A POLL THAT COUNTS

As a general rule, polls have limited value.  

As any expert will tell you, they are mere snapshots in time.  

When used to predict outcomes, their accuracy is off by anywhere from three to five percentage points.  Pollsters call this the margin of error. That error can be the result of a flawed or limited sample size.  In that case, the predicted outcome is likely to be erroneous because the people polled are not the people who later (and actually) decide. This is especially significant when the poll also reports the views of smaller sub-groups.  

There have been two polls much in the news this past week.

Over the weekend, the Washington Post/ABC News poll reported that 36% of its respondents approved of the job being done by President Biden. 56% disapproved and 8% had no opinion.  This was a decline of six points from its last survey conducted two months ago. The poll also showed both former President Trump and Florida Gov. Ron DeSantis beating Biden in a head-to-head by seven points.  And it reported that 54% of the respondents thought Trump did a better job handling the economy, an eighteen-point edge over those who favored Biden on this issue.

To say the poll immediately provoked sustained concern among Democrats would be an understatement.  

On This Week with George Stephanopoulos, former Democratic National Committee chairwoman Donna Brazille said the poll, which was released just after midnight on Sunday, "kept [her] up" and told the administration that it should "wake up".  On Monday, The New York Times reported that "the data has left many Democrats feeling anywhere from queasy to alarmed". And on Tuesday, Robert Hubbell, a Democratic partisan whose blog Today's Edition Newsletter is widely read, reported that he had received "dozens of emails from distraught readers over the last forty-eight hours" in reaction to the poll.  

The poll was not without its critics.  

As a matter of statistics, it is an outlier.  

Other polls give Biden higher marks and the average of all polls in Five Thirty-Eight's daily tracking charts the president's approval rating at 42.5%, which is just about where it has been for the past two months (or the same period surveyed by the Post and ABC).  

At least one pollster, Cornel Belcher, called the Post/ABC poll "trash". 

Belcher, who was Obama's pollster and otherwise claimed to have had "respect for the poll in the past", criticized its methodology as "problematic".  The sample size (1006 adults) was inordinately small, worse (only 900 registered voters) on the Biden/Trump match-up, and more or less useless in evaluating the views of relevant sub-groups (by race and age and of independents, the latter of whom will be critical in 2024). 

In the published poll itself, there was no report of the so-called "demographic" questions or of the numbers of respondents that fell into each demographic group, however defined,  and thus it was impossible to understand the significance of any of the so-called "cross tabs" reporting the approval and match-up results for those specific groups.

Its defenders are noting that the current Post/ABC poll is reporting the same trends the poll reported earlier in its January and September iterations and thus less an outlier than what might otherwise be thought. 

A Nate Cohn wrote in today's Times, "This consistent pattern requires more than just statistical noise and random sampling.  Something else is at play, whether that's something about the ABC/Post methodology, the underlying bias in telephone response patterns these days, or some combination of the above." Cohn then suggested what that "something else" might be, noting that "the ABC/Post poll is nearly the last of the traditional, live-interview, random-digit-dialing telephone surveys that dominated public polling for much of the last half-century".

The bottom line here is that the Post/ABC poll tells us a lot about today but nothing about November 2024.  

As for today, the poll tells us that a roaring economy with 3.4% unemployment and declining (but still annoying) inflation is not making Americans feel better about their particular circumstances (or their particular president).  This is probably because (i) the top-line employment number does not cure the reality that employment is not guaranteed and inequality is still rampant and (ii) unlike unemployment, inflation and high interest rates impose costs on a much wider swath of the public.

As for November 2024, however, the poll tells us nothing.  

Or . . .

Too much of everything . . .

Which gets you to the same place. 

Though the poll makes it clear that Biden is unpopular, it says Trump is too.  It reports that more than 63% believe Trump is a liar, more than half think he should be indicted in connection with the January 6 insurrection and his retention of classified documents, and about half deem Manhattan District Attorney Bragg's indictment appropriate. As for the man who gave us the Court that killed Roe, 78% believe abortion should be between a woman and her doctor.

The second poll conducted this week had none of the uncertainty of the first.

This one was conducted in New York City.

In a federal courthouse.

In it, nine jurors unanimously voted that Donald Trump had sexually assaulted E. Jean Carroll in a dressing room at Bergdorf Goodman in the 1990s and then defamed her after he left the presidency by calling her claim a "hoax and a lie" and a "complete con job" in October 2022.  (A separate defamation case for comments he made about the case while he was president is still pending.)

There are enormous differences these days between legal proceedings and presidential campaigns.  

The first are governed by rules of evidence and procedure and subject to review.  The second are pretty much no holds barred slugfests.  

In the first, facts matter and witnesses testify under oath and on pain of perjury.  In the second, media groups routinely fact check the claims of the candidates, many of which are found wanting or false.  

In the first, jurors deliberate under strict guidelines concerning the applicable law and the degree of certainty they must reach before rendering a verdict.  In the second, 35% (and before 2020 usually more) of those eligible to participate decide for one reason or another (or none at all) not to, and a sub-set (so-called "low information voters") do so with little thought or analysis.  

Following yesterday's verdict, Trump repeated his (now demonstrably false) claim that he did not know Ms. Carroll and did not assault her.  He attacked both the jury that heard the case and the judge who presided over it. He said the latter was biased because he was appointed by Bill Clinton, the former because it was empanelled from "an anti-Trump area."  This last charge was echoed by Senators Graham, Rubio and Tuberville, the last of whom said that with "a New York jury, [Trump] had no chance."

All of this is standard issue Trump talk and standard issue GOP denial.

In 2020 and the months following the November election, Trump and his GOP enablers twisted themselves into knots claiming the presidential election Trump lost was rigged but the House races the GOP won on the same ballots were legitimate.  

And yesterday, Trump and his enablers all sang from the same "no chance with a New York jury" chorus even though that same New York jury had decided Ms. Carroll did not prove Trump raped her.

As Yogi Berra would say, it was déjà vu all over again.

With Trump, only blue state votes . . .

And blue state juries . . .

Are rigged.

Even if they're not. 

When all the dust settles and smoke clears, two facts will be undeniable and featured prominently in any Democratic campaign for the presidency in 2024.

One is that, on January 6, 2021,  Donald Trump sponsored an insurrection to overturn a legitimate election.

The other is that, on May 9, 2023, a jury declared him to be a sexual predator.

And that is . . .

A poll that counts.