I guess I need to be nicer to you.

Very informative, Chris.

Does anyone have any clue why the Gallup poll on the front page of USA Today a couple days ago said that the Dems big congressional lead had evaporated, and that Bush was in the mid-40s? I haven't seen much individual information on races that bear that out, and every other national poll I have seen has also still had the dems with a comfortable lead.  

Minor point, but all four of those seats that switched R-D in 1994 were open seats.  No Republican incumbents lost that year.  Given the situation to date, I think we have reason to hope no Democratic incumbents will lose this time.  Not that we should take that for granted, but it does look unlikely at this point.

Chris,

Thanks so much for all the comprehensive, informative posts.  I learn so much here, and   really appreciate all you're doing.

I finally did the number crunching on the district lines, and it seems that the Republicans have a 3% advantage due to how the lines are drawn.  In other words, the Democrats will probably need to beat the Republicans by 3% in the nationwide popular vote to take back the House.  But if they can get more than that, they are likely to do that.

This is not just because of gerrymandering, its more a function of the Democrats piling up votes in urban and to a lesser extent majority minority districts, while losing by smaller margins in rural areas.

It gave us a 40-33 lead among all voters but 40-40 among likely voters: http://syndication.nationaljournal.com/o ncall/poll2.html

But it looked like the crosstabs were screwed up, giving Dems a 28-15 lead among indies. There can't be that many undecideds.

Hard for me to spend much time on House races because the polls are so infrequent and iffy, plus I don't know the dynamics of enough of the individual districts. Just give me 52% of the two-party vote.

Ive seen two generic polls recently that showed the GOP pulling even.  One was Hotline, the other was Gallup.  We need to make sure that this doesnt become a trend.

So, we could win a bunch of seats, and still see a small number of our incumbents go down.

If the right Dem incumbents go down (meaning those who vote with the GOP or are corrupt), then that's a gain for reforming the Democratic Party -- because in the next cycle good Democrats can be put into those offices.

You know what they say about breaking eggs to make an omelet.

It seems like the wind is shifting in favor of the Republicans for whatever reason, I don't know. The Abramoff scandal/Culture of Corruption theme seems to have died quickly. Rove indictment didn't happen.
Its gonna to be harder for Democrats to nationalize this election.

My prediction - House : Democrats pick up nine seats
                Senate: Democrats pick up four seats
         Governorships: Democrats pick up five seats

Yeah, I don't know why this is happening.  Probably something to do with the terror plot.  It almost seems like the wave crested too soon for us like in 1996.  However we should be able to pull this off:

House: +11   222R-213D
Senate: +4  51R-49D
Governorships: +6  28D-22R
State Legislative seats:  +150  3811D-3501R

If we can't manage this, we are in big trouble.  We need to work hard and get everyone we know to vote Democratic.  Let everyone know how important this election is.

Chris,

Check out the latest Keystone Poll. I don't recall them previously asking a question about if "YOU plan to vote for the Dem or R in YOUR US House district", but they did this time and the numbers are very interesting. 40% said they were going to vote for the Dem and 30% said they were going to vote for the R. Note also the majority of the survey respondents were in the central and west, the more conservative areas of the state.

Does anyone wonder that there may be a connection between Democrats CONSISTENTLY doing better on the generic than the general and the fact that we're now finding something like 75% of Democratic ads do not mention the affiliation of the candidate?

I did a series of three monthly posts looking at generic ballot lead compared to presidential approval, congressional approval and right track/wrong direction.  My point in doing that was to look at how the public was making connections in terms of overall political mood. My conclusion was that the Democratic Party wasn't doing anything effective to focus general discontent onto Congress.

This is a BIG failure, IMHO.  But it's no surprise at all.  The recent good news on the generic ballot front is, IMHO, simply a consequence of realworld forces--more bad news generally, and the elections getting closer.  

What really matters, as Chris notes, is how that generic lead translates into votes in specific districts, and the intermediate factor is the regional distribution.  This is why the news from Kansas, Nebraska and Montana has been particularly important--and why the Nevada Senate race matters a whole lot more than people seem to realize.

There is a real cultural split between the socially conservative South and the libertarian conservative West, with an even more Dem-friendly mix of the two in the prairie.  The most truly effective way to take advantage of this opening is via individual politicians, who bring the abstract possibilities down to earth in their specific actions, and personas.  The governors of Kansas and Wyoming are shining examples of this.  So, too, is John Murtha.

Creating Our Own House Battleground Poll Indices

In light of this analysis, I would suggest that we can somewhat mitigate the poverty of polls we face on the House by creating our own indices, composed by averaging the last poll from each district.  I would propose creating two of these, which would then be subdivided by region.  

One would be composed of all the polls taken in battleground districts.  We could either accept the Cook Report's list of districts, for example, or create our own list for doing this.  The second would be composed of a selected subset of districts where Dem candidates "get it."  This would be, in effect, an indicator of the success of the sort of political approach advocated here at MyDD and elsewhere in the blogosphere.

By aggregating polls from all these different races, we will create indices that we can then compare to other, less frequent polls.  It seems like a very sensible next step for us to take.

Anyone interested in joining in???

2. House campaigns, and even 1994-type realignments, are generally far less nationalized than understood. In 1992, Democrats won 46 seats with 55% of the vote or less. In 1994, when the national popular vote for the House shifted seven points in favor of Republicans, Democrats actually only lost 22 of those 46 seats. It was not a straight, national shift roughly equal across all districts--in other words, Republican gains were not equal across the nation. As I have documented in the past, it was basically Republicans completing the long, slow transition from the southern realignment of 1968-1972. If we have big gains, they probably won't come equally either. It will be district to district or, if it is a realignment, it will come from specific demographics that are concentrated in a few areas. It probably will not be not nationwide.

Watch Indiana this year, I think that

2. House campaigns, and even 1994-type realignments, are generally far less nationalized than understood. In 1992, Democrats won 46 seats with 55% of the vote or less. In 1994, when the national popular vote for the House shifted seven points in favor of Republicans, Democrats actually only lost 22 of those 46 seats. It was not a straight, national shift roughly equal across all districts--in other words, Republican gains were not equal across the nation. As I have documented in the past, it was basically Republicans completing the long, slow transition from the southern realignment of 1968-1972. If we have big gains, they probably won't come equally either. It will be district to district or, if it is a realignment, it will come from specific demographics that are concentrated in a few areas. It probably will not be not nationwide.

Watch Indiana this year, I think that a 3 seat pickup is going to happen absent a huge suprise, and it's the Great lakes region also.  I expect that people like Chris Shays and Santorum are going to get it, so that the Northeast becomes a monolithic block for the Democratic party the way that the Deep South (Texas through Georgia) has become for the Republicans.

And I think that Republican House majorities in Ohio and Indiana are going to switch to favor Democrats, and that Republicans in the Chicago suburbs are going to get canned.  

To parallel the comment about the Republican realignment of the South, I think that this year could be the year, the Democrats take the North, and make inroads in the  West (NV, NM, WY, CO, and AZ)  I wouldn't be suprised to see Democrats pick up 10-15 seats in the region that sweeps west from Buffalo to Chicago, and south from Detroit to Louisville.