Let's Talk About How Weird The End Of The 2020 Democratic Primary Was

A consideration of some of the metrics the media and public uses to judge candidates before elections.

Three sets of legs stand on a podium, the top labeled "3," the middle "1," and the bottom "2"

1. This is the idle time of the four-year campaign grind, the one period of peace in the apparatus we've constructed around electing presidents, even if this period of peace involves a global crisis with horrible consequences. But, since there are no immediate stakes riding on this, here’s a question: How do you know if a candidate is doing well? Right before an election, how do you determine who's winning?

2. I bring this up in the idle period, because it seems like there's a problem along these lines, of knowing who's really contending for elected office. You've probably seen the problems pollsters had in 2020 — a 3.3-point error in underestimating Donald Trump's support, an error that got worse in state polls and continued down the ballot. The polling experts who've studied the errors can't quite explain what happened there, though they have ruled out a variety of explanations. One unprovable theory is that a certain kind of Trump voter just doesn’t answer the phone or survey. This kind of polling error doesn't mean much for sussing out 60%–70% viewpoints among the American public, but it does cast some doubt over a tighter proposition or about a larger field of candidates. Is that systemic error a temporary feature of Trump’s rise, or something more permanent with the activation of a certain kind of mysterious (unreachable, unpollable) voter?

Nate Silver has argued the 2020 polls were more mediocre than historic disaster. More to the point — and this is either reassuring or confusing in terms of making sense of things overall — he argued that pollsters will adapt and change to try and correct what happened, and that modeling may become more important. So will the forecasters and modelers: FiveThirtyEight will no longer give extra weight to live-caller polls, which have been long considered the gold standard approach and treated with deference by the media and people on Twitter, in their model. Still, whether disaster or mediocrity, the point is that things are changing at least a little bit.

3. This maybe doesn't seem like too terrible a problem, or hardly a pressing one, because there is a fair argument that the media obsesses over Who's Up and Down, and The Drama Within and drags everyone else along with them into this vaguely nihilistic space where the only thing that matters is the theoretical scoreboard. Should the media and public let their wills be guided by the polled opinions of others, anyway? It's certainly true that the media and public have trouble dealing with what polls and especially forecasts really provide: probability, rather than prediction. The nature of campaigns is also nebulous; it’s hard to understand, exactly, why one candidate or issue connects with people and another does not, and then 10 or 30 years later, they do. And this kind of obsession with pinning down the nature of the future — like the people in 2022 who will only write about positioning for a 2024 race and about inside tracks and what advisers are privately saying — leads to the byproduct of disregarding the present, like the only reason we elect a president is to elect another one.

4. But take the recent New York mayoral primary. The city employed, for the first time, ranked-choice voting, which complicated things; people could pick up to five candidates on their ballot. Though former police officer and Brooklyn official Eric Adams won both the initial ballot and the final, ranked-choice version, the final ballot got a lot more competitive with two female candidates, one moderate and one progressive, nearly coming back to win. The overall tenor of the little-polled race was confusion, and that probably would have been the case even without ranked-choice. New York voters — particularly more liberal voters — plainly had trouble determining which candidates were in contention and worth their top vote if they wanted a candidate who valued certain issues or approaches to actually be mayor.

5. If you zoomed out to 100,000 feet and made someone produce one sentence about the 2020 election, they might write: Joe Biden, a well-known and decently liked center-left Democrat, was always going to win the nomination and presidency during the worst national crisis in decades.

But the end of the 2020 Democratic primary was totally bizarre.

For the first time since 1992 (and that depends on how you look at that one), an eventual nominee lost the first three nominating contests. And Biden didn’t just lose them, he got crushed. Those crushing losses didn’t even really congeal into a “comeback kid” reverberation — he left New Hampshire before his fifth-place results even came in; then Jim Clyburn endorsed Biden, South Carolina's Black and more moderate white voters voted for him, two candidates dropped out and endorsed him, and without much in the way of ads or increased organization, a few days later he essentially won the nomination in a big Super Tuesday sweep.

The things that the early states are supposed to symbolically illustrate — the pointed enthusiasm and momentum produced by electric events, a top-flight organization that knocks on voters’ doors and brings them out to polling places — didn’t seem to matter in the primary, nor the general election. Insofar as Biden held events after Super Tuesday, they were mostly in quiet gyms and theaters in Wilmington, Delaware, free from spectators; his campaign infamously did not do much in the way of in-person field organizing until the very end (though they were robust online and in ads). 2020 was probably unlike any year we’ll see, though, with a chaotic president whose sheer omnipresence defined the race, and a grim pandemic in which not holding events symbolized responsibility.

Still, most people who saw the Democratic field in person would probably agree that, during the primary, Biden and his early campaign did not match the precision and voter enthusiasm of the three candidates who beat him in Iowa: Pete Buttigieg, Bernie Sanders, and Elizabeth Warren. (Clyburn said as much, telling Biden to talk less and focus in more.)

In South Carolina, Nevada, and particularly New Hampshire and Iowa, if they want to, voters and reporters get a good long look at the candidates in different kinds of settings (town halls, rallies, wandering around in a store, small rooms where microphones are not needed); especially since Jimmy Carter won the Iowa caucus in 1976, this sort of thing has defined early coverage and set up either the narrative of success (Barack Obama, validated once in Iowa then twice in South Carolina) or a narrative to be twisted (Bill Clinton, the comeback kid). If you know a lot of reporters (and this is obviously not most people), they have different reasons for valuing events. Events are good for understanding what a candidate is like, what they’re emphasizing, what the self-selecting attendees connect with, and the kind of people who are turning up at events. Plus, travel is broadly good for meeting and building sources within campaigns. Bernie rallies in 2015 were an early indication that something big, which is still shaping US politics, was happening.

And it's obviously not just events that take on this kind of importance in and of themselves. There are all kinds of metrics that the media and public use to deem a candidate "contending" or "coming back" or "struggling," beyond the human basics of whether their campaign is tearing itself apart or running out of money. Reporters and staffers of all different ideologies assign symbolism to different elements in politics. Like: A strong early-state operation — with the number of volunteers and offices — means a candidate has real enthusiasm and discipline. Or small donors, which you can quantify, are better; even beyond the idea of nixing the smoky backroom, the small-donor marks a candidate with real grassroots support. Or certain kinds of endorsements and validators that carry more weight.

But the picture of Biden in the events wasn’t a decisive one in the 2020 primary. Nor was money. Nor was organization. Instead, the idea that Biden — without big money, momentum, electric events, grassroots support, or a particularly strong early-state record — would be a more electable, serious candidate on the ballot against Trump mattered more to voters. That Biden could win. His national polling, and his strength with older Black voters and more moderate white voters, mattered. In 2016, Trump’s lack of traditional organization didn’t stop his attrition attainment of the Republican nomination. And Hillary Clinton’s less electric events didn’t impede her from ultimately beating Bernie Sanders on the Democratic end, or winning the popular vote against Donald Trump during the general election. So what are we doing spending all this time at events in Iowa and New Hampshire? Do we just have to go somewhere together — candidate, staff, media, and voter — to try and figure out what’s happening?

5a. Counter to all of this: Maybe if Biden had been sharper as a candidate early on, in debates and at town halls in Des Moines, Iowa, the decisive margin through which he ultimately won the nomination would have been clearer earlier. Maybe if the Biden campaign had gone out and knocked on doors more in the general election, the general election margin would have been bigger in battleground states.

A problem with campaigns is you can't run them again and tweak the factors; Joe Biden and Donald Trump will only ever run against each other during a coronavirus pandemic that shut down the entire world during the biggest social justice period in our lifetimes once.

6. But other weird stuff happened in 2020 during the Democratic primary, and probably because of the overall weirdness and despair of 2020, we've sort of let that sail by.

Consider Mike Bloomberg and Tom Steyer for a second. The line on Bloomberg is that he wasted $1 billion for some delegates from American Samoa and little else. Elizabeth Warren took a chainsaw to him on the debate stage. Steyer spent millions for 61,140 votes in South Carolina and a show where Juvenile performed "Back That Azz Up." That's like the punchline.

But what if they were near-misses, rather than mistaken propositions? Bloomberg threw millions at ads and never participated in any of the normal presidential stuff (interviews, early state rallies, debates, events with multiple candidates). If South Carolina hadn't saved Biden, what would have happened on Super Tuesday? Could someone blow through a primary like this with a total money bomb assault? Or maybe that’s wrong, and there’s always someone willing to part with their money in this way.

Consider the more successful politicians on the presidential level of the last five years: Trump, Biden, Sanders, Clinton. With the exception of Sanders the first time (not the second), they all were extremely well-known qualities going in (leave aside whether the public properly knew Trump in 2015).

Let’s also consider Amy Klobuchar. She came in fifth in Iowa. She, however, appeared on television first to frame the narrative of what happened — i.e., that she'd had a good night. She followed it with a typically strong debate, then vaulted into third in New Hampshire, and she ended up as the upswing candidate whose endorsement mattered. Warren did better in Iowa! Buttigieg and Sanders did better than both.

8. There’s an element to this where it’s like a holographic house of cards and removing any one piece might topple the ephemeral structure. For instance: The media deeming a candidate "electable" or "serious" through scrutiny and elevation tends to confer a level of authority, and how people come to those conclusions — and what kind of systemic biases or blindspots play in – has been a major discussion topic the last five years. Beneath that basic layer, that dynamic can work in weird ways; "the media" combines thousands of different disparate individuals and publications, and any one's action can result in concurrence among others, blood-in-the-water competition, or a complete dead end.

For the candidate, achieving electable/serious status can unlock major fundraising but also trigger scrutiny that dismantles a person or campaign. For the supporters (and the stans, which have become an active feature of politics), the awareness of all this has created a life/death imperative to participate in the meta consumption of media that can loop back within the aforementioned ecosystem. The media can also elevate candidates unintentionally or as the result of bad incentives; cable news' obsession with Trump helped create a hurricane in the summer of 2015, assisted by some broader novelty coverage and an unwillingness among Republican candidates to attack him and the voters who were just really into what Trump was selling, even as most people in both parties assumed Trump was neither “electable” nor “serious.”

Some segments of the media continue to under-cover the voting preferences of older and more moderate Black voters — often the deciders when it comes to Democratic candidates — and to treat different kinds of groups as too monolithic. There are already clear and great exceptions to the following, but case in point for the broadening of coverage beyond monolith: Latino Democratic primary voters came out strong for Sanders, and, in parts of the South and Southwest in November, Biden still mostly won Latino voters, but Trump saw major gains over 2016. In addition to reporting changes, a major thing that would help with this in advance of (in particular) the next Democratic primary is significantly more investment in polls with samples large enough to more reliably distinguish the interests and preferences of Black, Latino, and Asian voters. Things can also be fluid, though: One of the weird aspects of the 2020 Democratic primary was that after South Carolina, Biden started winning with the kinds of voters he'd just lost to Buttigieg and Klobuchar.

8. Here are a few things I thought at various points the last six years: that Joe Biden would lose Iowa and New Hampshire and then the nomination, and I thought that beginning in summer 2019 until the second part didn’t happen; that Beto O’Rourke and Cory Booker had a real shot at the nomination; that the media was underrating Bernie Sanders, who would win Nevada then the entire thing, on the strength of Texas and California; that Donald Trump might fade in the Republican primary, even though I’d seen all the polls and the way that critical stories about Trump had little effect on those polls, until I was standing in a room with some Republican voters in the winter before the 2016 Iowa caucus and there was no consensus at all about who they’d vote for, and then it was like, oh, he’s gonna win.

9. Imagine, for a painful second, it’s fall 2023. The presidential primaries are a few months away, definitely in the Republican Party, but possibly for the Democrats too.

And let’s say there are a lot of candidates — like 10 or 12 people on each side, vying for the presidency of our fair, twisted nation. In Phantom 2023, only three or four candidates have reached double digits in polling of early states. But that polling varies a bit state to state — and a lot of that state polling deviates a bit from the national polls. Some of the candidates barely ever speak to the media; some throw electric rallies; some have robust Iowa or South Carolina organizations; many are running out of money. Nobody, in this hypothetical, is a 60%-of-the-vote-guaranteed, romance-with-the-nation kind of a candidate. So how, exactly, will we tell who’s doing well, and who’s “serious”?

The structure of what “contending” looks like can be amorphous and chancy and change over time. Big donors were a sign of cachet a decade ago; wealth taxes were a third-rail idea. The last few election cycles would indicate that it’d be a smart idea to consider what older Black voters or white moderates think on the Democratic end, but what if those groups were perceived to be divided about the candidates? What if there were a few "Trump" candidates, or they were all Trump candidates? What kind of polling would you trust to inform your view? Would you be comfortable betting that certain kinds of ads matter more than appearing in interviews and on TV a lot?

What should matter in politics is a different question from what has mattered when people try to measure success in campaigns. On some core level, polls and the perception of in-person energy have shaped the way media and public consume presidential politics the last few cycles. But, maybe, some of the metrics are getting a little fuzzy, not necessarily in a way that’s going to reshape primaries, but one that nevertheless poses a potential challenge for determining who’s worth the most scrutiny and attention.

And, again, you might reasonably feel that this is sort of a horse race distraction from the deeper issues of policy, because it is. In the wake of the 2010s elections, reporters (and the public) have clearly tried to go bigger with the more issue-thematic and investigative coverage of candidates and the direct impact of potential policy decisions. Which is great. And against the constraint of resources, maybe the only immediate solution is to continually try to become less predictive and more open to the idea that anything could happen while retaining a level of critical awareness.

But it’s also hard to entirely get away from the fact that campaigns cost real money, and real people give money, and real people vote. There are only so many stories any one person can write or read. People often want a risk-calculated understanding of a political field, in part so they can make an informed decision about what might otherwise be a wasted vote or dollar. The question you always hear is: Who do you think is going to win?

Skip to footer