Swing State Project

• KY-Sen: The online “moneybomb” technique seemed to help Rand Paul get a lot of traction in the early months of the Republican primary, but his latest few scheduled moneybombs have been fizzles. Yesterday’s was a case in point: $258K is still a lot of dough for two days, but it was far short of the planned $400K. (And every penny counts: Paul reported only $319K CoH at the end of June.) Are Paulists feeling tapped out these days, or are his early-adopting libertarian-minded nationwide fans chafing that he’s increasingly parting ways with his dad and becoming more of an NRSC sock puppet? (As seen just today with their divergence over Cordoba House.)

• PA-Sen: Two pieces of good news in Pennsylvania: first, Joe Sestak is rolling out an endorsement from Chuck Hagel, the former GOP Senator from Nebraska who was on the short list for an Obama cabinet position. Not that Hagel is probably a household name in the Keystone State, but it certainly burnishes Sestak’s bipartisan (and military) cred. Also, the DSCC is planning to spend even more on this race, letting Sestak, presumably, keep marshaling his resources for a big late push (the kind that helped him shoot past Arlen Specter in the closing weeks in the primary). They’re spending $546K on broadcast media this week, on top of $494K last week; the DSCC’s total spending and reservations so far in this race amount to $4.4 million.

• WI-Sen: Hmmm, maybe all that sunspot and Greenland stuff actually started to stick. Republican wingnut candidate Ron Johnson is trying to undo the crazy on his various climate change comments, offering the usual last refuge of the pathetic walkback, that his remarks were “taken out context.” Johnson is also trying to walk back his previous support for elimination of the home mortgage interest deduction, about as popular a provision of the tax code as there is. If you wanted to simultaneously set back both the entire real estate business and the middle class by several decades, eliminating the home mortgage interest deduction would be a good way to do it.  

• IL-Gov: The exodus from the field of broken dreams that is the Pat Quinn camp just continues. Today, the departee is CoS Jerry Stermer, and while it isn’t a rancorous departure (as was the case with Quinn breaking it off with David Axelrod’s media firm last week), he is resigning to avoid being a distraction with an ethics investigation (over whether he sent political e-mails from his government account) hanging overhead. With Quinn seemingly circling the drain, that ain’t gonna help.

• AZ-05: In the “ooops, spoke too soon” department, presumed Republican primary frontrunner David Schweikert opted for a last-minute ad buy for $14K on cable, apparently worried that he was losing momentum going into tomorrow’s primary. Last week, his camp had said he was so far ahead they were just going to go dark and start saving money for the general. Meanwhile, next door in AZ-03, Steve Moak, the only candidate who rivals Ben Quayle on the money front, is out with a new ad hitting Quayle over his pseudonymous tenure writing for the website that was the precursor to TheDirty.com.

• FL-24: Wow, Craig Miller actually went there: he sent out a last-minute mailer going after the widely-known, but not yet broached in the campaign context, issue of GOP primary opponent Karen Diebel’s mental stability. Maybe he’d been planning to do a last-minute boom-lowering all along, but it seems kind of strange, as Miller’s millions have seemed to have the race locked down, not requiring him to get his hands dirty. Was there a last minute Diebel surge (she’s been attacking him on being too soft on immigration)?

• TX-23: Republican challenger Quico Canseco is out with an internal poll, via OnMessage, giving him a small lead over incumbent Dem Rep. Ciro Rodriguez. The poll gives Canseco a 43-37 lead, a turnaround from Rodriguez’s lead of 48-45 in a Canseco poll from June.

• Ads: Bobby Bright’s out with his first ad, in AL-02. SSP southern correspondent Trent Thompson’s description of the ad says it all, so I’ll just quote him: “”Bobby” voted against everything Obama. Also, he kisses a baby.” Ben Chandler is also out with an ad in KY-06, touting his saving the local VA Hospital. Finally, Rick Boucher’s GOP opponent, Morgan Griffith, is also out with a new TV spot, reminding you that he’s a conservative. (As an interesting ethno-linguistic observation, one clear indicator that VA-09 is an Appalachian district, not a Southern district, is that Griffith cites “Warshington” as being the source of all our problems.) NWOTSOTB, in all three cases.

• We Ask America: The quirky little Republican-linked pollster that could, We Ask America, is out with an array of polls in Illinois (the state where they’re based). They find Mark Kirk leading Alexi Giannoulias 39-33, but they also look at three House races. In fact, they looked at these races back in March, so we now have trendlines. In IL-11, they find Adam Kinzinger leading Debbie Halvorson 52-32 (down from an already bad 42-30); in IL-14, they have Randy Hultgren leading Bill Foster 44-37 (also down, from 38-36). The good news: they have Dan Seals leading Bob Dold! in IL-10, 43-40 (Seals led 40-37 in March, so he’s holding his own). We’ve also discovered a few recent WAA odds and ends that we didn’t cover earlier: they also see Roy Blunt leading Robin Carnahan 47-43 in MO-Sen, and see Sean Duffy leading 42-33 over Julie Lassa in WI-07.

• Rasmussen:

• AL-Sen: William Barnes (D) 28%, Richard Shelby (R-inc) 60%

• AR-Gov: Mike Beebe (D-inc) 53%, Jim Keet (R) 33%

• MD-Sen: Barbara Mikulski (D-inc) 55%, Eric Wargotz (R) 39%

• TX-Gov: Bill White (D) 41%, Rick Perry (R-inc) 49%

Talk Business Research/Hendrix College for Talk Business (8/17, likely voters, no trend lines):

Chad Causey (D): 32

Rick Crawford (R): 48

Ken Adler (G): 4

Undecided: 16

(MoE: ±3.9%)

Ouch; what a brutal starting position for Chad Causey. Crawford, the beneficiary of a slam-dunk primary win, has an eye-popping 43-13 favorable rating, a much higher mark than the battle-scarred Causey’s 29-26. Each man is unknown to nearly half of the electorate, so there is still quite a bit of room for this race to evolve, but shifting it back away from the GOP will be a tough task. For his part, Causey will have to hope that Crawford’s past bankruptcy troubles become a bigger deal than that issue is today.

Note that this sample has a sample composition of 33% Republican and 31% Democratic — probably a more GOP-friendly sample than we’d be used to seeing from this ancestrally Democratic district, but maybe indicative of the folks who will actually be showing up to the polls this November.

UPDATE: The Causey campaign has responded by releasing their own internal poll (available below the fold) by Garin Hart Yang (8/16-18) showing the race at a much closer margin of 41-40 for Crawford. Details below.

Public Policy Polling (8/21-22, likely voters, 7/16-18 in parens):

Rick Scott (R): 47 (43)

Bill McCollum (R): 40 (29)

Undecided: 13 (28)

(MoE: ±5.6%)

Kendrick Meek (D): 51 (28)

Jeff Greene (D): 27 (25)

Glenn Burkett (D): 5 (6)

Maurice Ferre (D): 4 (4)

Undecided: 13 (37)

(MoE: ±5.4%)

With the Sunshine State’s primaries tomorrow, three major pollsters have rolled out last-minute predictions today. For the Democratic Senate race, everyone’s in agreement, except for maybe the magnitude of the victory. However, for the Republican gubernatorial primary, there are some divergent results, and PPP seems to be the odd man out this time. Contrary to the general trend of this race lately (I don’t know if there’s ever been a clearer illustration of “peaking too early” than this graph of Pollster trendlines), they give Rick Scott a 7-point lead over Bill McCollum. It’s still an improvement for McCollum over their July numbers, where he trailed by 14.

PPP’s GOP sample gives McCollum favorables of 38/45, while Scott’s actually above water at 46/33. Did they manage to find a group of voters who somehow have avoided the last few months’ worth of attack ads about Scott’s gigantic Medicare fraud?

Quinnipiac (8/21-22, likely voters, 8/11-16 in parens):

Bill McCollum (R): 39 (44)

Rick Scott (R): 35 (35)

Undecided: 22 (19)

(MoE: ±3.5%)

Kendrick Meek (D): 39 (35)

Jeff Greene (D): 29 (28)

Maurice Ferre (D): 3 (6)

Undecided: 28 (29)

(MoE: ±3.6%)

Quinnipiac finds a small lead for McCollum (smaller than they did a week earlier, showing that McCollum’s late surge seems to have maxed out); their pool of GOP voters gives 39/37 faves to McCollum and 31/40 to Scott. Interestingly, they also find a much smaller lead for Meek than did PPP, and freakishly high undecideds (28%) for an election that’s, y’know, tomorrow, indicating how little motivation the Dem primary seems to have generated.

Mason-Dixon for Miami Herald (pdf) (8/17-19, likely voters, 8/9-11 in parens):

Bill McCollum(R): 45 (34)

Rick Scott (R): 36 (30)

Mike McAlister (R): 4 (3)

Undecided: 15 (33)

(MoE: ±4.5%)

Kendrick Meek (D): 42 (40)

Jeff Greene (D): 30 (26)

Maurice Ferre (D): 4 (5)

Glenn Burkett (D): 1 (NA)

Undecided: 23 (28)

(MoE: ±4.5%)

Mason-Dixon’s numbers are a few days older than PPP’s and Quinnipiac’s; they’re pretty closely in line with Quinnipiac, although they see the biggest lead of all three for McCollum over Scott (and more momentum, compared with last week). They give McCollum 43/32 favorables, compared with 33/40 for Scott. Could this truly be the end of the line for shameless bald supervillain Scott? (The Lex Luthor comparisons have written themselves this cycle — but to me Scott’s always been the Bizarro World version of Peter Garrett, the very liberal, very earnest, very tall, very bald Australian Labor Environment Minister… who those of you who were listening to music in the 80s probably remember better as the singer for Midnight Oil.)

UPDATE: Was it worth it? Pre-primary campaign finance reports came out, and between Scott and McCollum, $70 million was spent on the GOP gubernatorial primary: $49.9 million from Scott, $21 million for McCollum (although the majority of the money spent on McCollum’s behalf was from allied outside groups). No worries: even after that spending, Scott still has a net worth of $218 million.

There’s a reason Greenland was called Greenland. It was actually green at one point in time. And it’s been, since, it’s a whole lot whiter now.

Someday I’ll have me a penthouse, stacks and stacks of folding green… Here are the July fundraising numbers for the six major party committees (June numbers are here):

SurveyUSA for KING-TV (8/18-19, likely voters, 4/19-22 in parens):

Patty Murray (D-inc): 45 (42)

Dino Rossi (R): 52 (52)

Undecided: 3 (7)

(MoE: ±4%)

Let’s see… we could believe this SurveyUSA poll, which seems to be one of a growing string of west coast outliers for them (leaving them fully 11 points to the right of Rasmussen), or we could believe the poll taken on Tuesday, that had 0% undecideds, 0% leaners, a MoE of 0%, and an n of probably 1.6 million (1 million so far, with only two-third counted). I’m talking, of course, about Washington’s Top Two primary, which, because of its unusual all-parties-in-one-pool nature, functions as essentially the most accurate poll you’re going to see taken all cycle.

The usual rule of thumb in the Top 2 primary is to project the total Dem and total GOP percentages out toward November. The current individual totals (with only 2/3ds of votes counted, though) are still Patty Murray 46, Dino Rossi 34, Clint Didier 12, but the real story is that the total Dem and GOP votes are essentially 50-49 right now, with a lead of 10,000 for all GOP candidates (Rossi, Didier, Paul Akers, and a whole bunch of anonymous weirdos) over all Dem candidates, out of more than a million votes. So for this SurveyUSA poll to be right:

a) Dino Rossi would have to consolidate every GOP vote behind him — every Didier vote, every Akers vote, every Norma Gruber vote, and so on — pick up every vote from every third-party or no-party candidate, pick up every vote for all the other hapless Dems who ran in the primary (including every Goodspaceguy vote and Mike the Mover vote), and then somehow turn around 1% of the electorate who voted for Patty Murray in the primary to vote for him instead,

a1) and that’s all presuming that the 46-34-12 percentages don’t change, although they most likely will, in a Murray-favorable direction by another percent or two, as the majority (1200 out of 2000) of outstanding precincts still to report are in traditionally slow-to-report King, Pierce, or Snohomish Counties… or

b) the number of Dems participating in November would have to be smaller, rather than bigger, than the number particpating in the primary… despite the fact that Dems had no major incentive to participate in Tuesday’s primary, seeing as how there weren’t any noteworthy Dem-on-Dem primary battles above the state legislative level, compared with the intensely fought Republican Senate contest in the primary and several others in House races. (In other words, the Democratic share in November is likely to go up from the primary, not down, when it’s actually for all the marbles instead of an academic exercise.)

Unlike a lot of SurveyUSA crosstabs, there isn’t the frequently-present quirk of young people loving the Republicans (here, the 18-34 set goes for Murray 50-49). Instead, the strangest number is that Murry and Rossi are tied in “metro Seattle” 48-48. That would be approximately true if “metro Seattle” were limited to suburban Pierce and Snohomish Counties (where current counts from Tuesday are 50 all GOP/48 all Dems in Pierce, and 50 all Dems/49 all GOP in Snohomish), but King County (which has a population greater than Pierce + Snohomish combined) is at 60 all Dems/37 all GOP. So, no, they aren’t tied in metro Seattle.

Also worth noting: SurveyUSA seemed to misunderestimate Murray’s vote share in their pre-primary polls (their last one saw a 41-33-11 primary), despite being close enough to the election to include a number of “actual voters,” i.e. those who’d already mailed in ballots. PPP, by contrast, came closer to nailing the primary from several weeks further out (predicting 47-33-10). SurveyUSA’s pre-primary sample didn’t also include general election matchups (their last general election trendlines are from April), but that same PPP sample that was quite close on the primary also projected a 49-46 race in favor of Murray in November (closely matching Rasmussen’s 50-46 win for Murray, with leaners pushed, from this week).

Also, PPP, in their sample several weeks ago, found that Rossi isn’t on track to consolidate all Didier and Akers voters; they were planning to go for Rossi by an 82-11 margin. That’s even more complicated by what seems to be an increase in tensions between the Rossi and Didier camps in the last few days, rather than any moves toward unity, after Didier gave a list of demands on Friday before he’d endorse Rossi. After Rossi shrugged that off, the Didier camp started dropping f-bombs in response to questions from local politics website Publicola:

Didier’s spokeswoman, Kathryn Serkes was more candid with us:

“So is Dino saying, ‘Fuck you’ to those people [who supported Didier]? ‘Fuck you, I don’t need your votes? I can win with 33 percent.'”

UPDATE: In my more cynical moments, I think that it could be that this whole conflict was scripted ahead of time, professional wrestling-style, in order to help Rossi burnish his moderate credentials by refusing to be held hostage by the teabagger, as he now has to sprint back to the middle. Somehow, though, it feels like it has that spark of autheniticity.

Public Policy Polling (8/14-16, likely voters, 6/19-21 in parens):

Dan Onorato (D): 35 (35)

Tom Corbett (R): 48 (45)

Undecided: 17 (20)

(MoE: ±4.1%)

In the other half of the Pennsylvania sample that saw a big drop in Joe Sestak’s numbers against Pat Toomey because of PPP’s inevitable jump from a registered to likely voter model, the news isn’t quite as bad for Dan Onorato… but that’s mostly because he wasn’t doing very well to begin with. He loses only three points’ worth of ground, compared to Sestak’s nine. We’re getting to the point where we can’t blame Onorato’s failure to close within single digits on his unknownness, as this sample has about the same number of no-opinions for him (30/28 favorables) as it does for Corbett (33/23); it seems to have more to do with the anti-Dem nature of the year (which seems disproportionately strong in Pennsylvania), as well as the downdraft eminating from Ed Rendell (current approvals of 27/63).

One other point of serious concern for Pennsylvania Democrats: PPP did a generic House ballot test within the state, and it has a 48-39 advantage for the GOP. (That contrasts with a generic House ballot with a 46-40 Democratic advantage in Illinois, also from this week’s PPP sample.) With a lot of the state’s Democrats concentrated in just a few districts in Philly and Pittsburgh, that points to serious potential trouble for more than just the most vulnerable seats (the open seat in PA-07, Paul Kanjorski in PA-11) to some of the other ones too. While a statewide generic ballot isn’t of much more predictive value than a nationwide generic ballot, it certainly suggests that, say, Kathy Dahlkemper and Patrick Murphy need to be at the top of their games this cycle.

I know it’s been almost 6 months since my last post. Time flies when you’re…working a lot; ok, I was having some fun, too.

A couple of people have sent in partisan data, which I’ve uploaded. North Carolina has been there since early June (thanks Peter!), and New Mexico is there today (thanks Neil!). And I fixed the VTD data for Indiana. (Thanks to all of you who reported that bug.)

Also, I’m ready to accept other partisan data from volunteers! Here’s how.

If you are able to track down 2008 presidential data that somewhat matches the 2000 voting districts (or census block groups) for a state, you can put that data in a .CSV file (comma separated values) and email it to me. The trick is that voting districts often change and many have between 2000 and 2008. Some states don’t have voting districts published by the Census Bureau, so that’s a problem. If you want to do more work, you can try to map the 2008 data back through any changes from the 2000 voting districts.

For CA, NY and TX, JeffMd created new shapes to (roughly) match 2008 voting districts. In general, I would not recommend this.

Anyway, check out the web page describing what to do and we can get more partisan data in the app so we can have more fun until the new 2010 census data comes out.

BTW, I will attempt to make some improvements to the app and will get 2010 census data in there after it comes out.

Thanks!