Last week I wrote about Where We Can Make the Most Progress in the House, where I tried to measure the biggest probable right-to-left swings that might result from the 2008 election in the House (the biggest one would be swapping out Dana Rohrabacher for Debbie Cook, in case you missed it). In the comments, the question came up of what would happen if I ran the same analysis for the competitive 2008 Senate races.
I have been reluctant to do so, because when I did the House, I relied on a demographic model for predicting which caucus new House members might join and, from there, predicting their likely DW-Nominate score. That just doesn’t work with the Senate: demographics-wise, states don’t break down as cleanly as do House districts. And Senators tend to disobey their states’ partisan lean much more so than Representatives: consider that we have two moderate Republican senators in one of our bluest states (Maine), two populist Democratic senators in one of our reddest states (North Dakota), and the swing states in the middle give us as wide a range of personalities as Russ Feingold and John Sununu.
So, I decided to try a different approach, more speculative than I generally prefer, where I tried to project prospective Democratic senators’ voting preferences by averaging out the scores of already sitting senators who seem to have some commonalities with them. As a vague rule of thumb, I tried to use one senator who had the most in common geographically and one who seemed to have the most in common ideologically and/or stylistically, although these categories pretty thoroughly blended. Here’s a case in point: Mark Warner. I decided he had a lot in common with Jim Webb (- 0.359), a fellow Virginian and someone else who knows how to connect with white working class voters, and also with Joe Biden (- 0.338), another mid-Atlantic senator with a mix of liberal intentions and pro-corporate leanings. Average those, and voila: Warner projects at – 0.349. (Some of my comparables, or the resulting scores, may strike you as completely misguided. This is all pure speculation on my part, so feel free to argue why in the comments, or ask for some clarification on a particular choice. My feelings won’t be hurt.) One exception: if the Democratic candidates have a House record, I used their most recent score from there.
| State | 110th Sen. | 110th Score | 111th Sen. | 111th Score (and Comparables) | Difference |
|---|---|---|---|---|---|
| OK | Inhofe | 0.766 | Rice | -0.392 (Dorgan + Casey) | -1.158 |
| CO | Allard * | 0.636 | M. Udall | -0.375 (110th Congress) | -1.011 |
| MN | Coleman | 0.178 | Franken | -0.746 (Klobuchar + Feingold) | -0.924 |
| NH | Sununu | 0.481 | Shaheen | -0.442 (Leahy + Feinstein) | -0.923 |
| TX | Cornyn | 0.557 | Noriega | -0.336 (Bingaman + Salazar) | -0.893 |
| NM | Domenici * | 0.281 | T. Udall | -0.525 (110th Congress) | -0.806 |
| NC | Dole | 0.451 | Hagan | -0.330 (Webb + Lincoln) | -0.781 |
| NE | Hagel * | 0.376 | Kleeb | -0.366 (Tester + Dorgan) | -0.742 |
| ID | Craig * | 0.457 | LaRocco | -0.242 (103rd Congress) | -0.699 |
| OR | Smith | 0.155 | Merkley | -0.698 (Wyden + Whitehouse) | -0.698 |
| KY | McConnell | 0.507 | Lunsford | -0.168 (Pryor + Ben Nelson) | -0.675 |
| AK | Stevens | 0.260 | Begich | -0.360 (Tester + Bingaman) | -0.620 |
| MS | Wicker | 0.465 | Musgrove | -0.147 (Landrieu + Ben Nelson) | -0.612 |
| VA | J. Warner * | 0.258 | M. Warner | -0.349 (Webb + Biden) | -0.607 |
| ME | Collins | 0.084 | Allen | -0.449 (110th Congress) | -0.533 |
| KS | Roberts | 0.376 | Slattery | -0.151 (103rd Congress) | -0.527 |
As you can see, swapping Jim Inhofe for Andrew Rice is the biggest gain (probably in terms of IQ points as well as in terms of voting record) even without factoring in that Rice may be more progressive than my score gives him credit for. However, unlike the House, where there are a fair number of opportunities to replace a right-winger with a progressive, in the Senate we’re pretty much limited to replacing right-wingers with moderates, or moderate Republicans with progressives, so the shifts are smaller.
Finally, you may notice asterisks for the GOP-held open seats. I’ve compiled a separate table that doesn’t focus on “progress” but the “what if,” i.e. how big a swing we’re looking at in terms of the potential replacement (each of whom I’ve projected to be more conservative than the guy they’re replacing, either based on their House record or on comparables). If you prefer to swap these numbers in for the numbers based on the retiring senator, this doesn’t change the order of the overall results too much, although it does highlight the importance of making sure we win our biggest races. (Especially in Colorado… Mark Udall is on the moderate side, and hasn’t endeared himself much lately with his Iraq and FISA votes, but he’s way to the left of Allard, and even more so than Schaffer, who it turns out was one of the most conservative members of the House during his time there.)
| State | GOP Sen. | GOP Score (and Comparables) | Dem Sen. | Dem Score | Difference |
|---|---|---|---|---|---|
| CO | Schaffer | 0.849 (107th Congress) | M. Udall | -0.375 | -1.224 |
| NM | Pearce | 0.557 (110th Congress) | T. Udall | -0.525 | -1.082 |
| VA | Gilmore | 0.543 (J. Warner + DeMint) | M. Warner | -0.349 | -0.892 |
| ID | Risch | 0.547 (Craig + Kyl) | LaRocco | -0.242 | -0.789 |
| NE | Johanns | 0.399 (Hagel + Grassley) | Kleeb | -0.366 | -0.765 |
that is a scary thought for Coloradans.
I really like your use of comparables — you have some great choices in there. The only one that I have a minor quibble with is Mark Warner being comparable to Webb/Biden. If anything, I might go with Biden and Bill Nelson — or maybe a guy like Tom Carper.
I’d imagine that Musgrove would actually be further right than either Landrieu or Nelson, and Hagan would probably be much more like Lincoln than like Webb, whilst LaRocco and Slattery would probably be a little more loyal than they were previously (they’ll have much less cover now that the boll weevils and co. have disappeared). But as a rough analytical tool, this is great.
I was under the impression that Johanns was a far-right. If he is in the Grassley/Hagel mode I can live with that. Of course I’d rather Kleeb, but that ain’t happening.
I agree with all of them for the most part. I could use different combos but I dont think it would change much.
The only thing that made me go, wtf was Mark Udall’s voting record! That really isnt that great. Was it more liberal before and he is just lowering it to run for Senate? If so, pansy…..
Great work. I know you didn’t want to do this but I think it looks spot on.
By the way. I won the OpenLeft American Blogger contest with a idea for a project to give PIE! (Progressive Investment Effectiveness) scores to candidates based in part on your “Where we can make the most progress” series. Is it OK if I use that as part of the formula? I’ll be sure to give you credit.
I’ve tried using both Save and Reset buttons now, and neither of them is saving my signature, for some reason.
Sorry if posting this as a comment is a bad place, but I don’t know where I should post technical questions about the site.
But I do agree that Johanns feels more like a Thune than a Grassley.
than Dorgan and definitely Casey. He would be a earth-shaking change from Inhofe.
Yeah, use whatever you need, and let me know if you need any other data or troubleshooting or anything! Nice netroots-self-referential name on the PIE scores too.
Great analysis as usual!
I’m surprised “Big John” Cornyn scored (relatively) low. I thought for sure he’d be a .7 or higher…
Does anyone know if American Political Polling ever released any info about the OH-1 poll they said they were gonna do in the release for their previous poll in OH-2? Granted, that poll was a total joke, but who here isn’t hungry for an OH-1 poll?
And for that matter, why haven’t there been any polls for Shays vs. Himes or anything in New Jersey?