rounded up or down to the nearest percentage point.

I don’t know who would win a Hochul-Stefanik race for governor, but I do know that any “internal poll” isn’t a survey that I would pay attention to. That is particularly true when the survey includes multiple ballot tests.

Several other issues to consider when evaluating this poll or others like it:

Surveys that ask respondents how they would behave if they knew something they don’t already know are usually a waste of time. 

Who knows what issues will decide the 2026 elections? 

Are respondents given both positive and negative information about each candidate? 

Does the poll test Trump’s standing in the state?

Why should I place great weight on a survey of “likely voters” when I don’t know how enthusiastic the two parties will be next year or what specific screens the pollster used to determine who is a “likely” voter and who isn’t. 

Regarding that last item, this is one reason why most pollsters sample “registered voters” or even adults early in a campaign before switching to likely voters later in a cycle, when they have reason to believe that they can identify “likely” voters.

The references to Mamdani in the questionnaire is not by accident. The Democratic mayoral hopeful is a contentious political figure, and the use of his name in the survey makes the poll as much about Mamdani as Stefanik. 

In fact, Mamdani’s presence in the gubernatorial contest might have damaged Hochul’s standing in New York state or in the polling. That might or might not be the case a year from now.

According to the NewsMax article, Stefanik said that Hochul “bent the knee to the communist antisemite running for mayor of New York City” and was the “worst governor in America.” 

Instead of ideological gibberish and talk about how weak Hochul is, I would have preferred more crosstabs from the questionnaire. 

The poll said little about Stefanik’s image, other than a question about whether she was viewed favorably or unfavorably. Nothing about whether voters identify her with Trump, or anything about her previous role as chair of the House Republican Conference, which, given the current shutdown, Democrats will likely be happy to remind voters of.

Consider also the most recent Siena University poll of a potential Hochul-Stefanik race, which works through these aforementioned issues, measuring each candidate against the backdrop of other public figures, from Trump to Mamdani to Speaker Mike Johnson to Health & Human Services Secretary Robert F. Kennedy Jr. to former New York Gov. Andrew Cuomo, who is also running for New York mayor against Mamdani. And the Siena pollsters are continuing to survey registered voters, a best practice at this point in the race.

The result? The most recent Siena poll had Hochul up by 25 points.

It is not hard to imagine the Stefanik internal poll as little more than a partisan effort to boost her gubernatorial prospects.