Pollster Stanley Greenberg says his latest Democracy Corps poll shows Republican support falling sharply in the Deep South, in rural areas and among downscale men -- groups among which Republicans have had big leads. Such a change wouldn't affect many House races, because these groups are concentrated in districts that are very heavily Republican. But it could put another dozen or so Republican seats in play, over and above the dozen or so where Democrats are making strong challenges (Republicans are making strong challenges in about half a dozen Democratic seats).
In years when voters have shifted sharply to one party -- Democrats in 1974, Republicans in 1994 -- the winning parties captured only about half the seats they targeted.
So even if the field of contested seats expands as Greenberg suggests, Democrats could take the House only if they picked off half their targets, while defending every one of their own contested seats. But few seats are captured without strong challenger candidates, and while Democratic recruiting has had some successes, it hasn't produced serious challengers in all these seats. Democrats have a chance to win the House, but it's far from a sure thing.
Of course, not all the factors have played out. At this point in the 1994 cycle, the Clinton healthcare plan had not yet collapsed, the Democrats had not yet embraced gun control and the Republicans had not yet rolled out their Contract With America. Will rural voters, however, cross with Bush and vote to install as speaker of the House a San Francisco Democrat?
Finally, the polls, whatever their bad news for Republicans, offer few clues about who's actually coming out to vote. In 2004, Republicans won because they did a better job turning out their party base than the Democrats did. They expanded the electorate and have a bigger reservoir of voters to draw on. My guess is that turnout, more than voter shifts, will determine who wins in 2006.