The Teamster non-endorsement was accompanied by several polls which showed a supposed drastic drop in support for Democrats after Biden’s step down. But a look behind the data shows they’re comparing apples to oranges.
Yesterday, the International Brotherhood of the Teamsters, one of America’s largest labor unions, announced that they would not endorse a candidate in the 2024 presidential election. This surprised some, who expected to see Teamsters throw their support behind their traditional ally, the Democratic Party. Others were less surprised given the friendliness Teamster President Sean O’Brien has shown to Donald Trump in the past year.
Either way, this was a momentous decision, the first time the Teamsters had not endorsed the Democratic nominee since 1996, and a repudiation of the administration that had bailed out the Teamster pension fund to the tune of $36 billion. But along with their decision, the Teamsters released a set of polls of their membership showing, apparently, that there had been a clear plurality of support for President Joe Biden before he stepped down from the top of the ticket, and that subsequently, there was a massive groundswell of Teamster support for Donald Trump. This narrative is very enticing for journalists, after all, it’s dramatic and plays into the persona of Scranton Joe. But it should be taken with a grain of salt.
One likely reason for this is that these polls are not of the same type. The “poll” that originally showed an eight point Joe Biden lead over Trump was a straw poll of members present at in-person events with their locals. The subsequent polls were a phone poll conducted by Lake Research Partners, a polling firm, and an electronic “poll” (it’s a survey) of members. The electronic poll was alleged to be not well advertised to membership by one Teamster on Twitter, but without confirmation, we cannot know if this is true. What we do know is that each of these “polls” have massive response bias issues. One required in-person presence, another required answering a phone (a horrifying idea to a Gen Zer like me), and the third is basically a Google Forms survey.
This is not to mean that we think there is some massive undercounting of Harris-supporting Teamsters (though reporting from Harold Meyerson at The American Prospect suggests the endorsement process was not as simple as following the “polls”) or that one of these polls cannot be accurate. But what we do know is that these polls cannot be accurately compared to each other. Unfortunately, the narrative about a supposed massive swing of union members away from Biden towards Trump is present both in bad news coverage, and from sources that are normally very reliable. But it is based on information that is, at best, without strong evidence.