Trouble, by the numbers

Sometimes the most important clues are hiding in plain view. That was the case in late June, when the Gallup Organization reported that the share of voters who describe themselves as conservative had increased from 37 percent to 42 percent in the past two years.

That does not sound like a big change. But given the long-term stability of these basic philosophical alignments, the reaction it measured to the economic troubles and the performance of the new Democratic administration is very significant.

The most recent number, a cumulative figure based on surveys during the first half of 2010, drew some attention because it was the highest percentage for conservatives in any such poll since Gallup started asking this question in 1992. The five-point gain came equally from the ranks of moderates and liberals, who fell to 35 percent and 20 percent, respectively.

What was less noticed at the time were the state-by-state Gallup figures, but thanks to the busy calculators at Third Way, the moderate Democratic advocacy and political action group, the implications of those numbers for the midterm election have become clear in a memo now circulating around Washington.

They explain why so many Democratic candidates are struggling in states such as Wisconsin and Washington, which have been kind to their party in the recent past. And they argue that President Obama may have been focused on the wrong target when he kicked off his fall campaigning at the University of Wisconsin in the liberal stronghold of Madison.

The message emerges from some pretty basic math calculations work done by Lydia Saad of Gallup and then overlaid by Anne Kim and Jon Cowan of Third Way.

Saad ran the Gallup numbers for individual states, with few surprises. Wyoming, Mississippi, Utah and South Dakota checked in with 50 percent or more conservatives. At the other end, Rhode Island, Connecticut, Vermont, Massachusetts and Colorado were the most liberal states but only in Rhode Island did the percentage top 30.

Then Kim and Cowan added their own assumptions: Suppose Democratic candidates run as well as Obama did nationally in 2008, taking 20 percent of the conservatives, 60 percent of the moderates and 89 percent of the liberals. And suppose, too, that turnout rates are the same for all three groups.

With the updated Gallup figures, a 2010 Democratic candidate who matched Obama's national percentages would still win Colorado, Connecticut, Delaware, Oregon and Washington. But, with more conservatives and fewer liberals in the mix, the Democrat would come up short in 13 other competitive states and barely break even in California, Illinois and New Hampshire. Among the big states where the numbers now break against the Democrats are Florida, Ohio and Pennsylvania.

As anyone who is following the election campaign knows, this kind of analysis makes no allowance for the possible impact of Lisa Murkowski's write-in effort in Alaska or the crash-and-burn Republican gubernatorial campaign in Colorado.

But the basic math shows why Democrats such as Sen. Russ Feingold in Wisconsin are struggling this year and why Obama may witness the defeat of his fellow Democrats running for governor and senator in his home state of Illinois.

And, in the view of the Third Way analysts, the math also suggests the limitations on the apparent White House strategy of concentrating the president's campaign efforts on young people and single women. To the extent that those groups delivered liberal votes to Obama in 2008, it makes sense to mine them again.

But if Gallup is right, and I believe its methodology is solid, there simply are fewer liberal votes to be won this time. And, as the Third Way memo says, "While the middle has always played a pivotal role in American electoral politics, where they swing this fall will certainly decide the fate of the Democratic majority."

• David Broder's e-mail address is

© 2010, Washington Post Writers Group