Earlier today, the International Brotherhood of Teamsters (IBT) released the results of a poll that showed a clear majority of Teamsters supporting Donald Trump. Just an hour later, the union announced it would not be endorsing any candidate for US president.

Many Teamsters are breathing a sigh of relief that the union didn’t endorse Trump after general president Sean O’Brien made an appearance at the Republican National Convention and after these polls indicate a fairly Trump-sympathetic membership. Others will be disappointed that Trump didn’t win the endorsement after the membership was polled. Still others will insist that the polls are rigged and that they should be discounted.

To some extent, those who take issue with the polls are not wrong. For instance, the original straw poll was conducted from April 9 through July 3 in union halls and was organized entirely by local union leadership. From anecdotal evidence, it seems that only a small minority of locals took time to mobilize and educate their membership on these votes, the stakes of the election, and the meaning of the polls. As a result, mainly active and engaged members (who tend to lean Democrat) showed up. This poll indicated that members backed then Democratic Party nominee Joe Biden 44.3 percent to Trump’s 36.3 percent.

Immediately following this poll, the IBT leadership announced another poll, this one online, that would supplement the in-person straw poll. There was a lot of speculation that this online poll was commissioned to counteract the results of the straw poll. There is no way of knowing whether that is true, but its results do indeed contradict the straw poll’s results. This poll went heavily for Trump. Conducted between July 24 and September 15, members voted 59.6 percent for Trump and only 34 percent for Democratic nominee Kamala Harris.

This pattern itself — a wild swing from Biden to Trump — doesn’t actually tell us very much. In fact, it confirms a suspicion held among many demographers and pollsters: that engaged and active members of organizations lean to the left. Votes that take place in person self-select for the kind of people who have higher levels of institutional trust and a higher commitment to political action — that is, they select for people who seem more like Kamala Harris voters. The online vote does just the opposite.

So, what about the phone research poll? This poll had Harris losing as well, by a wide margin — it went 58 percent for Trump and only 31 percent for Harris. Unlike the other two polls, this one was conducted on September 15 by a third-party polling firm, Lake Research Partners, which describes itself as “progressive.” If the firm is reputable, it would have taken a random group of the entire voting-eligible membership and produced a representative sample. There is no reason to doubt that it did this.

Therefore, it’s quite likely that 58 percent of Teamsters support Trump. This finding would be consistent with the fact that some 56 percent of non-college-educated voters overall support Trump, according to the Times/Siena poll from September 8.

Now the puzzle is, why? We don’t need to rehearse the litany of sins Trump has committed against labor — his National Labor Relations Board appointments, his closeness to union-busting billionaires, his comments encouraging the illegal firing of striking workers. The list goes on. So why would so many union workers still vote for him?

The answer is threefold.

First, inflation. There is no getting around it. Levels of inflation have been historic. Whether they are the fault of the president and vice president is almost irrelevant. Voters, especially working-class voters, are blaming them.

Second, the Teamsters (and they are far from alone in this) have not made it a priority to educate and mobilize its membership on political issues. In the past four years, the union has rarely made an effort to explain or motivate its distinct political interest in the upcoming presidential election. Most local union presidents (and this is true in almost all unions) avoid talking politics for fear of offending their members or causing arguments among their membership. The need for unity is incredibly important for a union, and so it makes good sense to avoid topics that might cause division. Still, the failure of union leadership to really articulate what working-class interests are and how the parties help or harm those interests is a factor. Such political education should be a monthly part of the union’s activities and not relegated to a few short meetings in the year of a high-stakes presidential election.

Finally, and most important, the Democratic Party has done a miserable job of trying to identify with the working class. Many, if not most, on the Left will offer an exasperated reply to this charge, insisting that the Democrats are so much better for workers and for unions. Yes, that is obvious from a policy perspective. But leading liberals take for granted that workers are people, too.

It is no secret that the Democratic Party’s image, interests, and value system are far closer to that of upper-middle-class liberals than they are to that of the average working-class person. On cultural issues, too, it is the case that many workers feel deeply alienated from the liberal perspective. In conversations with Teamster members and officers, it’s clear that even among Democrats and Harris supporters, there is a frustration and wariness toward many of the cultural positions Democrats endorse and how those positions are defended. Until the party is able to change this image and truly identify with the working class (on economic as well as cultural terms), we should expect to see a continued deterioration of working-class support for Democratic candidates.

Now, despite all of this, it would have been folly for the Teamsters to endorse Trump. Just because the union has failed to properly articulate its interest does not mean that interest does not exist. The union has a vital interest in preserving its legal rights — and those rights are explicitly threatened by the Republican Party and by Trump himself. For these reasons, the Teamsters ultimately decided to issue a “no endorsement” for the presidential election.

Such a decision might be prudent in order to avoid a civil war within the union, but unfortunately it only kicks the can down the road for fully articulating just what the union’s business ought to be when dipping its toes in political waters.

Worse, the “no endorsement” doesn’t bode well for Harris. The key states that she needs to win — Pennsylvania, Michigan, and Wisconsin — happen to be home to some of the union’s biggest membership strongholds. Teamster member-to-member political operations could have played a pivotal role in getting out the vote for Harris and influencing the outcome of the election. Worst of all, if the internal phone poll is anywhere close to accurate, it suggests that Harris is doing very badly among working-class voters.

Ultimately, the labor movement must develop a coherent political appeal to win its members over on terms that are relevant to the vast majority of the working class. That appeal cannot be a down-the-line progressive appeal, as found on Ivy League campuses or in nonprofit boardrooms, nor can it embrace any of the free-market conservatism and anti-tax populism of the Republican Party. In other words, the political appeal of labor must be one that can speak to the felt aspirations, fears, and needs of the working class while also outlining the ways that the government can improve the state of the economy and the workplace on a society-wide scale.