I didn’t notice this quote: “Experts estimate that more than 20 percent of voters nationwide will cast their ballots before Election Day by mail or at early-voting locations, a proportion of the electorate that is rising with each election. Some states and counties open the ballots before Election Day and keep the results secret; others count them with regular ballots.”

I’m one of the experts quoted in the article (and I supplied information for the graphic). My 2006 estimate is higher, though. A conservative guess would be 25%, but I wouldn’t be surprised to see 30% of ballot cast early this election.

Early voting rates are proportionately higher in lower profile contests. There are two reasons why this is the case. First, in lower profile contests, a larger proportion of turnout consists of habitual voters, and habitual voters, who are also more likely to vote early. So if we take the trendline (14% in 2000 and 20% in 2004), and further adjust for the midterm, that’s how I get to 30%.

But there is a second reason that there are a higher number of voters during low profile contests–though due to the lack of good survey evidence, here we are more in the realm of informed speculation. [Note: some of these ideas rely on Mike Alvarez’s excellent “Information and Elections”, U of Michigan press, 2000].

In a lower profile contest, there is less political information available to the electorate. The result is that many citizens opt not to vote at all. But among those highly informed citizens, the low profile contest makes less of a difference. The highly informed seek out political information on their own and are less influenced by media coverage. And these highly informed voters are more likely to vote.

Are the highly informed voters the same as the habitual voters? The groups overlap, but not as much as you might expect. Some vote habitually even though they are not well informed, relying on such information shortcuts as partisanship, incumbency, or evaluations of the “state of the times.” And some vote habitually because they care a lot about, and are informed a lot about, politics. But the two groups are not identical.

This is what makes 2006 difficult to predict. I have noticed some tendency among voters to hold their ballots during especially hard fought contests, presumably because they are contending with two-sided information flows from competing candidates, and are waiting until the last moment to make their decision. I have no idea what proportion of “late” voters are “informed but conflicted” late voters, and what proportion are simply disinterested or not paying attention. I hope to look at this question in the future.

Regardless, 2006 is, compared to the last few midterm contests, a hard fought, nationalized election. So it may look a lot more like a presidential contest than a typical midterm. While the trendline on early voting is rising, the impact of a hard fought campaign in 2006 may reduce early voting slightly.

We’ll find out a lot more in the next two weeks. My call?

30% early voting in 2006. If I overestimate, I’ll buy Hall and Alvarez a Goose Island in Chicago. They’ll hold me to that for sure!