D: 36,538,605 R: 33,452,888 I: 1,652,436
These are preliminary totals that do not include results from some 52 uncontested races, 43 of which were Democratic victories. When we have those numbers, we can expect a quite higher total for Democrats.
The two-party vote split (thus far, missing all those uncontested votes) is 52.2% to 47.8%.
As a very rough estimate, we'd be adding about 3.5 million more votes to the D column and about 740,000 to the R column.
As for the two-party vote: for now, including estimates for the uncontested races, it's somewhere around 54% to 46%.
This is based on [(average district turnout/2) x uncontested races per party]. The average district turnout was 164,699, based on preliminary results. I somewhat arbitrarily halved that number, given turnout will be quite low in these districts - unless a hot Senate race or something contentious down-ticket could bring out voters.
Stay tuned for the releases of Dubious Democracy, in which we analyze the competitiveness of the election and accuracy of resulting representation, and Monopoly Politics 2008, in which we'll predict the outcomes of over 340 U.S. House races... for 2008!