Very few people watch political polls more closely than I do. (Whether that's a good thing or suggests that I'm slightly neurotic is up for debate.) When you monitor surveys incessantly, you occasionally see results that you're unsure how to interpret. You don't know whether they signal a key turning point in public opinion or whether they're just a hiccup, a passing blip. Or perhaps the odd results are from an outlier poll, a statistical anomaly that is the political equivalent of a false positive medical test.
We're currently experiencing one of those periods of uncertainty. One interpretation of recent results is that the momentum in this critical midterm election has shifted and the Republican wave has subsided. Another interpretation is that it's too soon to tell whether much has changed at all.
For the weeks of July 12-18 and July 19-25, the Gallup Organization's weekly aggregation of daily tracking polls showed Democrats ahead among all registered voters on the generic congressional ballot test question by 6 points (49 percent to 43 percent) and 4 points (48 percent to 44 percent), respectively. Each poll canvassed more than 1,500 registered voters nationwide. For the uninitiated, the generic ballot test question tries to approximate what the popular two-party vote will be nationwide and, over time, it has closely corresponded to the outcome on Election Day.
If Democrats are running 4 or 5 points ahead among registered voters, it would mean a very, very close contest for control of the House.
Gallup noted that this was the first time that either party has held an advantage of this size for two consecutive weeks. In the 21 weeks that Gallup has asked the generic ballot test question this year, the two parties have averaged a tie. It should be noted, however, that polls of registered voters inherently tilt Democratic by 4 or 5 points compared with polls of likely midterm election voters. Voter turnout for midterm elections is about a third less than it is in presidential years, and midterm voters tend to be whiter and older, two problem population groups for Democrats this year.
For the four previous weeks, the two parties were tied at 46 percent on the generic ballot question. For the four weeks before that, Republicans averaged a 3-point lead, 48 percent to 45 percent. So, if Democrats really have turned up the heat and are running 4 or 5 points ahead among registered voters, the practical result would be about an even proposition among likely midterm voters and the national popular vote. If that were true, it would mean a very, very close contest for control of the House.
The November 2 election is obviously the ultimate arbiter. But before we get to the real thing, Gallup and other respected organizations will take polls that will either corroborate these findings or show that nothing has changed. Gallup is the most trusted name in national polling and its results should never be dismissed lightly. But no matter how terrific a pollster is, some statistical noise periodically pops up.
Since the generic ballot tracking started for the 2010 election, each party has seen its number drop as low as 43 percent and rise as high as 49 percent. Basically, both Democrats and the GOP have been at 46 percent and have traded within a 3-point range. What we are looking for is whether either party breaks out of that range and begins to consistently record an advantage, with "consistently" certainly defined as more than two weeks in a row.
What make me skeptical is that no defining event has taken place that would have triggered a significant shift in this year's race. I don't think that passage of the financial services regulatory reform bill would have caused a significant shift, nor would the capping of the Deepwater Horizon oil well.
I canvassed several pollsters who see large quantities of data from around the country; none seems to have detected any shifts in the past two weeks, and all seem to be at a loss in figuring out what would have triggered a change. With the president's approval ratings stable -- 46 percent one week in the Gallup Poll, 45 percent the next -- whatever might be happening does not seem to be helping him.
For now, people will have to just sit tight, wait for next week's Gallup release, and watch how other reputable pollsters weigh in.
This article appears in the July 31, 2010, edition of National Journal.