You Can’t Ignore The Polls, But It Helps If You Know How To Read Them
Democratic National Committee Chairman Tim Kaine, Obama administration strategist David Axelrod, and Pennsylvania Governor Ed Rendell, have all recently advised us to not pay much attention to the pre-election polls-- at least at this early date in the fall campaign. Washington Post columnist Dana Milbank correctly dismisses this dismissal of the polls as just so much political rhetoric: an "underdog's time honored strategy." It's impossible to believe that political pros like Kaine, Axelrod, and Rendell are not paying attention-- a lot of attention-- to the polls. But, more important than simply paying or not paying attention to the polls is getting a handle on what the polls mean and what they measure.
That's easier said than done. Over the years, the Gallup Poll's weekly generic House ballot has become almost the gold standard used inside the Beltway to forecast the outcome of the race to control Congress. The problem is that in 2010 Gallup's generic ballot numbers have fluctuated wildly, for no clearly discernable reason, almost from week-to-week. In mid-May the Republicans and Democrats were tied at 46% each. A couple of weeks later, the GOP held a 49% to 43% advantage. But, at the beginning of July, the two parties were effectively tied again (Democrats 46% and Republicans 45%) and just two weeks later, much to the joy of DC Democrats, their party held a 49% vs. 43% lead. By mid-August, however, it was the Republican's turn to cheer as the GOP led 47% to 44%, a lead that ballooned to ten percentage points (51% to 41%) as August turned to September. But then, just one week later, the two parties were once again tied (46% each) in Gallup's first post-Labor Day tracking poll, essentially the same numbers obtained when the Gallup results for the entire year are averaged.
The Post's Chris Cillizza posited several explanations for the most recent change in Gallup's generic ballot results. Possibly, Cillizza says, the poll giving the Republicans a ten-point lead was an "outlier," or one of those anomalous samples that occasionally fall beyond the normal range of sampling error. Perhaps it was. It is certainly true that many of Gallup's weekly numbers during 2010 were much closer to the generic ballot results of other pollsters than that one was. The problem with this reasoning, however, is that Gallup's generic ballot numbers have varied so dramatically from one weekly survey to another that it's difficult to know exactly which one might truly be an outlier.
A more plausible explanation is that, unlike many pollsters, Gallup does not "weight" its samples by the party identification of respondents. Gallup might argue that this methodology leads to purer, more accurate, results than predetermining the number of Democrats, Republicans, and Independents within its samples. However, at least in the current political era, this approach is likely to give the GOP an advantage on the generic ballot.
That is because the Democratic Party now has a clear and decisive lead party ID lead over the Republican Party. In a June survey conducted for NDN by communications research and consultation firm, Frank N. Magid Associates, the Democrats held a 47% vs. 33% edge over the Republicans with 19% of the electorate claiming to be totally unaffiliated independents. This distribution is matched almost identically by figures from the Pew Research Center's three most recent surveys, conducted from late July to late August. Across those three polls, a little under half of Americans (46%) identified or leaned toward the Democrats while about a third (36%) said they were Republicans and 14% were unattached independents.
None of this is to say that things will be easy for the Democrats this fall. Things are almost never easy for a newly elected president's party in the first midterm election of his administration. Historically, only once (in 1934 during Franklin Roosevelt's first term) did a new president's party buck the midterm odds and gain congressional seats. On average parties in the position the Democrats are in this year lose about 25 House and half dozen Senate seats in the midterm election after a new president assumes office. The odds against the Democrats in 2010 are only made longer by an economy slow to rebound from the deep recession that President Obama inherited from George W. Bush.
Still, this year, unlike 1994, when Bill Clinton and the Democrats were in the same position that Barack Obama and his party are in now, the Democratic Party has a clear party ID lead within an electorate that is demographically restructured and more open to activist government than at any time since the 1960s. This provides the Democrats with a base sizable enough to withstand a GOP wave if they can rally that base and get it to the polls in November. That is the big political question this year. The answer to it will determine the outcome of the 2010 midterm elections and will, in the end, be far more crucial than what happens with "angry independents" or "engaged Republicans."
A recent posting on Pollster.com indicates that recent surveys suggest the Democratic base may, in fact, now be returning home, possibly accounting for the apparent recent tightening of the Gallup generic ballot. It is for this reason that Democrats should be encouraged by President Obama's post-Labor Day campaign appearances that are so clearly designed to appeal to and rally Democratic voters.
- Michael Hais's blog
- Login to post comments
Comments