WOULD THE NEW YORK TIMES POLLING UNIT RATHER BE PRO-GOP THAN ACCURATE?

It's been obvious for quite a while that the folks who do polling for The New York Times feel really burned by their overestimate of the Democrats' chances in the 2016 and 2020 presidential elections. So it's no surprise that the polling numbers in the Times today are some of the worst the Democrats have seen recently:

Republicans enter the final weeks of the contest for control of Congress with a narrow but distinctive advantage as the economy and inflation have surged as the dominant concerns, giving the party momentum to take back power from Democrats in next month’s midterm elections, a New York Times/Siena College poll has found.

The poll shows that 49 percent of likely voters said they planned to vote for a Republican to represent them in Congress on Nov. 8, compared with 45 percent who planned to vote for a Democrat. The result represents an improvement for Republicans since September, when Democrats held a one-point edge among likely voters in the last Times/Siena poll.

Is this accurate? It might be. But in a sidebar, Nate Cohn of the Times inadvertently provides a reason to question the numbers. In a passage explaining the methodological changes Times/Siena has made, Cohn writes (emphasis in original):

We now use additional information about the attitudes of respondents in determining whether they’re likely to vote, including whether respondents are undecided; whether their views about the president align with their party; whether they like the candidate they intend to vote for; whether they back the party out of power in a midterm; and so on, all based on previous Times/Siena polls. At the same time, we now give even more weight to a respondent’s track record of voting than we did in the past.

So Times/Siena has decided that voters are less likely to vote if their views of Joe Biden don't "align with their party." The problem with this is that it's obvious in nearly every poll conducted this year that Democratic midterm candidates are outpolling the president, whose job approval numbers are still weak. What the folks at Times/Siena are telling us is that they don't really believe that people who dislike Joe Biden but like Raphael Warnock or John Fetterman or Mark Kelly will actually show up in November. Republicans don't have this disconnect -- obviously, they all despise Biden, so the poll is weighted in favor of their votes. That's an automatic bump for Republicans.

In addition, Cohn says, Times/Siena now gives "even more weight to a respondent’s track record of voting than we did in the past." But what Democrats are telling us is that there's been a registration upswing since the Supreme Court's Dobbs decision that's been disproportionately Democratic and female. Now, maybe those new registrants won't actually show up to vote. Maybe they'll be outnumbered by GOP voters fired up about immigration, inflation, critical race theory, and (nonexistent) schoolkids who identify as animals and use litter boxes. But Democrats are insisting that they'll turn out new voters, with some emprical evidence to back them up (the surprisingly strong vote in Kansas for abortion rights) -- and the response from Times/Siena is to say that this is less likely than usual. Again, a GOP bump.

I'm sure Times/Siena wants to be accurate, but, faced with conflicting signals, the pollsters have clearly concluded that they'd rather overestimate the GOP vote than underestimate it. If they overestimate the GOP's strength, they're sure Democrats won't "work the refs" by denouncing the pollsters' bias; if they underestimate the GOP's strength, they know that Republicans will never stop bringing it up.