December 13, 2007 -- It really is the early bird that gets the worm, at least in politics. New research out of Brown University shows that the American political process through which we choose the President is "front loaded," with voters in the early primaries having much more clout than voters in later primaries.
The researchers used a technique called statistical modeling to determine the effect of early primaries during the 2004 national election, and their results suggest that any candidate who ignores the relatively small state of Iowa does so at considerable political peril. The Iowa results, followed quickly by the New Hampshire primary, have an enormous impact on what the political pundits, and the researchers, call "momentum."
"Momentum effects will be generated when candidate performance in the actual primaries exceeds voter expectation," Brian Knight, associate professor of economics and public policy at Brown University, said in an interview.
That won't come as much of a surprise to pundits who have seen the fate of many a political candidate set in stone in the early primaries. But the research, carried out by Knight and a graduate student, Nathan Schiff, is a rare effort to study scientifically how momentum shifts as the sequence of primaries moves along.
Knight and Schiff came up with some surprising numbers. Voters in Iowa and New Hampshire have six to 20 times the influence of voters in later states in the selection of candidates.
Is it a bad thing? Not necessarily.
It may fly in the face of the democratic ideal of "one person, one vote," Knight said, but voters in Iowa and New Hampshire have a better chance of getting up close and personal with the candidates than voters in mega-states like California and New York. So the early voting record contributes to "social learning," an academic term that means we lean heavily on the judgment of those who came before us.
National primaries are usually spread out over several months, giving momentum more time to grow, or diminish. Thus it is quite likely that the winning candidates for both parties will emerge long before the last primaries are held.
At least that's the way it has usually been in the past. But maybe not this year.
The Iowa caucuses will be held on Jan. 3, and New Hampshire voters will go to the polls on Jan. 8. But "super Tuesday" will come less than a month later, on Feb. 5, when nearly half the states will hold primaries, including California, New York and New Jersey. That compression of the primary season will leave less time for momentum to build, so it's quite possible that no candidates will have either party's nomination locked up by the national conventions.
"We're putting ourselves on the line a little bit," Knight said, with that projection.
The researchers based their conclusions on data collected during the 2004 national elections. Only the Democratic primaries were studied because George W. Bush, running for re-election, had the Republican nomination locked up. The data consisted chiefly of polls conducted by the National Annenberg Election Survey 2004, located on the campus of the University of Pennsylvania. That survey was intended for academic research and the polls were conducted every day from Oct. 7, 2003 through the general election in November of 2004.
The survey gave the Brown University researchers the data they needed to measure momentum shifts throughout the primaries on a day by day basis. Knight and Schiff hope soon to publish their study in a peer-reviewed journal.
Prior to the balloting in Iowa and New Hampshire in 2004, the Democratic nomination seemed like it was already pretty well in the bag for Howard Dean, who had a huge lead in national polls. But Iowa turned that around. Sen. John Kerry was the big winner, at 38 percent, followed by Sen. John Edwards at 32 percent. Dean placed a distant third at 18 percent.
Dean was forced to drop out, and Kerry won the nomination. But it might not have been that way if the first primary had been in another state, say South Carolina, which Edwards carried. In that case, Edwards might have picked up the early momentum instead of Kerry.
So that suggests that Iowa voters, who will send only 56 delegates to the Democratic convention out of a total of 4,366 delegates, will once again have a "disproportionate" amount of clout, Knight said. And this time the Republicans will feel that clout as well. Both the parties are, at the moment, too close to call for Iowa.
The race in Iowa is close enough for someone in both parties to do a lot better, or a lot worse, than had been expected, thus gaining or losing momentum. And it will be a rare opportunity for the rest of us to watch and see if the researchers got it right.
But of course, politics can bring surprises. Who among us in 2004 expected the "Dean scream," that bizarre victory call on the heels of his loss in Iowa? It may have contributed as much to his defeat as the voters in both Iowa and New Hampshire.