Early Iowa Presidential Polls a Better Predictor Than National Ones

by Stuart Rothenberg October 8, 2015 · 11:18 AM EDT

Last month, national polls by CNN/ORC, Fox News and NBC News/Wall Street Journal got plenty of attention, and they certainly helped readers and viewers understand what is going on in the Republican and Democratic presidential contests.

But if history is any guide, early national polls are far less valuable in understanding what is happening in the presidential contest than are reliable surveys of Iowa voters, such as the NBC News/Wall Street Journal/Marist polls.

I looked back to surveys in the summer and fall of 2007 to see how the national surveys and Iowa caucuses polls conducted by reputable firms fared in documenting public opinion trends and in predicting the direction of the Democratic and Republican races — and Iowa polls were the clear winners.

On the Democratic side, polls of Iowa caucuses attendees were far ahead of national polls in documenting and predicting the rise of Barack Obama. National polling lagged far, far behind.

A July 29-Aug. 5, 2007, University of Iowa survey of those planning to attend the caucuses found Hillary Rodham Clinton at 27 percent and holding a narrow 5-point lead in the caucuses over John Edwards and Barack Obama, who were tied at 22 percent.

A July 26-31 ABC News/Washington Post poll of likely Iowa voters now looks even smarter, since it found Obama at 27 percent, 1 point ahead of both Clinton and Edwards, who were deadlocked at 26 percent.

At almost the same time, a July 27-30 NBC News/Wall Street Journal national survey found Clinton far ahead of second-place Obama by 21 points, 43 percent to 22 percent, and a July 25-29 national Pew Research survey showed Clinton over Obama by a very comfortable 19 points, 40 percent to 21 percent.

Iowa and national polls conducted two months later showed the same trend.

An Oct. 1-3 Des Moines Register survey of likely caucus attendees found Clinton at 29 percent and with single-digit leads over Edwards (23 percent) and Obama (22 percent). Yet at roughly the same time, an ABC News/Washington Post national survey conducted Sept. 27-30 found Clinton (53 percent) leading second-place Obama (20 percent) by 33 points.

An Oct. 12-14, 2007, Gallup survey showed Clinton leading Obama by 29 points. “Leads greater than 20 percentage points have been rare in past Democratic campaigns, but historically those who have enjoyed such a large lead in Gallup Polls this late in the year won the nomination the following year,” said the polling firm in its analysis of the results.

I checked to see whether this dynamic might be limited to Democrats, and of course it wasn’t.

A July 18-21 ABC News/Washington Post national survey found Rudy Giuliani with a commanding 18-point lead (34 percent to 16 percent) over John McCain, with Fred Thompson third at 14 percent.

But a July 26-31 ABC News/Washington Post survey of those planning to attend the Republican Iowa caucuses found Mitt Romney leading with 26 percent, followed by Giuliani at 14 percent and Thompson at 13 percent.

Two months later, an Oct. 1-3 Des Moines Register survey of the caucuses found Romney first with 29 percent, with 18 percent for Thompson and 12 percent for the fast-rising Huckabee.

On the other hand, both the Sept. 28-30 national NBC News/Wall Street Journal poll and the Sept. 27-30 ABC News/Washington Post national survey found Giuliani maintaining large leads, and neither showed Huckabee to be in the top four hopefuls.

A little more than a month before the Jan. 3, 2008, Iowa caucuses, the Des Moines Register poll of Nov.25-28 found Huckabee leading Romney 29 percent to 24 percent in Iowa, with Giuliani down to 13 percent and Thompson sinking to 9 percent.

In contrast, a national FOX News survey of Nov. 13-14 found Giuliani with a clear lead (37 percent to 17 percent over John McCain), as did two different Gallup polls for USA Today late in the year, one conducted Nov. 30-Dec. 2 and the other Dec. 14-16.

None of this should surprise anybody.

Iowans get the first glimpses of the candidates up close, and they take their role as the first step in the party nominating process remarkably seriously. Given that, their opinions are more informed than are those of most Americans, who pay far less attention to politics and who have not seen the candidates close-up.

Of course, Iowa doesn’t select party nominees. It is merely the first step in the process, and national survey results reflect factors — including name identification, money and perceived front-runner status (thanks to the media’s coverage) — that can aid a candidate later in the nominating process.

But the national polls exaggerate the importance of the large, most populous states and ignore the impact of the early contests on the larger electoral dynamic. Obama’s victory in Iowa, for example, changed the Democratic race dramatically, producing a shift in the national numbers.

Candidates who fare poorly in the early states can be all but eliminated from the race, as their money dries up and their scenarios for victory look unrealistic.

National polls certainly are not without value. But they tend to be over-valued by those in the national media and behind the curve compared to public opinion trends in Iowa. So if you want to know not only where the race is now but also where it is going, pay particular attention to Iowa surveys from proven pollsters over the next four months.