U.S. Rep. Ron Paul (R-Texas), a GOP presidential candidate, speaks to a crowd during a town hall meeting Sept. 27, 2011 at the Northeast Iowa Community College east campus in Dubuque, Iowa. (Jessica Reilly)

Ezra Klein: Iowa is not a rich state. It’s not a big state. It’s not a traditionally powerful state. So how did it get the enormous gift of kicking off the presidential primary process year after year after year?

David Redlawsk: By accident, is the answer. It’s sort of a funny story. In 1972, the revised Democratic rules coming out of the 1968 debacle required that notice be given of caucuses and primaries that would select party delegates. Prior to that, party bosses could schedule primaries without telling anyone. But the new rules required a 30-day requirement. Iowa’s system has four parts — the caucus, then the county convention, then the congressional district conventions, then the state conventions — so, to give a 30-day notice for all of them, Iowa had to start advertising early.

The second part, the state convention, is normally held in June. In 1972, they looked and found there were no available hotel rooms in Des Moines on the planned weekend. So they pushed the convention back. And that meant they had to push the caucuses back. And that’s how they ended up in January, in front of New Hampshire. It was not a plan, and in 1972, it made no difference. Edmund Muskie spent about a day there. But in 1976, Jimmy Carter’s campaign noticed Iowa was first and decided to invest some time. He ultimately came in second to uncommitted, but his win got him attention, and ultimately helped him get to the White House.

EK: And how has Iowa retained it? You would think that some other state, desirous of the same power Iowa has, would just leapfrog it.

DR: Iowa’s two parties realized very early on that this was advantageous, so they agreed to hold their caucuses on the same day. And then they wrote into Iowa law that the caucuses must be the first event. New Hampshire’s laws say they have to be the first primary, which is what lets Iowa stay in front. But both say they’ll go as early as they have to go.

EK: So if any other state tries to move its primary forward, Iowa and New Hampshire simply move their events forward even more.

DR: Right. In fact, every cycle we have this question of whether somebody will try and leapfrog or jump ahead and that’s how we get frontloading. The national parties try to penalize those who do that. This year, for instance, Florida is only getting half of its delegates because it violated the rules by trying to go early. So the system stays in stasis through a combination of Iowa and New Hampshire jealously guarding their privileges and the national parties penalizing anyone who tried to change the status quo. No one loves the current system, but no one has come up with something they would like better.

EK: A criticism you often hear of Iowa is that it’s tiny and relatively homogenous and that means, in letting it go first, we’re letting an unrepresentative state exert undue influence on the entire process. What does your research say about that?

DR: The first point I’d make is Iowa is not nearly as unrepresentative as people believe it. My colleague Michael Lewis-Beck did a study and found that on 51 social and economic indicators, Iowa was in the middle on 39 of them.

But even so, no one state can really be representative of the country. What I think we should look at more is that the set of early states are, as a group, fairly representative. The GOP has social conservatives in Iowa, economic conservatives in New Hampshire, and southern conservatives in South Carolina, and if you put them together, you pretty much have the Republican Party. In a sequential system, however, someone has to go first, and no one state will be perfectly representative.

EK: Is Iowa good at picking winners?

DR: It kind of depends on what you mean by picking winners. Probably what Iowa does best is winnow the field — eliminate the also-rans, the ones who just can’t build a campaign. That’s really what Iowa does. It teaches them to build a grassroots campaign. Those who do well get to move forward, and those who don’t drop out. That said, in the last few cycles, Iowa has played a very significant role. There’s no question that it launched Obama. But in the end, it’s not so much winning Iowa as it is generating attention because you beat expectations.

EK: That’s one of the odd things about the way Iowa actually ends up mattering, right? It’s not about winning it. It’s about doing better than you’re “expected” to do. But how are expectations set?

DR: They come from a mix of early polling, fundraising, days on the ground in Iowa, etc. So we can quantify in the models that if you spend, all else being equal, 10 percent more days on the ground, you’ll increase your Iowa numbers by 3 percent. So we can look at where those expectations come from. But that doesn’t quite tell us what everyone will be writing tonight or tomorrow. Some of that is just the media herd effect where someone writes something and then everyone else begins responding. But it’s hard to say if that will hold true this year. We didn’t have to look at Twitter conversations in 2008, for example. One of the challenges of doing political science is we’re usually looking at the past to build our models because we need the data.

EK: And if you beat those expectations, your win isn’t so much in the delegates from Iowa, but the media and political elites who are watching Iowa, right?

DR: Yes, candidates are showing not just voters but the media and party elites something about themselves, and the nature of campaigning in Iowa — particularly its grassroots focus — sends an important message about organizational capacity. But a lot of it is simply that Iowa goes first. Before Iowa, we haven’t seen people go out to vote. When people go out to vote, they sometimes tell us something different than what we thought we knew. That can shift attention. And in our research, the shift in attention can have a big effect on performance in New Hampshire.