Four years ago, Newt Gingrich led national polls of Republican voters, nearly 28 percent of whom indicated they supported the former House speaker. Mitt Romney was close behind at 24 percent, however, and in Iowa polls of prospective caucus-goers suggested that Ron Paul might beat them both—Paul getting 22 percent, Romney 21 percent, Gingrich 14 percent.

The polls were wrong. Gingrich would only win South Carolina and his home state of Georgia. He didn’t even make the top three in Iowa, where Romney and Paul placed second and third—and the surprise victor, who nine days before had been polling behind Rick Perry and Michele Bachmann as well as Paul, Romney, and Gingrich, was Rick Santorum. He, not Gingrich, would be Romney’s toughest competitor for the rest of the campaign, winning 10 more contests after New Hampshire.

This year Donald Trump polls better than Gingrich did in 2012, both nationally and in the early states, and no establishment contender in 2016 has the support that Romney did last time. But Santorum’s 2012 upset might still tell us something about what to expect on February 1. Santorum’s performance showed that Iowa voters were even more focused on social issues than pollsters and pundits had realized. Religious right caucus-goers voted their consciences, and when they asked themselves whether Romney or Gingrich or Ron Paul best represented their views, they disregarded the choices that polls and the media had given them and voted for Santorum instead.

This could be good news for Ben Carson, whose moribund campaign actually polls better in Iowa today than Santorum’s did nine days before his win in 2012. Then again, Trump has advantages this year that Gingrich did not have in 2012—most importantly, the endorsements of Phyllis Schlafly and Sarah Palin. But if the Santorum vote in 2012 really was a conscience vote, I have a hard time believing that Iowa evangelicals in their hearts of hearts identify more with Trump than with Carson. These voters—as their support for Santorum in 2012 and Huckabee in 2008 demonstrates—know who they are and what they believe in. They are rock steady, not swayed by media buzz or the showbiz glitz of a modern presidential campaign. They are also organized: their churches and religious groups are ready-made battalions that translate directly into political strength.

The left-wing canard about the religious right is that its adherents are—as a Washington Post reporter famously wrote—“largely poor, uneducated, and easy to command.” And among the non-evangelical right, there’s a reluctance to admit that Christian conservatives really do have fundamental differences with other Republicans. Elite centrists in the GOP who do indeed recognize the religious right as a separate species from themselves nevertheless tend to perceive the evangelical vote as basically populist rather than distinctly values-driven.

Left, right, and center thus all run the risk of underestimating how focused and distinct Christian conservatives can be, especially in an environment like the Iowa caucuses, where such voters predominate in numbers great enough that they need not worry about coalition-building. The Iowa caucuses are, in fact, practically the only opportunity these voters have to “send a message” in national politics—to testify in action to what they really believe.

That doesn’t mean that a surge for Ben Carson (or much further down in the polls, Mike Huckabee or Santorum himself) is the only possible outlet for evangelicals’ political intensity. My guess is that there is at least as good a chance that this intensity will lead to a bigger than expected turnout for Ted Cruz, who seems to have been successful at positioning himself as a plausible avatar for the religious right. An unexpectedly big win for Cruz could put a dead stop to Donald Trump’s momentum, robbing him of the limelight in which he has flourished and dampening, if not overcoming, his support in New Hampshire.

To be sure, the Santorum 2012 vote isn’t the only group that might yield a surprise in Iowa’s caucuses, where turnout is small enough that intensity and organization can trump—if you’ll pardon the expression—media exposure and general polling. Rand Paul has staked his campaign on getting younger and more libertarian Iowans to caucus; he’s confident in his ground operation, and he argues that his voters are not counted by conventional polling.

But if the religious right vote is historically even stronger in Iowa than analysts expect based on polls, the opposite is true of Iowa’s Paul-family vote: in 2012 Ron Paul underperformed in the caucuses relative to his polling beforehand. He had a shot at first place but finished third. No matter how good Rand’s ground operation may be, the supply of libertarian-minded voters in Iowa is simply unlikely to be enough.

Marco Rubio, meanwhile, has to hope that a large slice of Romney’s 2012 Iowa vote (roughly 25 percent of caucus participants) finds Trump and Cruz alike unacceptable—and that those voters pragmatically opt for Rubio instead of Jeb Bush. My hunch, however, is that the conventional wisdom is backwards: Christian conservatives are much more certain about how they want to vote than pundits give them credit for being, while middle-of-the-road Republicans are much more indecisive and impulsive this year.

(For Trump, the worst case scenario would be for the conventional wisdom to be half-right: a rally of electability-focused centrists to Rubio, coupled with stronger-than-expected evangelical support for Cruz. That would almost certainly set up a Cruz-Rubio race for the rest of the primary season, leaving Trump where Gingrich was in 2012.)

Donald Trump has proved all skeptics wrong so far. Maybe he really has changed the nature of Republican presidential politics, such that precedents from years past no longer apply. Maybe. But that notion has yet to be put to the test that counts—in real presidential primaries and caucuses.