The Hill (2/17/23) turns a 3-point change in the margin between Gov. Ron DeSantis and former President Donald Trump—11 months before the first Republican nomination contest—into a headline.
The Hill (2/17/23) announced last week, “DeSantis Approval Drops in GOP Primary: Poll.” The article, by Max Greenwood, went on to say:
A new Harvard CAPS/Harris Poll released Friday exclusively to The Hill shows Trump leading DeSantis 46% to 23%. That marks a 5-point drop in support for DeSantis since last month, when he trailed Trump by 20 percentage points in the same poll.
In fact, the poll showed no significant differences between January and February in that multicandidate matchup. (For poll information noted below, see a compilation by 538.) The actual difference in Trump’s lead between the two polls was just 3 points—48% to 28% in January (a 20-point margin), 46% to 23% in February (a 23-point margin). Yes, DeSantis’ support fell by 5 points, but Trump’s fell by 2 points, for a net marginal change of just 3 points. Such a small poll difference hardly proves DeSantis is losing support.
Later in the article, the author indicates that “perhaps more alarming for DeSantis” were results showing the Florida governor with only 39% support, when Trump was excluded from a matchup that included Mike Pence, Nikki Haley, Ted Cruz, Mike Pompeo, Marco Rubio and Tim Scott. In a similar poll the previous month, DeSantis had 49% support.
Yes, there was a 10-point decline, in part because Pence and Haley picked up 8 points between them. Still, the results were hardly “alarming.” Taking any of these early polls seriously is foolish.
11 months in advance
Early polling (Extra! Update, 6/07) told us that the 2008 presidential race would be between Rudy Giuliani (above) and Hillary Clinton—rather than John McCain and Barack Obama, the eventual nominees.
Should we really pay attention to national primary polls 11 months ahead of the first contest?
The obvious answer, of course, is no. A year from today, Trump may well have been indicted on one charge or another—and could conceivably be convicted, though even getting a case to trial will require navigating a long appeals process. How an indictment, much less a conviction, would change the political landscape is anyone’s guess.
Apart from that, it’s unclear which Republicans will run for president and, more importantly, how each individual will fare in the campaign. While DeSantis is clearly among the better known candidates, other than Trump, he has yet either to declare his candidacy, or to begin campaigning in the first two states in the delegate selection process—Iowa and New Hampshire. The potential challengers to the former president have yet to prove they are ready for the intense media scrutiny that comes with being a presidential candidate.
The current national primary polls thus reflect mostly name recognition of the candidates challenging Trump. Even then, the national polls don’t tell us what voters in Iowa and New Hampshire are thinking, though those two states can have an outsize role in screening out candidates and launching others into frontrunner status.
There is a long history of early polling being next to useless in predicting the results of competitive primary races (Extra! Update, 6/07). Still, looking at how news outlets read these tea leaves can tell you something about the media’s own preferences in political campaigns.
DeSantis the new frontrunner?
Over a year and a half ago, the GOP Daily Brief (6/21/21) jumped on Gov. Ron DeSantis’ bandwagon, announcing: “Republican 2024 Race Gets Fresh Frontrunner—Donald Trump Is No Longer Leading, Now Governor DeSantis Is.”
A PJ Media column (11/12/22) promoted DeSantis as the “greener pastures” Republicans were said to be looking for.
It took the GOP’s disappointing showing in the recent midterm elections, when Trump’s candidate endorsements mostly hurt the GOP, for Matt Margolis—conservative commentator and columnist for PJ Media (11/12/22)—to come to the same conclusion. “There’s a New Frontrunner for the 2024 Presidential Nomination,” blared his headline. “Many Republicans,” he wrote, “are starting to wonder if there are greener pastures for the GOP with someone like Ron DeSantis, according to a new poll from YouGov.”
That new YouGov poll (11/11/22) was a national survey of potential Republican primary voters that pitted DeSantis vs. Trump in a hypothetical head-to-head matchup. The Florida governor got 42% support, Trump 35%. The most recent Yahoo/YouGov poll (2/8/23) shows little change over the intervening three months, with DeSantis still beating Trump in a head-to-head matchup, 45% to 41%.
Perhaps the strongest advocate for DeSantis’ frontrunner status is Aaron Blake of the Washington Post (1/14/23), who listed the Florida governor as the most likely among ten possible GOP candidates to be the Republican nominee for president, a position DeSantis has occupied on Blake’s list for several months. Blake’s principal evidence consisted of national primary polls that included only DeSantis and Trump.
He downplayed the multi-candidate polls that did not support his theory. The vast majority of polls that matched Trump against several candidates simultaneously found Trump getting the most support, often by double digits.
That led Nathaniel Rakich of 538 (1/10/23), who analyzed the December 2022 polls, to write that “DeSantis is polling well against Trump—as long as no one else runs.” A recent poll by Yahoo News/YouGov (2/8/23) reaffirmed that conclusion: “DeSantis leads Trump for 2024 GOP nod—but not if Haley and others split the vote.”
But this new version of the conventional wisdom is not completely supported by the data either.
Head-to-head polls
Below is a list of all the national polls in January and February (as of this writing) compiled by 538 that report on a head-to-head matchup between Trump and DeSantis.
What’s most striking is the wildly contradictory pattern of results, which vary from a DeSantis advantage of 26 points (Marquette) to a 23-point advantage for Trump (Big Village)—a 49-point variation in the lead.
If the results are averaged (including the most recent results of Big Village, Premise, Harris/Harvard CAPS and YouGov, but excluding their earlier results, so they do not get disproportionate weighting), DeSantis averages 45.8%, Trump 44.3%—a 1.5 percentage point DeSantis advantage, not too far from a tie.
Misreading polls
It would be a mistake to interpret these results as though there is a volatile GOP electorate.
The only polling organizations that polled a head-to-head matchup in both January and February showed no significant change from one month to the next. YouGov/Yahoo News gave DeSantis a 3-point advantage in mid-January, and a 4-point advantage in early February. Harris/Harvard CAPS reported Trump with a 10-point advantage in mid-January and a 12-point advantage a month later. And Big Village recorded just a 1-point decline in Trump’s advantage from January to February, from 23 to 22 points.
NBC‘s subhead (1/26/23) told readers that “GOP voters slightly prefer Florida Gov. Ron DeSantis over former President Trump”—even as the story reported that Republican and Republican-leaning voters “preferred DeSantis as the eventual Republican presidential nominee over Trump by 28 percentage points.”
In addition, some polls conducted in the same or close time frames produce conflicting results. In fact, the most extreme numbers for the two candidates are produced in virtually the same time period. Big Village’s January poll ends on January 20, and shows Trump with a 23-point margin over DeSantis. Marquette’s poll ends on the same date and shows a DeSantis advantage of 26 points—a 49-point discrepancy.
Another example: YouGov/Yahoo News and On Message both show DeSantis in the lead (by 4 points and 15 points, respectively), at virtually the same time that Premise finds Trump in the lead by 20 points.
Opinion simply does not change that quickly.
The large fluctuations shown in the table thus reflect mostly the polling organizations themselves—their particular methods of sampling, question wording and data analysis. And it is indeed surprising, if not shocking, that polls of supposedly the same electorate, conducted over roughly the same time, should come to such contradictory findings.
The most obvious conclusion to draw from this table: We simply can’t trust any of these polls this early in the campaign to tell us which candidates might prevail in a head-to-head matchup—not “if the election were held today,” much less when it will actually be held many months from now.
The earliest polls that may have some meaningful results about the GOP candidates for president will be polls of the electorates in the four earliest states to choose delegates—the Iowa caucuses, and the primaries in New Hampshire, South Carolina and Nevada. But again, not for months, until shortly before the Iowa caucuses. Until then, arguing about the meaning of the latest national primary poll will have much in common with determining how many angels can dance on the head of a pin.
That will not deter the media from its obsession with early polls. Just in the past week and a half, several new national primary polls and poll stories have been reported—at Fox (2/20/23), NBC (2/16/23), Newsweek (2/19/23) Forbes (2/13/23) and PBS/NPR (2/22/23), among others.
But just because they are there doesn’t mean we have to waste our time reading them.
The post Early Polling Tells You Little About Next Year’s GOP Primary appeared first on FAIR.
This post was originally published on FAIR.