Branden Camp / Getty

Last Sunday, Marco Rubio voiced the conventional wisdom that guides much horse-race commentary about the GOP campaign: “Part of the dynamic up to this point,” Rubio declared, “is Donald [Trump] has been, you know, in the mid 30s to low 30s, high 20s, in most polls, and then you have 70 percent of the Republican electorate that says, ‘We’re not voting for him.’ But they’re divided up among five or seven people. So as that five or seven people continues to narrow down, I think it’s going make the race clearer and clearer.” Ted Cruz has said much the same thing: “Donald Trump … has a passionate, committed base of supporters, but he’s got a ceiling—between 60 and 70 percent of Republican primary voters.”

Unfortunately for both Rubio and Cruz, they’re wrong. The idea that most Republican voters reject Trump, and what he stands for, may be comforting to GOP elites. But it’s just not true.

Start with Rubio and Cruz’s contention that 60 or 70 percent of Republican primary voters won’t back Trump. Once upon at time, that was correct. When NBC and The Wall Street Journal began asking likely Republican primary voters if they could support Trump last March, 74 percent said no. When he announced his candidacy last June, 66 percent still held that view. But by September, with Trump atop the polls, the percentage of Republicans ruling him out had dropped to 52 percent. The last time NBC and the Journal asked the question, in January, the percentage who could never support Trump was a mere 34 percent—not much higher than the percentage who said they could never support Rubio or Cruz.

Polls in individual states tell a similar story. In a recent piece on FiveThirtyEight, Nate Silver noted that “about half” of Republican primary voters in New Hampshire said that “they’d be unhappy with [Trump] as their nominee.” That’s true: 46 percent of New Hampshire Republican voters told exit pollsters they would be “dissatisfied” if Trump won the nomination. What Silver didn’t mention is that the percentage that said they’d be dissatisfied if Rubio or Cruz won was even higher.

The point is that Trump’s “ceiling” is neither particularly low nor fixed. A mid-February Economist/YouGov poll found that 62 percent of likely GOP voters view Trump favorably, which is virtually identical to the percentage who approve of Rubio and Cruz.

In Ohio, according to a recent Quinnipiac poll, Trump’s favorability is 57 percent. In New Jersey, it’s 58 percent. And in deep-blue Massachusetts, it’s 64 percent. Trump is leading in all these states, but the percentage of Republicans who view him favorably far exceeds the percentage that currently support him, which suggests that he has room to grow.

In fact, in states like New Jersey and Massachusetts, which boast more moderate GOP electorates, the top-tier candidate with the lowest ceiling is not Trump. It’s the ultra-conservative Cruz. In New Jersey, Trump’s favorability-to-unfavorability margin is plus 28 points. Cruz’s is minus seven. In Massachusetts, Trump’s margin is plus 32 points. Cruz’s is an astounding minus 27. That high an unfavorability rating makes it hard to imagine Cruz winning 50 percent in the Bay State in even a two-person contest. Trump, by contrast, is polling at 50 percent in Massachusetts now—and that’s with four other candidates in the race!

Yet another way to measure Trump’s potential support is through hypothetical matchups in a narrowed field. But this doesn’t confirm the “hard-ceiling” thesis, either. I found three national polls that asked Republicans how they would vote if only Trump, Rubio, and Cruz were in the race. One had Trump in front by 11 points, a second had him ahead by 10 points, and a third had Rubio leading by one point. In two-way hypotheticals, Trump’s numbers are worse. An early February poll by Public Policy Polling showed Trump losing to Rubio in a one-on-one matchup by 12 points and losing to Cruz by 6 points. In a mid-February NBC/Wall Street Journal poll, Rubio and Cruz had an even larger edge. But that poll is something of an outlier: It’s the only national survey since last October not to show Trump leading overall. In a January NBC/Wall Street Journal poll with horse-race numbers that more closely track other polls, by contrast, Trump still lost a hypothetical matchup to Cruz, but he beat Rubio. In hypothetical one-on-one matchups in South Carolina, Trump defeated Cruz and Rubio (narrowly), as well.

Looking at who supporters of the other candidates name as their second choice also helps explain why Trump’s support doesn’t collapse when the field narrows. While Rubio voters tend to choose Cruz as their second choice, and John Kasich voters choose Rubio, many Cruz and Carson voters say their second choice is Trump. It’s the same in state polls. A new poll in North Carolina, for instance, finds that both Cruz and Carson voters pick Trump as their next favorite. Rubio voters choose Cruz but still prefer Trump to Kasich. Most remarkably of all, Trump almost ties Rubio among former supporters of Jeb Bush.

Trump’s not doomed in a two-way race because he’s not a niche candidate. In New Hampshire, he won women and men, young and old, college graduates and noncollege graduates, rich and poor, conservatives and moderates. In South Carolina, he narrowly lost college graduates and voters who called themselves “very conservative.” But he still won among voters of every age, income level, gender, and issue concern. His support was similarly broad-based in Nevada.

Republican voters also agree with Trump on many of his signature issues. Almost three-quarters want the United States to build a wall along its southern border. Almost 60 percent support a ban on Muslim immigration. (Among GOP voters in New Hampshire and South Carolina, the figure was even higher.) A plurality opposes the Trans-Pacific Partnership trade agreement. Most believe America’s campaign-finance system is corrupt.

In today’s Republican Party, it’s not Trump who’s outside the mainstream. It’s the consultants, donors, and pundits who fear him. He’s not bumping up against the ceiling of his support. They are.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.