The 2024 Polling Predicament
Don't Be Deceived. As We Know From 2016, 2020, and 2022, Polls Are Not Absolute Predictors of Election Winners.
To start, in the midst of commencements across the country, I want to extend my congratulations to the Providence College Class of 2024! Which includes my older brother David, the author of Salzillo’s Two Cents. You can see the full Providence College commencement event here.
On a side note, Gina Raimondo’s RhodeWorks program has proven to be quite a “success.” So much so that a RIDOT official says it will take even more time than expected to fix our roads. So much so that Gina’s efforts have boosted Rhode Island from dead last in infrastructure nationally to second-to-last in infrastructure. At this pace, At that pace, it will take Rhode Island around 400 years to reach that #1 in the nation spot.
Is this the record of “success” she has brought and will bring to the U.S. Commerce Department?
Anyways, we are now in late May. In recent days, I have heard what many of you have probably been hearing. That President Joe Biden’s path to victory looks slim and gloomier by the day. That Donald Trump’s candidacy is on the ascent, breaking into the Biden-Harris 2020 coalition. Pundits see the significant swings in voter sentiments and mood as an indicator that President Biden is on the path to defeat, being behind significantly in states like Nevada, Arizona, Georgia, and Michigan.
It is true that President Biden’s approval numbers are not great, if you look at the consistent trend line of the polls since late 2021. Of course, you can also argue Donald Trump’s own approval ratings are not that much better.
It is also true that President Biden’s reelection campaign is the underdog, given the public sentiment (something that, honestly, as a supporter, I really hope they have taken note of by now). There are many signs that show the White House has work to do in at least shoring up the majority coalition it had built back in 2020.
All that said, it is hard for me to see how the President is in an unusually perilous position. In fact, it can be argued the situation is not much different from Barack Obama’s reelection campaign in 2012, with Republicans in control of the House, and a disadvantageous Senate map for Democrats looking to retain their majority there. We know how it turned out in the end. A lot of lessons can be learned from that election season.
As for the corporate media, they are spinning their wheels over a recent New York Times/Siena poll that showed Biden way behind Trump across the board. The poll has been cited numerous times for almost endless amounts of time, courtesy of numerous TV media outlets. Frankly, it is probably the best example of how the corporate media landscape (i.e., the penny-pinching business executives and management, not the journalists and reporters stuck in the middle in those situations) is more interested in political entertainment and melodrama to boost revenue and ratings. You know, rather than promote fair, unbiased, credible, and tough journalism (a fact some readers are already aware of from my past takes on US Commerce Secretary Gina Raimondo). For some conglomerates like Sinclair Broadcast Group, shrinking local news studios and promoting anti-Biden or pro-Trump propaganda hit pieces speaks to a decades-long society neglect of offering a truly free and informative press for the American people and our democracy.
What many reports fail to mention about the New York Times/Siena poll is that many of the top polling experts and campaign strategists have bluntly questioned the poll’s accuracy. And no, I don’t mean partisan sycophants or rosy optimists. I mean widely respected people nationally and on-the-ground like Larry Sabato, Jon Ralston, Marc Caputo, Greg Sargent, and John Della Volpe. They have all questioned the methodology of the polling and the varying results from state to state.
And they have good reason to. How can the Wisconsin results show Biden ahead by 2 points (which is more than twice his 2020 margin), whereas the Nevada results have Trump ahead in the polls by 14 points (an approximately 16-point swing from 2020)? Even in these crazy times, who can believe those poll numbers provide a good sense of the current electorate?
Likewise, how could Biden get crushed with the young vote in Michigan, be tied with the young vote in Pennsylvania, and then trounce Trump with that same group in Wisconsin? Considering what the NYT-Siena poll once said about 2022, I would be skeptical of relying on it to help us determine who will win anything (vs. the gold standard Des Moines Register). It is a true outlier from most of the polling averages.
This is not to call out either The New York Times or Siena College. Because the fact is that many polls have not been as accurate as most people would expect, even though they still garner much publicity and headlines. In defense of the pollsters, it truly is getting harder to poll in a way that was doable even 10 years ago.
Still, let’s take it step by step. The 2022 Midterms predicted a red wave across the board. Instead, the real thing was a wash determined more by partisan polarization and where the country fits into its two ideological camps. Issues like abortion rights, gun safety, and democracy were bigger issues than were previously reported up to Election Day 2022. Not to mention that, per CBS News exit polls, Independents narrowly supported the party in power (Democrats) over the minority party (Republicans) 49%-47%. This was a trend not seen in many recent election cycles, resulting in what we can only call a Biden electoral victory 2 years ago.
We should also remember that some polling firms—whether unintentionally or even intentionally—hastily released junk polls skewing more Republican. We saw this in the national party preference polls for the House of Representatives and the U.S. Senate. We saw it in the races for Governor—like in Arizona between Kari Lake and Katie Hobbs. These polls featured Lake as the decisive favorite to win. This has real consequences. For instance, it led some national Democrats to contemplate abandoning the Senate race completely between Ron Johnson and Mandela Barnes because they viewed the contest as unwinnable. Contrary to their expectations, the race was not decided by 6 points, or even 3 points as most of the polls predicted in October and early November, but by 1 point. Taking polls too seriously, ironically enough, may impact, or even decide, crucial election contests.
Let’s look at some other past polling flubs. In 2020, despite the polling indicators, Trump overperformed expectations all across the country, making the election closer than many anticipated. In fact, polls back in summer of 2020 showed Trump behind in Michigan, Wisconsin, and Pennsylvania by double digits. Those results were not nearly that close; they were not even in the mid-to-high single digits.
And 2016 really was the first election to cause systemic doubt of the polls. As we all know, Hillary Clinton was “supposed to win.” But by 3 AM in early November of 2016, it would become very clear that she would not. However, these results should not necessarily have surprised Democrats as much as they did given Clinton’s ten-ton heap of electoral baggage.
To be sure, these trends are not new either. Polling in complex situations is hard. And the polling situation is especially difficult now. We are in a season where age is a big issue for both candidates (Biden is 81, while Trump is very close to him at 77), and Trump’s legal troubles will weigh heavily on voters. On top of that, respondents do not always give reliable answers, even on simple questions such as how likely they are to vote, or if they ignore the personal controversies of certain public officials.
In 2006, corruption tainted many congressional incumbents. Similarly, blunders like those of Virginia Senator George Allen in his very infamous “macaca moment” definitely influenced the course of that close race: he lost very narrowly to his opponent Jim Webb. On the other end, even though 1998 was another incredibly difficult cycle defined by the drama on Capitol Hill over the Lewinsky affair and the Newt Gingrich-led impeachment effort against Bill Clinton, voters were more likely to say the economy decided their voting choices more than anything else.
Last but not least, in 1948, some newspapers inaccurately projected Thomas Dewey as the election winner. As history buffs know, Dewey in reality lost to President Truman by a rather significant landslide.
Polls are snapshots in time—particularly if they are done in an effective way. That has been much harder to pull off since 2016 with Trump voter distrust of the polls, shy Biden voters (such as those in the 2022 midterms and the off-year special elections), the decreasing use of landlines, and the greater problems with selection bias. In all, the mission of polls to provide a spot-on assessment of the electoral landscape is an increasingly challenging one. Which is why junk polls and corporate media melodramas only create more confusion about the state of the election, and how polling data should really be analyzed.
Not that the fundamentals are necessarily off. We can look at past electoral history, voter data trends, and the sights and sounds on-the-ground. Right now, we can easily say the 2024 elections are a coin-flip based on those indicators, which is what most of the general polling averages do show. Yet as Simon Rosenberg has put it, “I would rather be us than them.”
Finally, anything can happen from now to November. There could be consequential Supreme Court rulings, or new proposed pieces of a congressional agendas from either party. There could be a Trump conviction, or something else. There could be improvements around inflation, or a resolution to the conflict in the Middle East. October Surprises (like the inexcusable Paul Pelosi attack in 2022) can still happen, even in this day and age.
In short, polls six months from the election should be taken with a grain of salt. As common-sense tells us, the only true poll that ultimately matters is the verdict that comes down on Election Day.