In the case of a memo released by President Trump’s reelection campaign on Sunday, the aim was something slightly different. In this case, the effort was to give a veneer of complexity and statistical rigor to a fundamentally weak set of claims.
Addressed to “Interested Parties,” the memo is titled, “Enthusiasm and Unity Update, Polling Methodology.” It attempts to downplay Trump’s weakened position in the polls by arguing, first, that Trump enjoys a distinct advantage in enthusiasm over former vice president Joe Biden, the presumptive Democratic nominee, and, second, that the campaign’s internal polling is more accurate than public polling.
Let’s evaluate the claims made by deputy campaign manager Bill Stepien.
The ‘enthusiasm gap’
In the abstract, this is actually an important point. Trump won the presidency in 2016 almost certainly in part because his voters were more motivated to go to the polls. His opponent that year, Hillary Clinton, was as unpopular as he is and, what’s more, was widely perceived to be likely to win. There’s little question that some voters, indifferent about her as a candidate and expecting an easy victory, simply didn’t cast a ballot. As we’ve noted before, Trump also benefited from support among many unenthusiastic voters: Those who liked neither him nor Clinton voted for him by a 17-point margin.
If the Trump campaign could demonstrate that Biden voters are far less enthusiastic about their candidate, it might show that Trump retains that advantage. But the memo doesn’t demonstrate that.
First, it walks through the number of primary votes Trump received this year, comparing them to past incumbent presidents and to his own 2016 totals. This argument can be summarized by Stepien’s comparison of Biden’s position to Clinton’s four years ago.
Consider “that in 2016, Hillary Clinton received 55% of the total votes cast in the Democrat primary,” the memo reads. “Today, Joe Biden has earned less than 50%. The real conclusion to be drawn is that, despite having no opponent actively campaigning for over two months, the former Vice President faces vibrant disunity within his own party.”
Clinton, of course, was facing one serious opponent in her bid for the nomination, Sen. Bernie Sanders (I-Vt.). Biden was facing a historically large field, divvying up the results of early contests. (And even some later ones, thanks to absentee ballots cast before candidates dropped out.)
It’s not terribly surprising that a nominee facing a large field would wind up with less than half of the votes cast in the primary. After all, that’s what happened to Trump in 2016.
It’s not the same situation, of course, because the Republican primaries in 2016 were contested until early May. Still, Trump’s campaign also had the ability to campaign, which Biden’s didn’t, given the coronavirus pandemic. The Trump campaign’s celebration of its robust support this year is due in part to its unusual decision to campaign hard and turn out voters despite not facing a serious opponent — and, in fact, working to limit any primary opposition.
Stepien also cites recent Fox News and New York Times-Siena College polls to demonstrate that Trump has an advantage in enthusiasm among voters this year.
“In the recent New York Times/Siena College polls of Arizona, Florida, Michigan, North Carolina, Pennsylvania, and Wisconsin, the enthusiasm gap widens,” he writes. “In those states, the President’s enthusiasm advantages range from 32% to 46% — meaning that in these states, far more voters say they are voting for President Trump because of President Trump, as compared to voters who say they are voting for Joe Biden because of Joe Biden.”
What’s being compared here is the percentage of respondents who support one of the two candidates who say that they are doing so mostly because they support that candidate. Trump does consistently have an advantage on this metric. But consider the flip side of this coin.
In each state, most Biden supporters say it’s more important to them that Trump lose than that Biden specifically win. That’s a form of enthusiasm, too: people who go to the polls to kick someone out. You may recall a similar sentiment in play in 2018, when voters went to the polls in part to send a message about Trump’s leadership. Democrats easily regained control of the House.
Obviously, Stepien doesn’t want to play up the anti-Trump vote, but it’s not insignificant. Many of those dislike-everyone voters from 2016 voted for Trump because they disliked Clinton more. In polling this year, those who say they dislike both candidates have consistently given Biden an edge of more than 20 points.
Public vs. internal polling
While there’s some validity to questions about how enthusiastic voters are, Stepien’s arguments about how trustworthy polling is are much weaker.
In essence, he argues that internal polls are more reliable than ones conducted by news organizations because they are better able to capture public sentiment. Stepien presents a number of factors that he says makes that argument true. Among them:
- Internal campaign polling looks at likely voters and not registered voters more broadly.
- Internal polling focuses on known voters, while other pollsters call randomly.
- Internal polls more narrowly tailor their geographic targeting.
- Internal polls have a better sense of the demographics of the electorate.
- Internal polls use live callers.
- Internal polls use professional callers instead of college students.
- Internal polling is conducted in multiple languages.
One advantage of Stepien’s argument here is that he can talk about media polling broadly while conflating different methods and tools the pollsters use. So, for example, The Washington Post’s polls with ABC News are live-caller polls using professional callers and are conducted in multiple languages. So are many of the other most recognized polls. Some polls do use more questionable tactics that haven’t necessarily been proved to be accurate, but conflating them with polls like ours is meant to muddy the waters.
Some of the advantages that Stepien cites aren’t advantages at all. Filtering for likely voters four months before the election necessarily makes assumptions that could turn out to be inaccurate. Calling random phone numbers is a tricky way to ensure that you’re calling registered voters, but it’s a good way to call random people for polling that’s got a broader scope than looking just at views of the election.
The simplest rejection of Stepien’s argument, though, is to point to the repeated success of high-quality polls in predicting election outcomes. While Stepien’s boss, the president, likes to claim that polls like Fox News’s were wildly wrong in 2016, the reality is that national polling four years ago was remarkably accurate. Most polls gave Clinton an advantage of a few points, which, you’ll recall, is actually how the 2016 election turned out.
One good indicator that Stepien understands that polls like Fox’s are quality polls is that he used Fox’s results to argue that Trump enjoyed an enthusiasm advantage. You can’t simply declare that methodologies like the one used by Fox News are bad or that college students can’t be trusted to poll effectively a few paragraphs after touting Fox and Siena College polls as proof of your point. That’s not how it works.
Sorry to inform the Do Nothing Democrats, but I am getting VERY GOOD internal Polling Numbers. Just like 2016, the @nytimes Polls are Fake! The @FoxNews Polls are a JOKE! Do you think they will apologize to me & their subscribers AGAIN when I WIN? People want LAW, ORDER & SAFETY!— Donald J. Trump (@realDonaldTrump) June 29, 2020
More broadly, the public should absolutely not trust internal polling — conducted opaquely and released strategically — over public polling in which methodologies and questions are released for review. Internal polls have repeatedly proved to be wildly inaccurate, as they did in 2014 when polling internal to former Virginia congressman Eric Cantor’s campaign showed him winning reelection by more than 30 points. He lost by 11 points. There are certainly examples of public polls getting results wrong (as in several states in 2016), but at least there’s some subsequent measure of accountability.
In his memo, Stepien argues that the media is intentionally making Trump look bad to “create an unfavorable scenario and to attempt to discourage President Trump’s supporters.” I guess if you are trying to identify likely voters in June, you’re also convinced that voters can be discouraged from voting four months early, too.
In fairness to Stepien, he may simply be echoing a line that has been used by Trump’s team repeatedly in recent weeks as polls have shown him losing the national vote by a double-digit margin. Three weeks ago, Trump tweeted another memo from a pollster arguing that a CNN poll showing him trailing Biden by a wide margin was similarly intended to suppress his November vote.
That memo was written by the guy who did Cantor’s internal polling.