You can’t talk about a U.S. national election — especially the Republican side of it — without a hefty discussion of what evangelicals want. But in the hurry to answer that question, the most basic of questions gets ignored: who are evangelicals? That definition can vary from person to person, or even from pollster to pollster. And at the center of it all is a term that, for all the attention it gets, is remarkably poorly defined.

How do you define it?

Here’s how squishy the term “evangelical” is: depending on the method of measurement, more than one-third of Americans are evangelical, or fewer than one-in-10 are.

That huge range comes from the different ways pollsters and other social scientists define the term. In a lot of surveys, a pollster simply asks people how they identify, often adding on the question of whether someone has been “reborn” as a Christian: “Do you consider yourself an evangelical or born-again Christian?”

According to the Pew Research Center, around 35 percent of American adults (that is, roughly half of all Christians) consider themselves evangelical or born-again. So when reporters and politicians talk about “evangelicals,” it can sound like they’re talking about a huge chunk of the population — more than a third.

But then, other national political pollsters, like CNN/ORC, add a modifier onto most of their evangelical polling, focusing on white evangelicals. (And this is the group most pundits are talking about, particularly when it comes to Republican primary politics.)

The idea, said one survey researcher, is to avoid lumping groups with clearly distinct political ideas into one bucket.

“White evangelical protestants are some of the most reliably conservative and Republican voters in the electorate,” said Greg Smith, associate director of research at Pew. “African-American protestants, on the other hand, are some of the most strongly and consistently Democratic voters in the electorate.”

“If you didn’t look at them separately,” he added, “if you lumped them all together, you would miss a big part of the story about the connections and the interrelations of religion, race, and politics in the U.S.”

Cut that pool of evangelicals or born-agains to white, non-Hispanic evangelical Protestants only, and they account for 19 percent of Americans, according to Pew’s data.

Beyond self-identification, there are more exacting ways of defining the group. In fact, Pew has two ways of counting evangelicals. In addition to asking people to self-identify, it sometimes uses a denominational system, creating a dividing line between “evangelical” Protestant denominations, like Southern Baptists, and “mainline Protestants,” like Methodists (“historically black” Protestant churches are in a separate category). By this definition, around 25 percent of Americans are evangelical.

Definitions can get even tighter — and with them come smaller estimates of evangelicals. The Barna Group, a research firm that specializes in religious issues, uses what may be the toughest definition of evangelicalism out there. It asks a series of nine questions about beliefs (Did Jesus lead a sinless life? Does salvation come from “grace, not works”?). Only 6 percent of Americans are “evangelical” by Barna’s definition, according to their latest count.

The entanglement between race and religion

Because political polls often focus on white evangelical voters (which is in turn in part because those evangelicals — however one defines them — are such a coveted demographic among GOP voters), white evangelicals end up getting a huge amount of media attention. But that means they can end up being portrayed as the face of evangelicalism, period. Indeed, articles about this polling sometimes end up conflating white evangelicals with all evangelicals.

Anthea Butler, an associate professor of religious studies at the University of Pennsylvania, believes that race ends up getting glossed over in the hubbub over the so-called “evangelical vote,” as she said in a February speech.

“The media does this all the time. You never hear them talk about black evangelicals,” she said. “Watch the 2016 election. When they begin to talk about evangelicals again, they won’t go to Bible-believing black evangelicals. They’re going to talk to white people. I know. I’ve watched them do this, and I have argued with people about this over and over again.”

Consider an imaginary pair of evangelicals — one black, one white — who sit next to each other in the pew every Sunday. They could have the same religious beliefs. But as Smith pointed out, they’re likely to have vastly different political beliefs: the black churchgoer is more likely to vote Democratic, while the white one will lean GOP.

(Pew’s polling on black Protestants focuses on that group as a whole, not on black evangelicals themselves. But 82 percent of attendees at historically black Protestant churches identify as or lean Democratic, according to Pew, and 72 percent of black Protestant churchgoers identify as evangelical or born-again. Clearly, a huge share of black self-identified evangelicals also tend Democratic.)

All of which means something important: when evangelicalism comes into the U.S. political conversation, it’s often also a conversation about race. The racial discrepancies in the numbers suggest that identifying as “evangelical” doesn’t necessarily make a person more likely to vote Republican.

The self-definition problem

The question at issue with measuring evangelicals is the question of what people’s religious beliefs mean for their political views.

Part of the problem here is that “evangelical” has a muddled definition, even when you strip away the politics and survey research.

“The term ‘evangelical’ has a very broad set of meanings in Christianity. In its origins, it refers to the evangel, which is a Greek word from the New Testament that refers to the ‘good news,’ or the gospel of Jesus Christ,” said John Green, professor of political science at the University of Akron and an expert in the intersection of politics and religion, in an August interview.

“In some sense, all Christians have an element of being an evangelical, because they all share to one degree or another those basic Christian beliefs,” he added.

Still, a few people and groups have tried to lay down some clear borders around evangelicalism. One of the better-known definitions (among religious scholars) comes from David Bebbington, a professor of history at Scotland’s University of Stirling, who identifies four key traits of evangelicals. Those are, in turn, similar to National Association of Evangelicals’ own definition. That definition itself has four parts — four beliefs that a person must have in order to claim evangelicalism. Under NAE’s rubric, an evangelical believes that the Bible is their “highest authority,” for example, and that it’s important to spread the word to non-Christians.

That NAE definition is the “most widely accepted definition” of evangelicalism, as the Atlantic’s Jonathan Merritt wrote earlier this month.

That’s not how specifically everyone defines their own evangelicalism, though. According to Jocelynn Bailey, who attends Centreville Baptist Church in Centreville, Va., it’s about evangelizing.

“What I think when I think ‘evangelical’ is, ‘I have good news about what I believe Jesus did for me on the cross, and I want other people to have that good news and that hope,’” she said, speaking in September. “An evangelical is someone else who desires to share that.”

One of her fellow parishioners, Tim Lemieux (himself a self-identified evangelical), had a different take about what’s most important for an evangelical.

“I define evangelical as someone who lives based on the beliefs of God and his authority for his purpose and his desires,” he told NPR in September.

It’s not that parishioners everywhere are likely to carry the same long, exacting definition in their heads. But Bailey and Lemeiux’s differing definitions are a subtle sign that the meaning of “evangelical” is different from person to person, making it a tough thing to measure.

“The term ‘evangelical’ is squishy because people use the term differently,” Green said in an email. “This is not uncommon — think of words like ‘middle class,’ ‘moderate,’ or ‘extreme.‘” (Indeed, in one recent survey, 87 percent of Americans saw themselves as some form of “middle class.“)

Consider that a Catholic could easily believe in spreading his or her faith, as Bailey does, or leading a godly life, like Lemieux does. And, indeed, Catholics will sometimes self-identify as “evangelical,” according to Smith. But by many religious or denominational definitions, Catholics are not evangelicals.

Even within the confines of Protestantism, “evangelical” does not always mean evangelical. Members of the Evangelical Lutheran Church in America — the largest Lutheran denomination in the U.S. — are mainline protestants, according to Pew’s denominational definition.

To add to the confusion, here’s another wrinkle: Missouri and Wisconsin Synod Lutherans are considered evangelical. (Another curveball: they don’t necessarily go to church in Missouri and Wisconsin.)

There’s one additional problem with the self-definition method, according to David Kinnaman, president of the Barna Group.

“The notion of a survey question asking, ‘Do you consider yourself to be an evangelical or born-again [Christian]?’ offends me as a researcher because it is a double-barreled question,” he told NPR in October. “It has two very vague concepts.”

He’s not the only social scientist complaining about this: Pew demographer Conrad Hackett has likewise complained about this way of wording the question: it “implies that ‘born-again’ and ‘evangelical’ are interchangeable labels, which may not be true for all respondents,” he wrote in 2008. “It does not offer respondents alternate ways of expressing religious identity, which no doubt inflates estimates of the evangelical population.”

Politics may be blurring the lens

You could dismiss this all as pedantry — that using “evangelicals” as a catch-all term for a certain group of Christians is a harmless shorthand, like calling all tissues Kleenexes or all sodas Coke.

But then, consider how pollsters and pundits often separate white and black evangelicals based on their political views. That’s one piece of a bigger problem: the degree to which “evangelical” may be becoming redefined by its political associations.

“While evangelical, in this traditional sense, is really a religious word,” Green said, “it’s become very strongly associated with Republican and conservative politics, because since the days of Ronald Reagan up until today, that group of believers have moved in that direction politically.”

Indeed, that association has grown stronger in the last couple of decades. In the late 1980s, around one-third of white evangelicals identified as Republican, according to Pew. Earlier this year, Pew found that 68 percent of white evangelicals do.

“For the most part, the concept of being an evangelical has been used so much within the last three to four election cycles … as a key demographic that we find that there’s a lot of perceptions that the term evangelicals means ‘Christians who vote Republican,’” said Kinnaman.

That means American culture may be moving toward a mushy, self-reinforcing idea of who evangelicals are. The term becomes not a nuanced religious concept but a flat heuristic for the idea of “politically conservative Christians.” If this is indeed how some Americans view evangelicalism, their responses to pollsters would border on meaningless — at least, in terms of measuring the relationship between religion and political leanings.

“It may very well be that when people hear those words, if they have conservative perspectives, they may feel, ‘That’s my group, maybe I identify with that group,’ whereas that may not be an accurate measure of their religion,” Green said.

So why measure?

1976 was the first year Gallup asked Americans if they had been “born again,” as Hackett wrote in a 2008 paper. The organization’s measurement methods varied over the next decade, but in 1986, the organization first asked the “born-again or evangelical” question that it uses today.

Over that time, self-proclaimed born-again Christians and evangelicals helped reshape the political landscape. In 1976, the born-again former Georgia Gov. Jimmy Carter was elected to the White House. After that, political interest in evangelicals and born-again Christians remained, but Rev. Pat Robertson’s 1988 second-place finish in the Iowa caucuses in particular made it clear that white evangelicals were swinging Republican. Outspoken Christians like George W. Bush continued the trend of winning over these conservative Christians, and targeting those voters is still a key campaign strategy for politicians like Mike Huckabee and Rick Santorum.

Green acknowledges that it’s a hard term to pin down, but he believes there’s real value in studying evangelicals.

“A lack of common definition doesn’t mean that the realities behind these terms are unimportant — just that measuring the realities is challenging,” Green added. “Behind these definition issues are real groups of people with distinctive values and behaviors. The trick is how to define measure the group of people accurately.”

Still, as with the term “middle class,” it’s possible that people’s self-definition is so clouded that it’s obscuring what’s really going on in the intersection between American religion and politics.

And Kinnaman believes there is one other danger in the range of measures of evangelicalism out there — the more ways there are to measure this group of people, the more opportunities there are for spin.

“For different purposes I have found that evangelical leaders might say, ‘We’re so small and such a small minority, and we’re overlooked, and woe is us,’ and other times they might say, ‘Don’t forget about us! We’re huge and we’re as many as a quarter or 40 percent of the population,’” Kinnaman said. “It’s easy to be elastic about these numbers when they suit our purposes.”

The most obvious lesson from any of this is that political reporters and readers need to know what they’re looking at when they’re reading news about “evangelicals.” Green and Smith both agree on this point — because surveys can be done a few different ways, those paying close attention to the results need to know that “evangelicals” are not always evangelicals.

“From a certain point of view, any kind of information is probably better than nothing, but we have to be very careful when we interpret these findings,” Green said.

Of course, to Christian voters themselves, the term itself isn’t what matters; it’s how politicians relate to them. Just as “evangelical” has been reduced in some political rhetoric to “conservative Christian,” some self-identified evangelicals fear being treated as one-dimensional Bible-thumpers.

For her part, Jocelynn Bailey’s top issues include national defense and her self-described constructionist view of the Constitution. And based on those issues, she says Kentucky Sen. Rand Paul is currently her top choice as a presidential candidate. So when she hears that a particular politician is courting the “evangelical vote,” she bristles.

“It frustrates me, to be honest, because I think that I’m more than just that,” Bailey says. “Certainly that flavors the way I would vote, but I want them to tell me who they are, and all of who they are, not just the stuff that they think I might want to hear.”

She added, “My vote is about more than my faith.”

— via NPR