Why Iowa caucuses don't matter: some data

By Sharon Machlis

Intense media coverage of "first-in-the-nation" Iowa caucuses gives the impression that the results of those caucuses matter.

They don't.

Results won't indicate anything beyond who's popular among Iowa Republican activists ... although not for the reason many skeptics cite.

Only 100,000 or so caucus-goers will be weighing in tonight, compared with more than 130 million Americans who will likely vote on Election Day. And more than one journalist has cited that small number as a reason why the caucuses are, as New York Times columnist Gail Collins wrote, "really ridiculous. ... Maybe the Republicans will hit 150,000! That is about the same number of people in Pomona, Calif. Imagine your reaction to seeing a story saying that a plurality of people in Pomona, Calif., thought Newt Gingrich would be the best G.O.P. presidential candidate. Would you say, 'Wow! I guess Newt is now the de facto front-runner?' Possibly not."

But it's not the number of people participating that makes Iowa meaningless as a predictor, but its skewed sample. In other words, Iowa voters are not a random sample of American voters and thus do not reflect all U.S. voters in ways that matter to the ultimate outcome.

If a caucus (or primary) were a true reflection of all American voters, you'd need a sample of less than 10,000 to accurately reflect a population of 130 million voters (that's with a margin of error of 1% using the standard 95% confidence level and Macorr Research Solutions' Sample Size Calculator). A vote of 100,000 would be considerably more than needed.

But whatever you believe about the heartland, Iowans reflect only a portion of our nation. While 40% of Americans live in a top-25 metropolitan area, zero Iowans do -- and it's certainly possible that city-dwellers are interested in some different issues than the average Iowan. More than one-third of Americans are Hispanic, African-American, Asian or other minority, and four states including California and Texas have "majority minority" populations; yet only 8.7% of Iowans are minorities. Certainly minorities groups may have different voting patterns.

Collins does acknowledge this issue as well, noting: "On Tuesday, there will be a contest to select the preferred candidate of a small group of people who are older, wealthier and whiter than American voters in general, and more politically extreme than the average Iowa Republican."

This skewed sample problem shows up in its failure to reflect the eventual winner. Of the five "first-in-the-nation" Iowa Republican caucuses where there was no clear front-runner -- i.e. no sitting president or vice president seeking the GOP nomination -- the Iowa victor went on to win the Republican nomination only twice: Robert Dole in 1996 and George W. Bush in 2000. Also in Iowa, George H.W. Bush beat Ronald Reagan in 1980, Robert Dole beat George H.W. Bush in 1988 (Bush finished a distant third) and Mike Huckabee won in 2008. Based on the data, it's a better argument that winning Iowa predicts an overall loss -- perhaps because tailoring a campaign message to Iowa's electorate isn't necessarily appealing to the national Republican party as a whole.

The only reason Iowa has become important is its effect on subsequent candidate fundraising success and front-runner claims -- and that's because a media echo-chamber has made it so. But think about it: Even with that edge, Iowa caucuses have less than 50% accuracy in predicting the ultimate Republican nominee. That's worth keeping in mind when reading or watching various breathless news reports about tonight's caucuses.

The Iowa caucuses are certainly interesting to political junkies, much like spring training and Opening Day are enjoyed by baseball fans. But few knowledgeable sports enthusiasts would argue that success in the first game of the season is an accurate predictor of how a team will do the rest of the year. Knowledgeable voters -- and political reporters -- should view Iowa caucuses the same way.

See more from the Data Avenger series.

Sharon Machlis is online managing editor at Computerworld. Her e-mail address is smachlis@computerworld.com. You can follow her on Twitter @sharon000, on Facebook, on Google+ or by subscribing to her RSS feeds:
articles | blogs .

Related:

Copyright © 2012 IDG Communications, Inc.

  
Shop Tech Products at Amazon