« Netroots August Fundraising Push Round-up Charts | Main | TX-23, TX-10: Ciro Dropping Out; Ankrum Poll »

Wednesday, August 30, 2006

New PVIs from Cook Political Report

Posted by DavidNYC

Every four years, the Cook Political Report publishes a new Partisan Voting Index based on presidential election results at the congressional district level. What exactly are PVIs? Here's the explanation (sub. req.):

In August of 1997, The Cook Political Report introduced the Partisan Voting Index (PVI) as a means of providing a more accurate picture of the competitiveness of each of the 435 congressional districts. Using the 1992 and 1996 major-party Presidential voting results, the PVI measured how each congressional district performed compared to the nation as a whole.

Using the results of the 2004 elections, we have updated these PVI ratings and have even more information to draw upon to understand the congressional level trends and tilts that will help to define upcoming elections.

Developed for The Cook Political Report by Clark Benson and Polidata Inc., the index is an attempt to find an objective measurement of each congressional district that allows comparisons between states and districts, thereby making it relevant in both mid-term and presidential election years. While other data such as the results of senatorial, gubernatorial, congressional and other local races can help fine tune the exact partisan tilt of a particular district, those kinds of results don't allow a comparison of districts across state lines. Only Presidential results allow for total comparability. A Partisan Voting Index score of D+2.3, for example, means that in the 2000 and 2004 presidential elections, that district performed an average of 2.3 points more Democratic than the nation did as a whole, while an R+3.8 means the district performed more Republican than the nation.

PVIs are only available to owners of or subscribers to the Almanac of American Politics or the Cook Report. (However, the underlying data - presidential vote by CD - is available publicly both at CQ Politics and at PoliData.) For political junkies, though, a subscription to one (if not both) are practically a necessity. (Just as long as you ignore Michael Barone's increasingly wrong-headed conservative CW in the Almanac.)

Because PVIs rely on presidential vote by CD, whenever redistricting takes place, the PVIs have to get adjusted. As you well know, this used to be a once-a-decade affair. But Republican chicanery in Georgia and Texas has been keeping the number-crunchers busy. Fortunately, Cook has released new PVIs based on the mid-decade redistricting in Georgia and the most recent round of re-redistricting (or should that be re-re-re?) in Texas necessitated by violations of the Voting Rights Act.

While I won't republish all the new PVIs here, I can tell you how things have changed in the three most competitive districts in TX & GA.

GA-08: Marshall (D)
Old PVI: R+3
New PVI: R+8

GA-12: Barrow (D)
Old PVI: D+5
New PVI: D+2

TX-23: Bonilla (R)
Old PVI: R+13
New PVI: R+4

In any other year, the changes in Georgia would make me pretty nervous. But as you can see from Superribbie's newest compilation, the pros don't think all that highly of the GOP's chances in either GA-08 or GA-12. (GA-08, by the way, used to be numbered GA-03 until the redistricting.) However, you can bet that Barrow will get a serious challenge one of these years, when the GOP's fortunes change.

As far as TX-23 goes, well, I'm just not sure how much of the new district is actually new to Bonilla. And let's not kid ourselves: R+4 may look small compared to R+13, but it still means the district went for Bush by about 55-45 over Kerry. So I wouldn't get your hopes up for a surprise pickup there.

Overall, 12 of 13 districts in Georgia changed their partisan composition, and 5 of 32 in Texas did as well. The biggest remaining changes are in GA-1, where a possibly swing district has become extremely Republican, and TX-25, where Lloyd Doggett's heavily Democratic seat is now in swing territory. In fact, I'd expect a hefty challenge to Doggett soon, too. And while the news isn't good for John Courage - Lamar Smith's very Republican seat did become a little bit moreso - but I doubt it's going to make much of a difference.

But if you want the full run-down on all the changes, you're going to have to subscribe yourself. And I highly recommend doing so.

Posted at 04:40 PM in 2006 Elections - House, Redistricting | Technorati

Trackback Pings

TrackBack URL for this entry:


Jim Marshall's new district is extremely ugly. I can't believe he's no longer considered very vulnerable against a top-tier challenger like Mac Collins.

How much worse did Doggett's district become? I didn't realize things got that hairy for him after the redraw.

Posted by: Mark [TypeKey Profile Page] at August 30, 2006 08:24 PM | Permalink | Edit Comment | Delete Comment

Doggett got pretty hosed - D+14 to D+1. A safe Dem district to a clearly swing district. He caught a break this year, because it's too late to find a serious challenger. But in 2008, he's going to be a prime target.

I wonder if people just don't realize how much more GOP Marshall's district became. And like I say, if he survives this year, he's going to have a hell of a time going forward.

Posted by: DavidNYC [TypeKey Profile Page] at August 30, 2006 10:49 PM | Permalink | Edit Comment | Delete Comment

I know Georgia's gotten really ugly, but if he (Marshall) can solidify himself this time around, I bet (hope?) he can turn into one of those solid Southern Dems that win year after year in a conservative district. Examples: Bob Etheridge, Mike McIntyre, John Spratt, Lincoln Davis, Bart Gordon, Bud Cramer, Gene Taylor, and Ike Skelton.

Posted by: nada [TypeKey Profile Page] at August 31, 2006 12:34 PM | Permalink | Edit Comment | Delete Comment

Marshall's numbers illustrate the insanity of using Presidential Election results to compare districts across the country. In 2004, John Kerry was a Masschusetts liberal running against a good old boy from the South. In addition, national Democrats and the Kerry campaign made no secret of the fact that they had conceded Georgia (and much of the South) immediately after the primaries ended and the general election began.

Of course the Democratic presidential number is going to suck. These people have not seen a Democratic presidential ad or a Democratic president visit their area since 1996!

I have long argued on Dailykos and MyDD (where they are obsessed with PVI) that broader Democratic performance is much more important, as it takes into consideration the local and statewide character of the parties and politicians that actually run competitive campaigns in the state.

To illustrate the PVI is pretty useless, a good question is how can Jim Marshall be considered a pretty safe bet to win a +8 seat, but a better question is why can't we win all those D+ seats in Connecticut and Pennsylvania? And the answer is, the local Republicans have a much better reputation than the Texas version. And you quickly see how useless this index is.

Posted by: chris [TypeKey Profile Page] at September 1, 2006 05:18 PM | Permalink | Edit Comment | Delete Comment

Gosh, I never suggested that PVIs were the be-all, end-all of electoral analysis. But in any event, calling them "insane" or "useless" is not exactly the best way to convince people of the rightness of your position.

PVIs are neither the be-all, end-all of political analysis, nor are they worthless. But it's hard to take seriously an argument that chastises one extreme alleged view and replaces it with another.

Posted by: DavidNYC [TypeKey Profile Page] at September 1, 2006 07:06 PM | Permalink | Edit Comment | Delete Comment

I'll tone down my rhetoric a little. Of course they aren't completely worthless, we aren't winning many R+15 seats and they aren't winning many D+15 seats. But once you get inside the +/- 5 region they are a lot less useful than actual Democratic performance. There are so few seats that are actually in play that Democratic performance wouldn't be too hard to come by.

I can tell you in Georgia the Marshall district's Dem performance is actually about 48% (over a 3 cycle average) and the Barrow district's performance is about 54% (over the same period). Those are the only two competitive seats in Georgia, now what are Connecticut's actual Dem performance, factoring in how the local Republicans do there. I'd be much more interested in that, or even a project that tries to collect all of this stuff.

Posted by: chris [TypeKey Profile Page] at September 1, 2006 10:32 PM | Permalink | Edit Comment | Delete Comment