In the wake of June’s Democratic presidential debate, most polling agrees that former Vice President Joe Biden is still the front-runner (although the size of his lead has taken a hit), followed by a rough three-way tie among Sens. Kamala Harris, Bernie Sanders and Elizabeth Warren. But as we already have seen numerous times this primary season, the standings are fluid. Many voters probably will change their minds between now and when voting begins. So it’s also valuable to take note of whom people’s second choices are.
When I first took a look at second-choice polling in February, there was little sign that the race was breaking down into “lanes” — a way of thinking about crowded primary fields that says smaller groupings of candidates are fighting over specific voting blocs within the party, such as progressive voters or black voters. Candidates like Biden and Sanders who dominated overall were also voters’ most popular second choices, no matter who their first choice was. However, fast-forward through four months of campaigning and one debate, and the polling indicates that those lanes may be starting to emerge — although there are three caveats that I’ll spell out shortly.
Let’s look at the data we have. In the past two weeks, at least three pollsters have asked Democratic voters nationwide about their first and second choices. First, a Reuters/Ipsos poll conducted from June 28 to July 2 found that Biden was the first choice of 22 percent of Americans, followed by Sanders at 16 percent, Harris at 10 percent and Warren at 9 percent. As for respondents’ second choices, Biden and Sanders were each other’s supporters’ top second pick, but Warren was Harris supporters’ top runner-up, and Harris was Warren supporters’ top runner-up, too.
In particular, the 38 percent of Warren supporters whose second choice was Harris stands out in comparison with the much smaller number who preferred Biden (14 percent) or Sanders (17 percent). Also interesting was that Sanders’s and Harris’s supporters seemed especially incompatible; for example, only 8 percent of Harris supporters chose Sanders as their backup.
Another national poll, taken almost concurrently (June 28 to July 1) by the Washington Post and ABC News, yielded similar results. Unfortunately, only the sample size for Biden supporters was large enough (152 respondents) to be reported: 37 percent of people who said Biden was their first choice picked Sanders as their second choice, 17 percent picked Harris and 14 percent picked Warren. (Voters’ first-choice picks broke down as Biden at 29 percent, Sanders at 23 percent, and Harris and Warren tied at 11 percent.)
Finally, a Morning Consult poll conducted July 1 to 7 found that 31 percent of potential Democratic primary voters chose Biden as their first choice; 19 percent chose Sanders, 14 percent chose Harris and 13 percent chose Warren. That poll also had a very large sample size, which allows us to look at even more candidates, such as South Bend, Indiana, Mayor Pete Buttigieg (the first choice of 6 percent of respondents), former Rep. Beto O’Rourke (3 percent) and Sen. Cory Booker (2 percent). And as you can see in the chart below, the Biden-Sanders lane is still evident in voters’ second choices. The Harris-Warren one also becomes clearer: Buttigieg is also a member of it, as 27 percent of his supporters pick Harris as their second choice and 19 percent pick Warren.
Perhaps what is happening here is that this is the “well-educated white liberal” lane, as polls have generally shown these three candidates doing well with those demographic groups. It may also be the “fresh face” lane — people who don’t want the 46th president to be older, straight, white and/or male.
Now for some caveats. First, the fact that all three polls put Biden and Sanders in the same lane teaches us that lanes may not be what you assume. Indeed, conventional wisdom would probably put Biden and Sanders on opposite wings of the party. Sanders uses revolutionary language and has some of the most progressive views in the field, while Biden is a pretty establishment-friendly candidate who has been pegged as a moderate. But so far at least, it doesn’t look like the race is breaking down along ideological lines. Maybe this is the “experience” lane — Biden and Sanders have each served in public office for more than three decades. Maybe it is the “electability” lane, as those two are generally seen as having the best chance to beat President Trump. Or maybe name recognition still matters, and Biden and Sanders are the only two candidates these voters know a meaningful amount about.
Second, lanes aren’t mutually exclusive; there are bound to be some candidates who “belong” to multiple lanes. For example, despite the Biden-Sanders vs. Harris-Warren-Buttigieg dichotomy identified above, Warren supporters’ second runner-up is Sanders, and vice versa. That may suggest that some Warren and Sanders supporters have ideological overlap and are voting based on the progressive economic policies and anti-corporation sentiment that they share. Similarly, Biden is actually the second-ranked second choice of Harris supporters, and vice versa. That could represent the thinking of black voters, who really like Biden but also are starting to show support for Harris, too. Cross-currents like these can make voters’ second choices less straightforward.
Related to this complicated network of preferences is my third caveat: Lanes are easy to overstate. In most cases, the share of Candidate X’s supporters who say Candidate Y is their second choice is closer to Candidate Y’s topline vote share in the poll than it is to 0 or 100 percent. For example, Biden ranks at least third among supporters of all the top candidates — even someone clearly not in his lane like Warren — simply because he is a strong candidate overall. Among Harris and Warren supporters, he even ranks ahead of Buttigieg, who is in their lane but is not polling nearly as well as Biden overall.
In other words, it’s not like all, or even most, voters follow “lane” logic, especially at this stage in the primary when the field is still so crowded. There are also many drawbacks to lane analysis, such as voters not knowing enough about candidates to taxonomize them, voters not universally perceiving candidates in the same way, a candidate’s “brand” changing over time, and the bandwagon effect of winning an early primary or two. And because there is so much subjectivity in identifying lanes, they don’t always accurately predict where a candidate’s support will go after he or she falls out of fashion or drops out of the race. Lanes can be a useful way to analyze support in a presidential primary — just don’t rely on them too much.
From ABC News: