"Although the balance of power in the U.S. House of Representatives will be determined by 435 individual congressional elections, Gallup's generic-ballot measure of national support for the two major parties -- more specifically, the final pre-election generic ballot based on likely voters -- has proven to be a strong predictor of the actual percentage of votes cast nationally for all Republican and Democratic candidates. This, in turn, bears a close relationship to the number of seats won by each party."
...
Results are based on telephone interviews with 1,625 national adults, aged 18 and older, conducted June 15-19, 2008. For results based on the total sample of national adults, one can say with 95% confidence that the maximum margin of sampling error is ±3 percentage points.
For results based on the sample of 1,460 registered voters, the maximum margin of sampling error is ±3 percentage points.
Results for likely voters are based on the subsample of 1,310 survey respondents deemed most likely to vote in the November 2008 general election, according to a series of questions measuring current voting intentions and past voting behavior. For results based on the total sample of likely voters, one can say with 95% confidence that the maximum margin of sampling error is ±3 percentage points. The likely voter model assumes a turnout of 60% of national adults. The likely voter sample is weighted to match this assumption, so the weighted sample size is 974.
http://www.gallup.com/poll/108454/Democrats-Favored-Retain-House-November.aspx
