Polls Raise Red Flags

By SAM URETSKY

“Half of the American people have never read a newspaper. Half never voted for President. One hopes it is the same half.” —Gore Vidal

If Robert Kennedy Jr. has your email address and/or telephone number you can be sure of getting all the attention that modern technology can provide. Just in case you don’t have enough email to meet your needs, there will be two or three in your in-box, along with a text message and a telephone call the same day. Most of the email is deleted unread, but one is really worth looking at; it’s headed: “New Poll: Only One Candidate Can Beat Donald Trump.” It starts, “A massive new poll of the 2024 election cycle surveyed three head-to-head scenarios. The results are shaking up the political establishment”:

• Biden vs. Trump. Result: Donald Trump wins by a comfortable margin.

• Trump vs. Kennedy. Result: Kennedy wins in close race.

• Kennedy vs. Biden: Result Kennedy wins in a landslide.

This 50-state poll from Zogby, distinguished by its unprecedented scale and precision, surveyed over 26,000 likely voters nationwide, achieving a remarkably low margin of error of just 0.6%.

Of course if you want to read something else, for example the Drudge Report, you might have seen this headline on July 10; SHOCK POLL: KAMALA BEATS TRUMP. HILLARY BEATS TRUMP. This headline leads to a report by Bendixen & Amandi International, a Florida polling firm with a reputation for polling Cubans. The shock poll was based on “1,000 interviews with registered voters who are likely to vote in the November 2024 General Election representative of current national electorate.”

That should wave red flags. The population of the United States is approximately 333.3 million, give or take a few million. Also, the material in a 10-year census can’t (won’t) be released for 72 years (US Census bureau: the U.S. government will not release personally identifiable information about an individual to any other individual or agency until 72 years after it was collected for the decennial census. This “72-Year Rule” (92 Stat. 915; Public Law 95-416; Oct. 5, 1978) restricts access to decennial census records to all but the individual named on the record or their legal heir.) That means that the latest census available is from 1950, and things have changed since then.

Also, there is no simple formula for calculating the sample size for a study to produce a statistically significant result. The American Statistical Association publishes 17 journals discussing different kinds of studies covering business and economics, biopharmaceutical research, agriculture, biology, the environment, sports, surveys, and more. In other words, asking 1,000 people whom they expect to vote for, or whom they would prefer, is not to be taken seriously. End of that discussion.

The massive new poll that RFK describes queried 26,000 likely voters, but, given the universe — the number of people who are likely to vote in the next election — that’s not much better than 1,000. Also, it was conducted by Zogby, which is not a good sign. The web site 538.com analyzes polls and gives appropriate weighting based on a number of factors affecting credibility. Their most reliable polls are:

The New York Times/Siena College

ABC News/The Washington Post

Marquette University Law School

There are a number of polls and sponsors that come close to the top rating (2.9 versus 3.) 538.com is owned by ABC but that doesn’t guarantee a place in the top tier, and ABC polling is different from the ABC News/Washington Post rating. They look at hundreds of polls and John Zogby Strategies rates 104 while Zogby Analytics comes in at 226. It’s not clear what the difference is. Actually, that’s not terrible, considering that they list 277 polling firms, followed by an amorphous bunch of polls that are listed alphabetically with no rating. Zogby uses an atypical polling method although it’s used by many other companies: organize a large number of respondents and ask them questions via email. YouGov, a company with somewhat similar methods, is #4 on the list. Too many media reports omit the name of the pollster, or the methodology. It’s too easy to believe that the Kennedy gave three choices and Kennedy was a surrogate for “none of the above.” Similarly, the Bendixen poll might have included Vice-President Harris and Ms. Clinton but not Kennedy.

The Pew Research Center, 40th on the 538 list, has an excellent discussion of the methodology of polling at PewResearch.org (See the link with the online version of this article). Also, Dana Milbank wrote a brilliant column (see page 23) on the choices we seem to have – every Democrat should read it and share it. The fact is polls don’t tell the story – and it’s difficult to get the facts, but try. Before November.

Sam Uretsky is a writer and pharmacist living in Louisville, Ky. Email sam.uretsky@gmail.com

From The Progressive Populist, August 15, 2024

Populist.com