There are currently two competing, and demonstrably false, claims about the 2010 midterm elections:
- The Republican claim that they would win the House if the election were held today is not true.
- The counter-claim, or at least implied counter-claim, by some pro-Democratic bloggers that Republicans have made no gains is also not true.
The truth is, instead, in the middle. Republicans have made gains, but Democrats still lead.
This can be easily demonstrated through a look at congressional generic ballot polling over the last four months of both 2005 and 2007, as well as a look at polling over the last two months of 2009 (that is, August 15th forward). Using archived data from Polling Report (for 2006), Real Clear Politics (for 2008), and a combination of Pollster.com and Polling Report for 2010 (Pollster.com for most polls, but Polling report for Daily Kos, CBS, and NBC polls), we can quickly see that Republicans are in a better position than they were in either 2006 or 2008, but that they are still clearly behind Democrats:
Democratic Lead, Generic Congressional Ballot, Autumn 2009, 2007 and 2005
| Methodology |
2010 |
2008 |
2006 |
| All Polls |
3.8 |
11.2 |
9.3 |
| 1 Poll per Pollster |
5.3 |
10.6 |
9.9 |
| All Polls, no Rasmussen |
6.0 |
10.0 |
9.3 |
| 1 Poll per pollster, No Rasmussen |
6.0 |
9.5 |
9.9 |
- "All Polls" means every poll from every pollster, including multiple polls from pollsters with multiple polls.
- "1 Poll per Pollster" means the most recent poll from every pollster that conducted one in the given time frame
- the next two lines simply repeat #1 and #2, eliminating all Rasmussen polls
According to every measurement, even when Rasmussen polls are removed from the equation, Republicans are in a better position than they were in either the autumn of 2005 or the autumn of 2007. Even when it comes to the minimum gain of 3.5% gain for Republicans--represented in the fourth metric--there is still a greater than 90% chance that it is a real gain and not a statistical fluke.
At the same time, in every measurement, Democrats still hold a significant advantage well beyond the normal margin of error for polling averages of this sort. Even when all polls from every polling firm are included in the average--a method that currently shows Democrats only ahead by 3.8%--there is still a greater than 90% chance that Democrats would win the national House popular vote if the election were held today.
The data used in these calculations can be viewed here: comparing generic congressional ballot polls, one year out, 2010, 2008 and 2006. Simple polling averages were used. However, as I will argue in a few days, if enough polls are used in the averages, such simple means can actually produce 20-30% most accurate results than other poll-based forecasting / snapshot methods.
The bottom line is that it is very, very likely that Republicans are in a better position now than there were in 2006 and 2008, and that it is equally likely they are still losing. This may not be a breakthrough conclusion, as most political observers probably intuitively sense this to be state of the current political environment. Still, I find that providing numbers to ground such discussions is always a useful exercise.
|
STATE BLOGS
