What We Can Learn from the 2004 Presidential Race

by Stuart Rothenberg July 12, 2019 · 1:59 PM EDT

Beware of reading too much into presidential polls. Take, for example, the 2004 race.

An August 2003 CNN/USA Today/Gallup national survey found Connecticut Sen. Joe Lieberman, the 2000 Democratic nominee for vice president, leading the party’s presidential field with 23 percent. He was trailed by former House Majority (and Minority) Leader Richard A. Gephardt (13 percent), former Vermont Gov. Howard Dean (12 percent) and Massachusetts Sen. John Kerry (10 percent).

A month later, another CNN/USA Today/Gallup poll found the newest entry into the race, retired Gen. Wesley Clark, topping the pack with 22 percent, while Dean (13 percent), Kerry (11 percent), Gephardt (11 percent) and Lieberman (10 percent) were bunched together.

There were others in the race, of course — North Carolina Sen. John Edwards, former Illinois Sen. Carol Moseley-Braun, Florida Sen. Bob Graham, Ohio Rep. Dennis J. Kucinich and activist Al Sharpton — but they barely registered in the polls.

By December, Dean had pulled ahead in the CNN/USA Today/Gallup poll, opening up a considerable lead over Lieberman, 31 percent to 13 percent. Clark and Kerry drew 10 percent, and Gephardt received 8 percent.

(Multiple other polls, including those for CBS News, Newsweek, ABC News/Washington Post and NBC News/Wall Street Journal showed the same movement to Dean.)

A little more than a month later, on Jan. 19, Kerry won the Iowa caucuses with 38 percent, while Edwards surged into second place with 32 percent. Dean and Gephardt finished much further behind.

Clark and Lieberman did not really compete in the caucuses, preferring to devote their time to New Hampshire.

The first CNN/USA Today/Gallup national survey after the caucuses, conducted Jan. 29 to Feb. 1, 2004, showed Kerry leading nationally with 49 percent to Dean’s 14 percent, Edwards’ 13 percent and Clark’s 9 percent. The race for the Democratic nomination was effectively over.

In a matter of weeks, Kerry went from a political basket case to the de facto Democratic nominee.

Edwards become his major opponent and eventually his running mate.

The early polls weren’t “wrong.” They just weren’t predictive of how things would develop.

16 Years Later
But 2020 is very different from 2004, isn’t it? In 2004, neither party had ever nominated a person of color or a woman to be president. The  Democratic field then was large but nothing like the almost two dozen hopefuls this year.

The Democratic Party of 2004 was very different from the one we see today. And the political environment was very different then, given its proximity to 9/11.

One thing that is similar is the timing of the first debate. In 2003, Democrats held a debate on May 3, about six weeks earlier than this year’s June debates.

Nine Democrats — including one woman, Moseley-Braun, and two candidates of color, Moseley-Braun and Sharpton — participated in the debate, which was held at the University of South Carolina and broadcast on C-SPAN and 59 ABC affiliates. George Stephanopoulos of ABC News was the moderator.

The More Things Change …
So, what might 2004 teach us about 2020?

First, as Gallup’s Frank Newport noted shortly before the 2004 Iowa caucuses, an early lead in national polling doesn’t guarantee anything. 

Sure, some hopefuls (including George W. Bush and Al Gore in 1999-2000, and Donald Trump in 2015-16) have gone wire-to-wire, but most contests show surges and lead changes.

Throughout 2003, voters were doing just what they are doing this year: learning about the candidates, sampling them and sorting them into those they might support and those they won’t.

Polls from before the 2004 Iowa caucuses showed Kerry virtually eliminated from the race — until he wasn’t.

As Democratic pollster Diane Feldman told me more than a decade ago, it’s only weeks before the Iowa caucuses that caucus attendees start asking themselves which of the hopefuls could handle the office of president. That’s a different test from “Who do I like?’ or “Who do I agree with the most?”

Second, every day is “crucial” for cable TV hosts and political spinners, but not for campaigns and candidates.

Fundraising numbers are worth noting, as are news reports about controversies and polling. But all the hoopla — countdown clocks, Super Bowl-like pre-debate introductions, massive panels of experts dissecting every new public opinion survey — is mostly about drawing eyeballs, not educating viewers.

Third, Iowa can change everything.

The winner of the Iowa caucuses doesn’t automatically win the nomination, but the results invariably change perceptions about the front-runner and the entire field, creating new expectations and weeding out the also-rans.

“The [nomination] process is not a simultaneous process but a sequential one,” noted Feldman, whose clients have included the late Minnesota Sen. Paul Wellstone, Ohio Sen. Sherrod Brown, Washington Gov. Jay Inslee and South Bend, Indiana, Mayor Pete Buttigieg.

Fourth, the Hawkeye State will be more uncertain than ever with the addition of a “virtual caucus” that allows Iowa Democrats to caucus up to six days before the Feb. 3 traditional in-person caucuses. 

People undoubtedly are paying more attention to the current race than they did at this point in 2003-04, in part because Democrats are so focused on defeating Trump next year.

And Trump’s nomination and election may well have altered the way people look at the political process and campaigns.

But the coverage of the Democratic contest — including contradictory polling results and multiple media narratives — tells us less than you may think about the party’s eventual nominee.

Voters have plenty of time to evaluate the crowded field before Iowa and Nevada caucus-goers and New Hampshire and South Carolina primary voters make their choices and remake the field.