We haven’t written as much about Illinois as some other Senate contests — but it may be taking on an increasingly important role as some other races that had once looked to be competitive, like Connecticut and Missouri, now show a clearer advantage for one or another party.
The race in Illinois has long been within the margin of error in most polls; our forecasting model last week characterized it as a near-tie, with the Republican Mark Kirk projected to win by 0.4 points over Democrat Alexi Giannoulias.
What’s unusual about Illinois is the number of voters committed to minor-party candidates — like the Green Party’s nominee, LeAlan Jones, and the Libertarian, Mike Labno — or who haven’t committed to a candidate at all yet. Collectively, the third-party candidates have held between 5 and 11 percent of the vote in recent polls, while about 10 to 15 percent of voters remain undecided.
Such a circumstance is quite unusual. I searched our database of Senate race polls and came up with only four examples since 1998 in which neither major-party candidate had more than 42 percent of the vote in the polling average with a month to go in the campaign:
- In Michigan in 2000, Democrat Debbie Stabenow led Republican incumbent Spencer Abraham 41-37 in the polls with a month to go; Ms. Stabenow wound up winning the race by 2 points.
- In Colorado in 2002, Democratic challenger Tom Strickland led Republican incumbent Wayne Allard 40-37 in the polls. But Mr. Allard came back to win by 5 points.
- In an open-seat race in Oklahoma in 2004, Republican Tom Coburn led Democrat Brad Carson 42-41 with a month left. Mr. Coburn eventually won by a considerably larger 13-point margin.
- In Minnesota in 2008, Democrat Al Franken and incumbent Republican Norm Coleman were in a 40-40 tie in the polls with 30 days to go, with the Independence Party candidate, Dean Barkley, also drawing a significant percentage of the vote. That race, of course, remained just as close through election day, with Mr. Franken eventually declared the winner by 312 votes after a recount that lasted for months.
I can’t say I see much in the way of a pattern there — other than that the error in both the Colorado and Oklahoma races was fairly high, suggesting that such races may behave unpredictably. Indeed, our model accounts for this property, assuming there is a larger margin of error in the forecasts when the number of undecided voters (and voters claiming to support third-party candidates) is higher.
Illinois residents have some recent history of casting protest votes for third-party candidates when they are dissatisfied with the major-party nominees, as they appear to be this year. In the gubernatorial race there in 2006, for instance, the Green Party’s nominee Rich Whitney won around 10 percent of the vote, although the incumbent Rod Blagojevich was nevertheless re-elected with 49.8 percent of the vote to Republican Judy Barr Topinka’s 39.3 percent.
The race between Mr. Blagojevich and Ms. Topinka, however, was only marginally competitive — and so voters might have been more comfortable casting a “protest vote” there. This year, by contrast, every vote figures to count in both the Senate race and the race for governor, which had appeared to favor Republican Bill Brady but now has tightened in some polls. Third-party candidates sometimes under perform their margins in the polls, especially in races where the polling between the major party nominees is tight.
It might be reasonable to surmise that Mr. Giannoulias, the Democrat, is the favorite to pick up some votes from the Green Party’s Mr. Jones. Still, the same logic could hold for Mr. Kirk and the Libertarian candidate’s votes — and in general, the polling has shown no clear relationship between the number of third-party votes and the margin between between Mr. Giannoulias and Mr. Kirk.
Another question is whether, in a state where voters seem to have little inspiration to cast their ballots, campaigning by President Obama might make some difference. But some polls show Mr. Obama, who won his home state with 62 percent of the vote in 2008, as now only having an approval rating of about 50 percent there among likely voters.
In the absence of a clear trend in the race, both Mr. Kirk’s and Mr. Giannoulias’s campaigns, perhaps growing impatient, have released their own polling in the state.
Mr. Kirk’s poll, from the firm Fulcrum Strategies, gave him a 42-33 advantage. But a poll by the Democratic Governors Association, obtained by FiveThirtyEight, gives Mr. Giannoulias a 40-37 edge. The Democratic poll — conducted Sept. 23rd through 26th among 604 likely voters by the firm Global Strategy Group, also showed a tight governor’s race, with Mr. Brady leading the Democratic incumbent, Pat Quinn, with 36 percent of the vote to 35.