Rothenberg: Iowa Caucuses Have Their Own Arithmetic

Posted January 8, 2004 at 9:41am

As the Iowa caucuses approach, it’s wise to remember that when it comes to candidate strength and caucus results, nothing is quite as simple (or straightforward) as it first seems. And I’m not referring to the so-called expectations game. [IMGCAP(1)]

For example, while Iowa’s Jan. 19 caucuses select delegates to county conventions based on the number of Democratic votes during the most recent presidential and gubernatorial general elections, caucus attendees in the state’s larger, often urban — and Democratic — areas don’t have the disproportionate influence that you might assume.

Individual caucus attendees who live in sparsely populated rural areas can have more of an influence on the selection of delegates to the Democratic National Convention than those who live in Des Moines, Dubuque, Davenport or other cities.

The relative importance of rural attendees comes from the fact that a caucus in Des Moines might attract a couple hundred attendees, while one in a rural area might draw only a dozen people. Both caucuses select delegates, but caucus attendance is often lower per delegate in rural areas, giving rural attendees more clout.

And geography is a factor in other ways.

Democrats familiar with the state note that each of the top Democratic presidential candidates is particularly strong in certain parts of the state, and that this can lead them to “waste” caucus attendees in some areas.

For example, if former Vermont Gov. Howard Dean’s support turns out to be disproportionately concentrated in a handful of areas, including Des Moines and college communities, such as Iowa City (University of Iowa), Ames (Iowa State University) and even Grinnell (Grinnell College), he will have more caucus attendees than he needs to elect delegates to the next round in those areas.

In effect, he could end up wasting some of those supporters, who would be more valuable to his efforts if some of them attended caucuses in other parts of the state.

Sen. John Edwards (N.C.) figures it will be more difficult to compete with his Democratic rivals in Iowa’s cities, so he is concentrating his efforts in rural areas. But while in some caucuses he may exceed the 15 percent threshold he needs to receive delegates, he may fall short in others, essentially “wasting” support in those areas.

Finally, the caucuses’ results are susceptible to being misreported, since two separate things are really going on at once.

Unlike Iowa Republicans, the Iowa Democratic Party does not report a “straw vote” or “presidential preference” of caucus attendees statewide. Edison Media Research/Mitofsky International, however, is doing an “entrance poll” this year for a consortium of media companies.

That data will allow the media to dissect Democratic caucus-goers’ preferences by age, income, gender, union membership and other variables. To avoid affecting turnout, the poll data won’t be released until after all of the caucuses begin. (But given cell phones and BlackBerrys, caucus attendees may not completely be in the dark about the entrance poll’s numbers.)

While entrance poll data will tell us a good deal about caucus attendees, including their initial candidate preferences, they won’t exactly reflect delegates won (what analysts refer to as “state delegate equivalents”). That’s because delegates will be decided only after some candidates fail to meet the 15 percent threshold at individual caucuses and attendees engage in a round of horse-trading. Only then will attendees re-caucus at each site, producing the final results.

The state party will release delegate figures after those final numbers are available, and in 1984, 1988 and 2000 the entrance poll figures and delegate figures differed by a few percentage points. In 1984 and 1988, the top three finishers in the “initial candidate preference” entrance poll increased their showing when final delegate equivalents results were calculated.

In 1988, for example, Rep. Richard Gephardt (Mo.) “beat” then-Sen. Paul Simon (Ill.) 27.4 percent to 24.3 percent in the entrance poll’s results, but had a 31.3 percent to 26.7 percent advantage in delegate equivalence reported by the state party. In contrast, Jesse Jackson drew 11.1 percent in the entrance poll, which measured initial presidential preference of caucus attendees, but drew only 8.8 percent in delegate equivalence.

Any discrepancy between the entrance poll and delegate equivalence won’t matter, of course, if the Iowa results produce a clear winner. But in a squeaker, the two different sets of “results” could affect interpretations of the results. In any case, observers should recognize that both numbers could be “true” and “accurate,” since they count different things.

So beware about any early horse race numbers from the entrance poll. They may not tell the whole story.

Stuart Rothenberg is editor of the Rothenberg Political Report.