I'm putting this on the front page just to remind everyone of this new, regular feature on Open Left. The National House ballot is updated nearly everyday, although usually the updates do not appear on the front page. Make sure to come back to Open Left everyday, to check out the latest numbers! You can find them in the middle column--Chris
Ras has R's plus 7 while nearly every other poll is closer to the exact opposite (D's +6 or 7). The fact that Ras contributes 4 numbers to this really skews things. Wouldn't it make more sense to just use the latest Ras instead of all?
A: Rasmussen's extensive weighting in the average (four of the thirteen polls) does not bother me. According to the research I have done, including multiple polls from the same firm produces much more accurate results than including only the most recent poll from each firm.
Admittedly, some of the tightening can be explained by The Economist discontinuing their weekly poll. As such, I know there is a weird ontological problem with my method, in that the polls are creating reality in addition to measuring it. It's not like reality changed just because The Economist stopped measuring it, after all. (Then again, The Economist sometimes seems so haughty, they might disagree with that assessment.)
Still, I have yet to see a better forecasting method anywhere, so I am sticking with it.
Until the final six weeks of the election, take the simple mean of virtually all polls where the majority of interviews were conducted over the last 30 days.
During the final six weeks of an election, use the simple mean from the last 15 days.
Do not include Zogby Interactive polls and Columbus Dispatch polls, due to their horrendous past performance and questionable methodologies.
Do not include Strategic Vision polls, as it is starting to seem likely those are not real polls.
Include campaign-funded polls. Further, if there is more than one poll from a single organization, include all of them.
The basic idea is to cram as many polls with sound methodologies into the averages as possible, and weight them evenly to include more overall data in the sample. Because voter preferences don't really change that much in high-profile elections, I thought this method might produce a more accurate result through logic of regression to the mean. It seems to work pretty well, as my research has shown so far.
This is different from my 2006 and 2008 methodology in that it includes polls from 15 days out from an election, instead of only 8. Further, campaign funded polls, and multiple polls from a single polling firm, are now included. All of these changes were made to include more polls in the averages, since my previous methodology was about 10-20% less accurate than Pollster.com and fivethrityeight.com. Since they had already raised the bar so high, and since they will probably improve their methodologies for 2010 even more, it was time for Open Left to step it up.
STATE BLOGS
