Both Gallup and Rasmussen show a wide split between the two parties with three weeks to go:

Among voters Gallup estimates to be most likely to vote at this point under either a higher- or lower-turnout scenario, Republicans maintain substantial double-digit advantages. In Gallup's higher-turnout scenario, Republicans lead 53% to 41%. In Gallup's lower-turnout scenario, Republicans lead 56% to 39%. These likely voter estimates are based on respondents' answers to seven turnout questions, with the results used to assign a "likelihood to vote" score to each registered voter and, in turn, to create hypothetical models of the electorate based on various turnout scenarios.

In addition to turnout, independents' voting intentions are a critical determinant of the midterm election outcome - particularly relevant, given that more than 90% of Democrats and 90% of Republicans say they will vote for their party's candidate in the elections. At this point, independents tilt strongly toward the Republican candidate in their district, helping shift the race in the GOP's direction.

Rasmussen has the GOP with an 8 point advantage - bigger than their last survey.

Prognosticators have GOP landslide totals for gains inching upwards into the mid 30's in the House and 6-8 seats in the senate, although there are still another 4 Democratic seats "in play." At this rate, it would appear that prior to the election, almost all major pollsters will show a GOP gain sufficient to take back the House.

As long as no one on the Republican side gets complacent...