In a beige reception hall in a Des Moines suburb, over paper plates piled with the remains of a Monday morning continental breakfast, Sen. Bernie Sanders urged a packed house of Iowans to manifest their dreams. Imagine an America where cancer only kills you, rather than also rifling through your wallet. Visualize a future where no American child has to pay off her grandmother’s student loans. Cynicism is high and more than a quarter of us believe the American Dream is unattainable, but Sanders’s stump speech offered hope. “Everything is impossible until it’s not,” he said. The crowd went wild.
This speech was about issues but it was also a pitch for the semi-improbable Sanders campaign, itself. Before you stands a 78-year-old Jewish man, a self-described democratic socialist and an independent interloper in the Democratic Party, who is making his second try at the presidency after 30 years as a professional gadfly.
Yet here we are, five weeks later, and that same old People’s Grandpa is the candidate most likely to win a majority of delegates in the Democratic primary. Everything is impossible, after all, until it’s not. But the pundits say Sanders is risky, and the pundits are honorable men. Already, you can find headlines full of concern that even if he was viable in Iowa, Sanders won’t be able to win when it really counts. In other words: Sanders isn’t electable.
This whole concept of electability is frustrating to voters like Brooks Vander Kopsa of Granger, Iowa. Standing in the back of the crowd, he told me who is and isn’t electable makes no sense to him — but he’s not even sure it matters. “There’s all this talk about electability. ‘Oh, this person is so much more electable than that person.’ But when I look at policy and I look at track record, I don’t know who is more electable than Bernie,” he said. “So, electability. I guess I don’t know what that is.”
Truth is, nobody does. For all the hands we’ve wrung dry over it in recent elections, electability isn’t a thing you can measure. It’s subjective, not objective — which is why Sanders isn’t the only candidate whose persona can be twisted one way to fit a narrative of unelectability, and another to tell a story of certain success. (Sen. Elizabeth Warren can attest to that.)
Political scientists study electability, but electability ain’t no science. Instead, researchers say, it’s basically a layer of ex post facto rationalization that we slather over a stack of psychological biases, media influence and self-fulfilling poll prophecies. It’s not bullshit, exactly; some people really are more likely to be elected than others. But the reasons behind it, and the ability to make assumptions based on it, well …
“[Electability] is this vague, floppy concept,” said Nichole Bauer, a professor of political communication at Louisiana State University. “We don’t know who is electable until someone is elected.”
“I’m not sure I’m who you want to talk to,” said Julie Brown of West Des Moines, arching her eyebrows and flashing the Elizabeth Warren button hidden under the flap of her canvas purse. She came to the Sanders rally with her teenage daughter, curious to understand why he was polling better than her favored candidate. As Sanders proxies worked the crowd, we huddled against a wall, talking about the ways electability and psychological biases overlap. “I think he is electable and that frustrates me,” she said. “It frustrates the female inside me. If Elizabeth Warren had had a heart attack, they would have put her six feet under.”
Determining who is electable inevitably pits candidates against each other, especially in an election year when the top priority for primary voters — by a long shot — is nominating someone who can take down the sitting president. Brown is a voter who sees “electability” as basically a reflection of whether a candidate can clear the hurdles presented by the electorate’s prejudices.
Months of talking about the primary — and wondering whether candidates will eventually win the general election — has made electability a hot buzzword of the 2020 election. But, scientists say, we’ve not put as much work into clearly establishing what it is.
When physicists suspect a thing exists, but can’t observe it directly, they start studying the stuff around it. You can’t see the particles, you can’t look at the black hole, but you can see what happens when they crash into something else. And that’s basically what political scientists have ended up doing with electability. To understand it better, researchers have looked at a couple of different kinds of social collisions: What voters like in a politician, and what those voters think other people like.
And, in that way, Julie Brown isn’t wrong about electability and bias, Bauer told me. Social scientists do use voters’ biases to understand what electability is and what it might look like.
A lot of this comes from experimental studies — contrived situations where researchers present participants with information about hypothetical candidates and ask them questions about how likeable that imaginary person is, or how much leadership ability they assume the candidate would possess. It’s not the real world, but it does tell us something. Specifically, Bauer told me, voters’ conception of who can get elected appears to be based on who has been elected in the past. “And we always think about men,” Bauer said.
For example, men generally have lower pitched voices than women — and there’s a lot of research suggesting that people are more willing to vote for somebody whose voice pitch is more, well, manly. In a 2016 paper, researchers made recordings of five men and five women speaking the same sentence: “I urge you to vote for me this November.” They played these recordings for 393 men and 411 women, all of whom were participants in the Cooperative Congressional Election Study — a nationally representative survey that’s used to track all kinds of voter behavior and opinions.
Participants were randomly assigned to listen to either five pairs of male voices or five pairs of female voices, and were asked which of each pair they’d prefer to vote for. Across the board, participants preferred to vote for the candidate with the lower-pitched voice, regardless of if that candidate was male or female. And the effect was clearer for participants over 40 — you know, the people most likely to turn out to vote.
But it’s not like someone’s voice means much when it comes to actually governing. The people who did this study of voice pitch later went back and analyzed whether the voice pitch of sitting members of Congress correlated with their legislative activity, the holding of leadership positions or their influence in setting legislative priorities. Lo and behold, having a deeper voice does not make you a better politician. Voters just apparently sorta think it does.
Studies like this run somewhat counter to actual electoral outcomes, though, said Cindy Kam, professor of political science at Vanderbilt University. Yes, studies suggest that voters hold female candidates to higher standards than their male counterparts — women who get elected to public office tend to be more qualified for the jobs they hold than men who get elected, for example. And women are significantly underrepresented in public office. But that’s not the whole story because, while biases exist, women who do run seem to do about as well as men when it comes to getting elected.
Racial bias, on the other hand, more clearly factors into outcomes of who actually wins elections, Kam and other experts said. Studies have found that white voters see black and Latino candidates as more ideologically extreme and less competent. There’s also evidence white voters resist coming out to vote for black candidates even when they share an ideology with that candidate. And black women still rely on the black electorate to win their races.
Even Barack Obama, who won the presidency, probably didn’t get the votes a similarly positioned white candidate would have. In their 2012 book “The End of Race?” political scientists Donald Kinder and Allison Dale-Riddle analyzed voter data from the 2008 presidential election. Based on party identification, you’d have expected any candidate put forth by the Democratic party that year to pull in 55.5 percent of white voters. Instead, Obama got 43.3 percent of the white vote. He won the presidency, but with lower enthusiasm and turnout among whites than a similar white candidate would likely have had, Kam said.
So it’s fair to say that our notion of electability is, at some level, related to our individual knee-jerk social biases — things like the color of a person’s skin, or the way they present their gender to the world. We take those ingredients and we make assumptions about that person. We make assumptions about what other people might think about that person. We make assumptions about what researchers want us to say when they ask about our biases. We make a stew — reactions and reactions to reactions. It’s virtually impossible to avoid bias in perceptions of electability, said Alan Abramowitz, professor of political science at Emory University. “Just about anything that affects how you feel about a candidate could affect assessment of electability,” he said.
Media narratives, in turn, often prey on these biases, which only makes them stronger. In lifting up electability as a marker of fitness, we’ve inadvertently created a system that caters to whatever our imagined lowest common denominator might be. You might want to vote for a black, female candidate, goes the narrative … but other voters are racist and sexist and so you can’t.
Because, of course, electability isn’t just about individual feelings.
When voters like Julie Brown and Brooks Vander Kopsa talk about whether Bernie Sanders is electable, they aren’t really talking about their own feelings. They’re talking about what they think other people feel, which is where polls come in.
“The average person knows a little about politics, but not a ton,” Stephen Utych, a professor of political science at Boise State University, said. And voters use polls as a source of information to fill in the gaps. “If I’m a Republican and other Republicans don’t like this person, I don’t know what it is, but there must be something wrong with them,” Utych said. We American voters really like to believe we’re independent, Kam agreed, but the reality is that we take a lot of cues from the herd.
But polls can become a bit of an ouroboros. Kam and Utych’s 2014 study found that candidates who were behind in the polls were rated less favorably by voters — and voters were less interested in seeking out information about those candidates.
The interaction of polls and media becomes its own self-fulfilling prophecy, Abramowitz and Utych both said. And candidates can shift the perception of how electable they are by striking back at the media and crafting their own narratives. In a 2018 study, the share of voters who, after reading a candidate’s defense of their own electability, were willing to think the candidate could win the election more than doubled, rising from 15 percent to nearly 34 percent.
This early in the election season, there’s still an opportunity to change the narrative – to grasp electability out of the jaws of defeat. And that’s the paradox that leads candidates like Sanders to spend months traversing the early primary states – breakfast to breakfast, handshake to handshake. Winning Iowa allowed Barack Obama to craft a narrative of electability around himself in 2008. Conversely, Bill Clinton lost Iowa and took second place in New Hampshire in 1992. But, from that, his campaign was able to spin a narrative of being the “comeback kid”, said Seth McKee, a professor of political science at Texas Tech University. “I think Iowa and New Hampshire matter so much in how the media portray the horse race after the votes have been cast,” he told me.
But building those narratives and harnessing those horses are dependent on the idea that voters have a good idea of what other voters want, or what other people’s deal breakers might be. And the psychology gets very tricky here. Frankly, experts said, voters aren’t great at knowing what’s going on in their own heads, let alone those of strangers.
A June 2019 Ipsos poll, for example, found that 74 percent of Democrats and Independents said they’d be comfortable voting for a female president, but only 33 percent of those same people believed their neighbors would be as open-minded.
The effect captured in that Ipsos poll is so common, social scientists use it in their research to make sure participants aren’t just blowing some woke-sounding smoke. “People aren’t stupid,” Bauer told me. If you just ask who they like as a candidate, they’ll figure out that you’re trying to see if they’re sexist. “But asking if you think they’ll win is asking if you think other people will vote for that candidate. It takes social desirability pressure off the individual.” But when polls turn up results like that, are they showing that Democrats are secretly more sexist than they let on? Are they showing Democrats are unfairly contemptuous of their fellow Americans? Maybe a little of both? It’s hard to say, but it does demonstrate how hard it is to predict electability based on what you think other people think.
Then there’s the issue that electability is not a fixed idea. What makes a candidate likable to the nation, as a whole, is in flux — tracking, experts say, with hardening partisan lines.
And voters see it, too. James Muhammad, a Californian visiting Iowa, was one of the other people I spoke to at the Sanders rally. When I asked him about electability, he just laughed. “Was Trump electable?” he said.
That’s a question academics are also asking. And it’s one that’s deeply tied up in attempts to understand what electability looks like to Democrats now. From what we can see in research on congressional races, which are more numerous, there’s something about electability that is shifting. Something fundamental.
“I think there is an idea in the media of a centrist, usually white, not necessarily college educated voter who is the one at play and that probably has influenced the way the media is covering it,” said Joshua Darr, a FiveThirtyEight contributor and professor of political science at Louisiana State University. That assumption of the power of the centrist voter is, to some extent, evidence based. Historically, being moderate and appealing to centrist voters was a great way to win congressional elections, Utych and Abramowitz both told me. But that’s been changing. Abramowitz’s analysis of the 2018 House elections turned up evidence that an incumbent candidate’s past voting record — whether they were more moderate or not — didn’t really make much of a difference in whether they won or lost, regardless of party. What’s more, he told me, the number of moderate members in Congress has been falling for decades. Forty-eight percent of the 95th Congress (1977-79) fell within the moderate range of ideology,1 compared to just 16 percent of the 115th Congress (2017-19), Abramowitz found.
Ideologues are elected more often than they used to be. Outsiders are elected more often, too. And the percentage of true swing voters is shrinking, Utych said. So does that mean someone like Sanders is more electable and someone like former Vice President Joe Biden is less electable? Electability here becomes a game of divining which group is more important to winning — swing voters or the partisan base. But that’s no more accurate than trying to estimate how sexist your neighbors are. “Which segment is bigger … there’s not great information on that,” Utych said. “Anything you say is just guessing.”
Even attempts to pin electability down subjectively leave you chasing your own tail, said Elizabeth Simas, a professor of political science at the University of Houston. We know from decades of research that voters have a tendency to line up their assumptions about who is electable line with the person they want to be elected. Maybe that means people just want to maintain some kind of cognitive consistency. “But it’s just always going to be impossible to parse out whether someone supports a candidate because of electability, or if a candidate is perceived as electable because they are the preferred candidate,” Simas said.
And there’s no better place to see that ambiguity than at a primary campaign rally. Skirting the edges of a cheering crowd, Brown and Vander Kopsa basically both want the same things — a candidate who cares about average people, a candidate who will be a game-changer and think outside the box. They both suspect other voters aren’t engaged or doing the research necessary to know who meets those criteria. What they don’t agree on is whether Bernie Sanders is inside the box, or out of it.
As measured by DW-NOMINATE scores.
Maggie Koerth is a senior science writer for FiveThirtyEight. @maggiekb1