There’s just one problem: Those polls are likely to be exaggerating any differences in electability among the candidates.
To show this, I draw on data from Democracy Fund + UCLA Nationscape — a weekly survey of more than 6,000 Americans. In this survey, as in most national surveys, Americans generally favor the major Democratic candidates over Trump.
For example, in the Oct. 24-30 weekly survey , those who say they are registered and intend to vote in 2020 prefer Joe Biden over Trump by 52 percent to 40 percent. Several other candidates for the Democratic nomination are also leading Trump, including Bernie Sanders (51-42), Elizabeth Warren (49-42), Pete Buttigieg (46-40), and Kamala D. Harris (46-42). These numbers haven’t changed a great deal since mid-July, when this survey project began.
If the candidates who perform better in match-ups are actually more electable, this survey suggests that Biden is most likely to defeat Trump.
But here’s why that is misleading: How Democratic candidates perform in these match-ups depends a lot on name recognition. While Biden and Sanders have spent considerable time on the national stage, Americans know less about other candidates.
For example, when asked whether they have a favorable or unfavorable view of these candidates, a significant number of Americans say they “haven’t heard enough” about Buttigieg (40 percent), Harris (25 percent), and Warren (18 percent). By contrast, fewer than 1 in 10 Americans say the same about Biden (9 percent) and Sanders (6 percent).
As it turns out, there is a strong negative relationship between the percentage of people saying they “haven’t heard enough” about a candidate and that candidate’s performance against Trump. In weekly data from mid-July through October, it’s clear that the better-known Democratic candidates are generally performing better than their lesser-known counterparts.
In other words, Warren, Harris, and Buttigieg’s low name recognition is making them look less electable than they truly are. If their name recognition were to increase — as it certainly would if they won the Democratic nomination — polls could show that they would do as well against Trump as the candidates who are now better known.
Does this mean that all the candidates would do just as well if they were equally well known? Not necessarily. Biden and Sanders provide an excellent example. Despite being slightly better known, Sanders actually performs somewhat worse against Trump than Biden does. At this point in the race, Biden appears to be more electable than Sanders. This raises a question for candidates like Warren, Harris or Buttigieg: If their name recognition increased, would they look more like Biden or Sanders when matched against Trump? In each case, about a quarter of those saying they hadn’t heard enough about a candidate for the Democratic nomination also did not know how they would vote in a match-up between that candidate and Trump. In other words, these voters are doubly undecided — they don’t have an opinion of the Democratic candidate and do not have a preference between this candidate and Trump.
As a thought experiment, I looked at whether these voters had a more favorable opinion of Trump or “Democrats.” I assumed those with a more favorable opinion of Democrats would ultimately vote for the Democratic candidate once they learned more about the person. I assumed that those who had a more favorable opinion of Trump would ultimately vote for him.
This exercise increased Harris’s margin over Trump slightly (from four points to five points) but increased Buttiegieg’s margin even more (from six points to 10 points). Warren’s stayed the same (eight points). Thus at least some Democratic candidates may be more electable than they appear. At the same time, Biden retains a larger lead over Trump (12 points).
As the Democratic primary campaign continues, match-up polls will undoubtably continue to fuel the debate about electability. But we should take those polls with a grain of salt. Although they may reveal something about electability, they’re also exaggerating the differences among the Democratic candidates in how they would perform against Trump.
Robert Griffin is a political scientist and research director of the Democracy Fund Voter Study Group.
Read more of TMC’s analysis about the 2020 presidential election: