Iowa Caucus Goers are late deciders: Still Plenty of Room for Change

Posted: December 2, 2015 | By: Iowa Caucus Project Staff Tagged: About the Caucuses

By David Redlawsk, Harkin Institute Mabry Fellow

Professor of Political Science at Rutgers University in New Brunswick and Director of the Eagleton Center for Public Interest Polling

It’s just over 60 days to the Iowa Caucuses. Polling continues to show Donald Trump on top of the GOP pack here in the Hawkeye State. But, as I write this, two recent polls show something changing under the Trump umbrella. Ben Carson, who had been running a close second to Trump appears to be falling; perhaps his time is over. Taking his place is Ted Cruz, who has doubled his support to just over 20%, while Carson has fallen below that mark.

Trump himself remains in the 25%-30% range, where he has been stuck for months. Nothing seems to change for Trump in Iowa; all the action is in second and third.

Meanwhile the media remains obsessed about whether Trump’s support is real, or whether it will fade as voters get “serious.” Nate Silver just suggested the media needs to “stop freaking out” over Trump. He argues, as I have since at least August in Twitter comments and on the news app Sidewire, that Trump’s numbers remain stagnant at about a quarter of GOP voters. We’re seeing this in Iowa as well as nationally. In fact, a poll in August had Trump at 23% in Iowa; today he is around 25%.

The basis of Silver’s sanguine attitude toward a Trump nomination is the claim that voters in places like Iowa do not make up their minds until quite late. Silver uses public exit polls from the 2008 and 2012 caucuses to show this.

Here I want to reinforce that point, using a completely different dataset. In 2008, Caroline Tolbert, Todd Donovan (my co-authors on Why Iowa?), and I administered an in-caucus survey of both parties. This is NOT an exit poll. Instead, with the cooperation of both parties, we placed a single survey instrument in EVERY Iowa precinct with instructions to the Caucus Chair to give it to the person whose birthday was closest to a random date on the packet. This allowed us to randomize the survey and also to potentially cover every precinct in Iowa. While we didn’t get them all back, our return rate was over 60% for the GOP and over 70% for the Democrats.

So what can we learn from the GOP and Democratic caucus goers of 2008, the last time both parties had wide open nominations?

First, YES, Iowans do not rush to make Caucus decisions. Across both parties in 2008, 54% of those filling out the survey told us they had made their candidate choice only in the final month, and 5% came in the door that night undecided. Just a quarter had decided at least three months before the caucuses.

And first-time attendees were not any faster or slower making up their minds: 56% had done so in the final month, compared to 52% of repeat attendees, an insignificant difference. Not surprisingly though, those who had caucused before were a little more likely to be early deciders: 27% decided before October 2007, compared to 22% of first timers.

We also see no differences in gender – men and women were equally likely in 2008 to make late decisions.

Of course, some Iowa Caucus goers are party activists, but in 2008 a surprising number (nearly 60% in our survey) were not, something I would expect will be the case again this year. No one should be surprised that activists make up their minds earlier: across both parties nearly 60% of the most active has made a decision more than a month before, as had over half of those who called themselves “somewhat active” in their party. In contrast only slightly more than one-third of less active voters made an early decision.

While the above combines both parties, we can dig further into our data and look for differences between Republicans and Democrats. What we find is that in 2007-08, GOP voters were slower to decide. While 47% of Democrats with a preference entering the caucuses, waited until the last month to decide this rises to 62% of GOP voters that year. Nearly 30% of Democrats had decided before October, but only 19% of Republicans had settled on a choice that early.

A group of particular importance to GOP candidates is Evangelical Christians. But guess what? Once again we see no real differences in decision time. Evangelicals were just as likely to make a late decision as any other Iowa Caucus goer in 2008.

What’s the takeaway from this deeper dive into Iowa Caucus goers’ decision timing? Simply, there is a lot of room for candidates to play in the final weeks before Iowans cast their First in the Nation votes. The lesson for 2016 may be exactly what Silver and others are saying. There remains fluidity, and the inevitability of Donald Trump in Iowa (or anywhere else) is not at all certain. If the past holds, any indicators for today, at two months ahead of the caucuses, a very large share of the vote is still in play, no matter what people tell pollsters today. And in particular, given that GOP voters appear to make later decisions even in 2008 when both parties had large fields, there is every reason to think we have a lot more ups and downs to watch before this whole thing is over. Some people are committed, but most are still shopping, even as their choice set gets smaller and the shopping days fly by.

Final note: The data we collected with our 2008 in-caucus, along with the survey instruments, are available for anyone to examine at