Numbers wizard Nate Silver on the race for control of the US House:
In the House, by contrast, Republicans are reasonably clear favorites to maintain their majority.
Still, the House is probably more difficult to analyze than the battle for the Senate or the presidential race. Democrats, in my view, do retain something of an upside case that revolves around the possibility that anti-incumbent sentiment could manifest itself in a large amount of Congressional turnover.
A macro-level view of the House of Representatives usually begins with the generic Congressional ballot. Right now, that shows Democrats and Republicans roughly tied; about as many people say they plan to vote for the Democratic candidate for Congress as the Republican one.
It might be thought that if the generic ballot is tied, the seats in the House would be split about evenly, therefore implying that Democrats had about even odds of retaking the chamber. But that probably is not the case. Generally, the incumbent party in the House over-performs its vote share in terms of the number of seats it carries because of the way the vote is distributed from district to district, and therefore has a little slack in its numbers. A crude estimate we made last year is that Democrats might need to have a 3 or 4 point edge on the generic ballot - not a mere tie - to have even odds of winning the House.
It is plausible that the seat-by-seat breakdown of the House vote will be different this year if the incumbency advantage is smaller than normal.
However, Republicans have a very nice hedge against this contingency: redistricting. Republicans, who were in control of the redistricting process in far more places than Democrats, were able to shore up some of their vulnerable incumbents. In a true wave election, these defenses might prove feeble or even counterproductive, but in an election where there were merely a slight edge to Democrats, they could make the Republicans' majority considerably more robust.
Unfortunately, the Dems got the better of the GOP as far as redistricting in Texas, Illinois, and California and will pick up nearly a dozen seats. But that won't be enough and given the chances of a GOP breaktrhough in some seats where Dems are retiring, the Republican loss on election day should be minimal.