The PVI-Voting Pattern Index

(Phenomenal work. Promoted from the diaries with an edit to the title by DavidNYC.)

Lots of people in the blogosphere want to primary Bad Dems. Used wisely, that’s a good thing. It can wake up a Blue Dog on cruise control in a dark blue district (Jane Harman); sometimes, it can even lead to a victory and a Better Dem (Donna Edwards).

Unfortunately, I’ve never seen any sort of metric that helps us decide who’s a worthwhile primary target, i.e. who’s way out of whack with his or her district, versus someone who’s an odious Blue Dog but who’s the best we can manage in a red district and someone to be welcomed into the big tent (if sometimes secretly grumbled about). So, here’s my stab at it.

It’s based on this proposition: the representative in the most liberal district should have the most liberal voting record. The representative in the most conservative district should have the most conservative voting record. The representative in the 217th most liberal district should have the 217th most liberal voting record. And so on. Of course, in reality, it doesn’t always work like that. But finding the representatives whose voting records are severely mismatched with their districts’ lean helps us find the Bad Dems who need some prodding. And as a bonus, it also helps us find the Republicans who should theoretically be extra vulnerable in a general election: the wingnuts hiding in moderate districts.

To that end, I’ve developed the PVI-Voting Pattern Index. It simply rates every district from most Democratic to most Republican in its presidential preference, and rates every representative from most liberal to most conservative, and looks at the differentials.

Let’s start with the Dems who are underperforming their districts: in other words, the ones with super-safe districts who are voting in the middle of the pack, or ones with bluish districts who are voting conservatively.

Rep. District PVI PVI rank PP NJ Liberal rank Difference
A. Davis AL-07 D+17 62 91.77 61.3 179 -117
Meek FL-17 D+35 10 94.76 76.5 117 -107
Rangel NY-15 D+43 2 94.85 80.3 105 -103
Jefferson LA-02 D+28 26 95.14 73 121 -95
E.B. Johnson TX-30 D+26 29 93.97 77.7 123 -94
Fattah PA-02 D+39 4 94.95 82.2 96 -92
Lipinski IL-03 D+10 102 90.93 54.7 191 -89
S.T. Jones OH-11 D+33 16 95.28 79 104 -88
Doyle PA-14 D+22 39 94.58 75.2 125 -86
C. Brown FL-03 D+16 64 94.39 67.7 142 -78
R. Brady PA-01 D+36 8 95.35 84 83.5 -75.5
Rush IL-01 D+35 11 95.61 82.8 86.5 -75.5
Meeks NY-06 D+38 6 96.24 83.3 76 -70
Engel NY-17 D+21 44 93.12 83.3 111 -67
Lynch MA-09 D+15 68 94.07 72.3 135 -67
G. Green TX-29 D+8 124 91.21 56.5 189 -65
Towns NY-10 D+41 3 96.37 84.2 67 -64
T. Ryan OH-17 D+14 73 94.36 70.3 136 -63
M. Udall CO-02 D+8 121 90.94 60.2 184 -63
Serrano NY-16 D+43 1 97.54 78 63 -62
J. Cooper TN-05 D+6 139 85.7 54.7 200 -61
Berkley NV-01 D+9 115 92.01 61.8 173 -58
P. Stark CA-13 D+22 40 96.24 75.2 98 -58
Reyes TX-16 D+9 111 92.54 61.8 166.5 -55.5
Sires NJ-13 D+23 37 95.67 80.3 92 -55
Cummings MD-07 D+25 32 96.51 77.5 86.5 -54.5
Maloney NY-14 D+26 31 96.06 80 85 -54
Capuano MA-08 D+33 20 97.16 79.2 73.5 -53.5
Emanuel IL-05 D+18 57 95.96 72.3 110 -53
Barrow GA-12 D+2 178 68.22 45.8 228.5 -50.5

Frankly, I was surprised to see so many members of the Congressional Black Caucus on the list, especially since most of the ones on the list have pretty progressive records (although it certainly does shine the spotlight on the ones with more centrist records, like Artur Davis and Kendrick Meek… those two, plus Greg Meeks, are the only members of both the CBC and the New Democrat Caucus). The formula is pretty unforgiving, and it can be hard to live up to having one of the most Democratic leaning seats in the nation.

It may be more interesting to focus on people who have the less safe seats, but not so dangerous as to justify sketchier voting records. Mr. Lipinski you’re all familiar with, but this list also points, for instance, to Jim Cooper, a Blue Dog in a D+6 seat based in Nashville, or John Barrow, the second most conservative Democrat in the House, who sits in a D+2 seat.

Now, let’s turn to the good news: the Democrats who are vastly overperforming their districts. These are a mix of Blue Dogs holding down the fort in some of the nation’s most conservative districts, and flat-out progressives based in light-blue or swing districts.

Rep. District PVI PVI rank PP NJ Liberal rank Difference
C. Edwards TX-17 R+18 406 89.55 53.7 196 210
Pomeroy ND-AL R+13 360 90.61 61.3 183 177
Matheson UT-02 R+17 397 77.65 49.2 220 177
G. Taylor MS-04 R+16 393 73.46 46.3 225.5 167.5
Lampson TX-22 R+15 380 71.24 47 227 153
Skelton MO-04 R+11 339 89.50 54.3 195 144
Hinchey NY-22 D+6 142 98.02 92.5 12 130
Herseth SD-AL R+11 328 86.43 53.7 204.5 123.5
Price NC-04 D+6 141 97.42 90.7 24 117
Boucher VA-09 R+7 297 91.91 58.3 181.5 115.5
Spratt SC-05 R+6 279 91.54 63.7 170.5 108.5
Capps CA-23 D+9 112 98.29 94 6.5 105.5
Chandler KY-06 R+7 294 90.48 55.3 193 101
Filner CA-51 D+7 132 97.23 91.5 32 100
Holt NJ-12 D+8 128 97.42 89.7 28 100
Boyda KS-02 R+7 301 85.66 53.8 204.5 96.5
Holden PA-17 R+7 293 89.75 53.2 197 96
Mollohan WV-01 R+6 281 92.00 56.7 185 96
Shea-Porter NH-01 D+0 197 95.39 78.5 102.5 94.5
Grijalva AZ-07 D+10 107 98.53 89.2 15 92

Now, let’s look at the Republicans. A low OWI score shows which GOP representatives are, as it were, overperforming their lean of their districts, by being hardcore dead-enders in districts that should elect moderates, or better yet, Democrats.

Rep. District PVI PVI rank PP NJ Liberal rank Difference
Kline MN-02 R+3 228 2.11 9.3 400 -172
Feeney FL-24 R+3 236 2.18 12 396 -160
Chabot OH-01 R+1 201 5.25 17.5 353 -152
Mica FL-07 R+4 248 3.2 8.3 394 -146
Bachmann MN-06 R+5 266 1.48 10.8 405 -139
P. Ryan WI-01 R+2 218 3.97 20.3 352 -134
Garrett NJ-05 R+4 253 3.10 14.7 383 -130
Shadegg AZ-03 R+6 284 2.23 6.7 409 -125
Walberg MI-07 R+2 226 3.80 20.7 350 -124
Roskam IL-06 R+3 232 5.05 16.8 355 -123
Putnam FL-12 R+5 274 2.62 12.3 390 -116
Franks AZ-02 R+9 315 0.74 6.7 421 -106
S. King IA-05 R+8 311 1.12 8.8 414 -103
Gallegly CA-24 R+5 260 6.19 14 360 -100
Tiberi OH-12 R+1 205 8.12 25.3 304 -99
Fossella NY-13 D+1 186 11.51 29.7 285 -99
H. Wilson NM-01 D+2 173 19.26 31.7 272 -99
P. King NY-03 D+2 179 18.78 27.8 277 -98
M. Rogers MI-08 R+2 216 6.53 22.2 314 -98
Latham IA-04 D+0 190 13.10 26 287 -97

It’s heartening to see that a lot of these guys are already being targeted by the DCCC this cycle (Feeney, Chabot, Bachmann, Walberg, Shadegg). Also, I see more names that are on the “Races to Watch” lists of a lot of people here (Kline, Garrett).

Finally, let’s look at the Republicans who are more moderate than their districts would predict. There isn’t that much we can do with this list, other than thank them (or sit back and watch as the Club for Growth primaries them, as we saw with Wayne Gilchrest this year), so I’m just holding it to 10 names. It’s an interesting mix of representatives from red districts who’ve turned against the war, and very conservative representatives from super-red districts who just aren’t as crazy as the voters in their districts.

Rep. District PVI PVI rank PP NJ Liberal rank Difference
W. Jones NC-03 R+15 383 27.48 42 236 147
Simpson ID-02 R+19 408 18.92 33.2 270 138
J. Moran KS-01 R+20 413 12.69 34.7 275.5 137.5
Paul TX-14 R+14 379 16.72 39.8 259 120
Platts PA-19 R+12 357 21.76 38.5 251 106
Gilchrest MD-01 R+10 327 39.71 48.3 224 103
D. Young AK-AL R+14 368 16.05 35.8 270 98
Coble NC-06 R+17 403 5.61 27.8 305 98
Aderholt AL-04 R+16 387 10.20 26.8 290 97
Emerson MO-08 R+11 342 22.80 39.8 245.5 96.5

Some of the more observant of you might be wondering about how there’s one name missing… the name that’s come to signify that Bad Dems can, in fact, be beaten? Some of the others of you might be wondering who are the representatives who exactly match their districts, with almost no difference between their record and the district’s lean? Well, the answer to both of those questions is: Al Wynn. Somewhat surprisingly, for his 2007 voting record, he was right on the mark. MD-04, at D+30, is the 23rd most Democratic district. He had a Progressive Punch score of 97.14 and a National Journal score of 95, making him the 22nd most liberal representative. That’s a differential of 1.

So what does that mean? Did we target the wrong man? Nope… turns out, savvy politician that he is, he veered sharply to the left in 2007, worried by his previous narrow primary victory and anticipating the rematch. Here’s a chart that illustrates the U-turn: Progressive Punch also puts together lifetime scores, so I compared those against district lean too. Observe how most of the Lifetime Bad Dems are the same as the 2007 Bad Dems, but check out who’s #2 on the list:

Rep. District PVI PVI rank Lifetime PP Liberal rank Difference
Doyle PA-14 D+22 42 76.23 200 -158
Wynn MD-04 D+30 24 86.23 142 -118
Jefferson LA-02 D+28 27 86.54 140 -113
A. Davis AL-07 D+17 65 82.49 172 -107
Crowley NY-07 D+28 28 87.72 134 -106
Meek FL-17 D+35 10 88.70 116 -106
R. Andrews NJ-01 D+14 77 80.69 182 -105
R. Brady PA-01 D+36 8 89.25 108.5 -100.5
J. Moran VA-08 D+14 81 81.40 178 -97
Meeks NY-06 D+38 6 89.56 100 -94

Now, Wynn’s turnaround of 10.91, from a lifetime score of 86.23 to a 2007 score of 97.14, wasn’t the biggest gain of any Democrat… not by a long shot. The biggest jumps, perhaps unsurprisingly, were among the Blue Dogs, who had lower scores to begin with, and an opportunity to make up more ground by voting with the new Democratic majority. The biggest jump belonged to Ike Skelton, who went from 63.46 to 89.5, and the 3rd biggest jump was John Murtha, shooting up from 70.85 to 94.67.

In fact bigger jumps were shown by many of the other Dems subject to internal challenges or the threat thereof, often at the behest of the netroots: Jane Harman (who’s #13 on the Lifetime Bad Dem list) went from 79.08 to 93.32, Ellen Tauscher went from 84.01 to 95.97, and even Henry Cuellar beat Wynn, going from 74.25 to 85.24. (In case you’re wondering, only one Dem went in reverse: Nick Lampson went from a lifetime score of 75.28 to a 2007 score of 71.24, perhaps to compensate for his new very red district.)

So, there’s something here for both people from the “more” Democrats school and the “better” Democrats school to chew on. Primary challenges can sometimes be effective, but they have to be chosen wisely, and the best targets may not always be the most conservative. Also, there’s a lot of low-hanging fruit in the form of unpleasant right-wingers in light-red and swing districts; let’s get picking!

Methodological notes: Districts are rated according to PVI. There would be a lot of ties if I just used the Cook PVI ratings, so I took the actual vote totals in each district and used the PVI formula to recalculate PVIs to enough significant digits where aren’t any ties.

Representatives are rated according to the 2007-08 Progressive Punch scores and their 2007 National Journal composite scores, to smooth out any hiccups caused by one or the other compilers’ methods. To further increase the sample size I would have liked to also include DW-Nominate scores (which don’t cherry-pick votes but include all votes and measure representatives’ deviations from each other), but those won’t be published for the 110th Congress until it’s over in 2009, and to include ADA scores, but there isn’t enough differentiation there (for instance, 78 different representatives got a score of 95). Rather than averaging their two scores (where you run into an apples & oranges problem) and ranking the averaged scores, I ranked each representative from 1 to 422 on each score, then averaged their two rankings, then ranked the averaged rankings.

Here’s an example: Mike McIntyre is the midpoint for the House, at the 211th most liberal (or conservative) representative. He has a Progressive Punch score of 84.31, which is 208th. He has a National Journal score of 49.5, for a rank of 217.5 (a tie at 217th). The average of those is 212.75. An averaged score of 212.75 makes him the 211th overall. His district, NC-07, has a PVI of R+3.5271, the 230th most Democratic district in the nation. Subtract 211 from 230 for an OWI score of +19. Not much difference; he’s very slightly overperforming his district.

Wait… why are there only 422? Well, there are four open seats (those formerly held by Lantos, Jindal, Baker, and Wicker). Also, there are nine representatives for whom there wasn’t enough information to calculate a National Journal score (Richardson, Broun, Foster, Carson, Tsongas, Latta, and Wittman are all too new, while Cubin has missed too many votes and Pelosi ordinarily doesn’t vote). For each of these missing representatives, I also didn’t rank the district lean, so that the lean of open districts wouldn’t cause any distortion. For the last table on lifetime PP scores, there was a pool of 431, for which I re-ranked the lean of their 431 districts, so there may be some slight difference in district rankings compared with the other charts.

I fully acknowledge that these rankings don’t incorporate where their money comes from, who their back-room friends are, what they say in front of the cameras, how they frame things, or anything other than how they vote, but that’s the only thing that can be fairly quantified.

UPDATE: There have been some calls to make the whole database available, so I’ve done just that. It’s at Google Documents.

38 thoughts on “The PVI-Voting Pattern Index”

  1. Great idea and fine execution.

    One of the quirks with Progressive Punch is that it penalizes people who miss votes.  It is not a one-for-one thing but Lampson had a bypass operation, Eliot Engel missed a lot of time due to his mother’s illness and Tim Johnson is all the way up to a 49 this session (he was at 18 at one point).  Yvette Clark missed some time and went from the top of the pack to a poor score and now is rising again.

    I think they have worked with the formula a bit to reduce the effects of absences but I am not sure.

    One way or another, there is a good chance that Bill Jefferson will be out.  He has little money, lots of debt, and legal problems up the gazoo.  The new system with separate Democratic and Republican primaries rather than one “jungle” primary that incorporates all candidates with a possible runoff is the wild card.  Jefferson’s chances in a SDemocratic primary may be better but there is a good chance he will get a runoff in the primary (if he runs again)ratjher than a straight 50% + win.

  2. not only is he corrupt,  he’s a crappy fir for his district too.  Lets get this guy out of here.  

  3. And something I’ve wanted to do myself for a long time. I would love to see all the numbers graphed out on a chart, so that we can get a visual representation of who the most out-of-whack members are.

  4. … to Daily Kos as well?

    Peter Flom and I were working on a project about who should be primaried, and I think this is excellent to add to the list of criteria and provide some substitutions to the list that he came up with based solely on one Iraq vote (out of the dozens that were cast at different points in the evolution of the political climate in regards to the war).  If you don’t plan to make a diary about this (and that many of these House members deserve serious primary challengers) on Daily Kos, may I do have your permission to do so?

  5. skews the data in any way. I would guess there is a lot more spacing in PVI and PP/NJ at the ends then there is in the middle. I would think a minor change in the middle of these numbers could change a ranking 20-30 spots, and I wonder if anyone suffers or is helped by this.

    Maybe as a followup you could assign everybody a 0-100 value, one for PVI and the other for PP/NJ, based on whatever scale makes sense, and see how the numbers come out that way.

    Great job.

  6. I knew Feeney would be on the list. The guy actually carries around a Conservative scorecard in his pocket everyday and that area is only slightly Republican. It’s too bad we’re not going to get a really good Dem there, but I think Kosmas will probably win the seat.

  7. There are a few exceptions to these calculations. Mo Udall in Colorado is measured against his home district, but he’s known for many months — when did Allard announce his decision not to run again? — that he would be running for Senator. He’s been calibrating his votes to an entire state, not his 1 district of 7. Likewise Artur Davis, down home in Alabama, has ambitions to run for Senate. He’s looking at a constituency much more conservative than his 1 district of 7. To be fair, you’d really have to work your calculations for Colorado and Alabama, for their next elections, not the past.

    In Texas (and perhaps in Georgia), the Democratic members in swing districts may live in constant fear of still another gerrymander. I mean, the mid-decade thing worked once, why not try it again? With the Repubs controlling the entire state government, they could in theory once more redraw the lines against Lampson and Edwards and even Ciro Rodriguez. There’s a real nightmare! Now I know the political climate has changed, but if our guys’ votes give some signs of post-traumatic gerrymander syndrome, who can blame them?

  8. Aren’t the National Journal rankings the ones that magically determine that the Democratic presidential candidate just happens to be “the most liberal senator”? I’ve always had the impression that they were pretty bogus. What’s the advantage of adding them into the mix rather than just using Political Punch?

  9. It’s interesting to compare the effects of an incumbent protection re-districting to one where a lot of Republicans have at least theoretically competitive districts.

    California has a group of almost entirely progressive Democrats and reactionary Republicans.  Only one Democrat, the newly elected Jerry McNerney scores under 90 on Progressive Punch (88.63) and a bunch of Democrats like Ellen Tauscher and Jim Costa veered sharply to the left this session (Costa is still at a relatively conservative 901.10 score).

    California Republicans are in a word horrible.  The most interesting scoring came from Brian Bilbray who was initially moderate and now has veered far right (5.18 score) after his seat felt safe.  Bilbray’s lifetime score of 13.17 belies the fact that at the beginning of the session he was up around 17.  The only two Republicans to score over 10 include the ethically super challenged Jerry Lewis (14.77) who may be feeling unaccustomed heat and Mary Bono Mack (13.40).Doolittle, another endangered GOPer (who entered forced retirement) had a 9.32.

    People like Bilbray (5.18), Rohrabacher (5.08), Dreier (4.95)and Campbell (3.21) seem too conservative for their districts.

    Compare it to New Jersey where the seven Democratic districts are safe but five of the six Republican districts have at least some threat and even NJ-11 is not a super safe district.  Five of the six Republicans in Jersey score over 20 on Progressive Punch with four (all from dicey districts) scoring over 28 (Chris Smith, 36.43; Frank Lo Biondo, 33.86; Mike Fergusin,32.31; Jim Saxton, 28.35).  Frelinghuysen is at 20.20 and super out-of-touch Scott Garret is at an incredible 3.10.  

    The Jersey Dems all score at least 94.99 but Rob Andrews is another Al Wynn type who is voting more left to beat back possible challenges (94.99 vs. 80.61 for his career and that number was in the 70s).

  10. of course, any such method will have it’s flaws; one problem is that I am not sure there should be a linear relationship between Cook PVI and liberalness or safety of a district.

    If you have this in an Excel data set or some such, I could make some graphs….

  11. I’m bummed but not surprised in the least to see Bachmann and Kline on the list.  Kline’s district, while being conservative (conservative Republicans and somewhat conservative Democrats), his voting record is completely off.  His congressional district also has a majority of democrats sent to the state house and state senate, this one congressional district trending our way.

    Bachmann, being the 405 most liberal is not surprising either.  She is absolutely bat-shit crazy and I am so embarrased that she is from my state.  The district is definitely conservative leaning, Democrats and Republicans a like but God is she a weirdo.  The absolutely BEST attack ad that will be played ALL election season will be one of her molesting the president.  While I am not sure if this race will really be that close (she won 50%-42%, a comfortable margin for a top top top tier race), I really hope the DCCC at least airs an ad with that.  No one likes Bush and an ad with that will be quite powerful.

  12. This is an interesting & useful post, but I’m not at all surprised to see lots of the Congressional Black Caucus on the list.  You need to remember this when you make a list of this kind that:

    Just as not all Democratic representatives are not equally liberal, not all Democratic voters are equally liberal.

    Imagine that you have a “deep blue” district in San Francisco and another, even deeper blue district in the black belt of Alabama.  The Employment Non-Discrimination Act (ENDA), protecting gay rights, comes up for a vote.  It could take real political courage for the representative of the latter district to vote for it, while in San Francisco, it would be political suicide not to vote for it.

    Although that’s an extreme example, I think there are really two different variables at work in a given district: how uniformly Dem the district is (PVI being a decent measure of that) and how progressive those Dems are, which is not the same thing.  I don’t know what would be the most useful measure of the latter variable.  But you shouldn’t assume that just because a district is highly Dem-gerrymandered, those constituents are particularly liberal.

    In any event I tend to think that it’s more useful to use numbers as a starting point for real case-by-case analysis of whether we have an Al Wynn on our hands, and whether a more progressive Democrat could win in the district.

    But I’d be interested to hear about ideas for measuring how liberal (as opposed to how Dem-heavy) a given constituency is.

  13. what JR12 wrote about the PVI not necessarily being correlated to how progressive a district is, as a caution to those who will want to blindly use this excellent information you gave as a justification for a primary challenge, rather than a starting point of discussion.

    Going on a bit of a tangent here, this was one of the main problems I have with Howie Klein, great activist that he is, when he goes off on a rant of all the Blue Dogs, and throws people like Gene Taylor and Jim Matheson into the mix, telling his readers to flood their offices with angry calls, and yet he never seems to take into account the people they represent.  If we were to run someone who completely shares Howie’s values in some of those districts, watch that person get less than 20% of the vote.

    Not that they should get a free pass by any means, but geez, it seems like some in the blogosphere jump in automatic disgust anytime they don’t tow the Democratic Party line.  For me, I basically hold judgment on any social issue, because I’d assume those generally reflect what their constituents believe, whether we in the netroots agree or not with that point of view.  Something like voting against SCHIP is less forgivable, IMO.

    And I do agree with the above poster who said we should use z-scores, especially for the one where you said there was a big clump around 94-96%.  It would then be misleading to use a simple linear ranking system when you have clumps like that, as you stretch data out when maybe it shouldn’t be stretched out in the first place.

    But this is an excellent starting point for discussion on who we in the netroots should focus on.

  14. Folks,

    I think it would be interesting to filter the Bush Dogs list at Open Left… against the out of whackness index here.

    Just a few bits –

    The following “Bush Dogs” outperform:

    C. Edwards TX-17 R+18 406 89.55 53.7 196 210

    Pomeroy ND-AL R+13 360 90.61 61.3 183 177

    Matheson UT-02 R+17 397 77.65 49.2 220 177

    G. Taylor MS-04 R+16 393 73.46 46.3 225.5 167.5

    Lampson TX-22 R+15 380 71.24 47 227 153

    Herseth SD-AL R+11 328 86.43 53.7 204.5 123.5

    and I think should be given a break on the Bush Dog list. Perhaps someone with a bit more time can get into more detail and diary this over there.  

  15. I know ADA ratings have that problem; Clarke (who missed 2 of the 20 selected votes) and Engel (missed 3) had surprisingly low scores from them, and Lampson (missed 3) was down in the basement for Dems, although he voted the wrong way a lot of the time too. There isn’t enough transparency in the Progressive Punch and National Journal scores to see how missed votes affect the scores, but there are also enough votes thrown into their blenders that absences don’t seem to have much effect. (Clarke still managed to be the 43rd most liberal representative.) DW-Nominate scores, like I said, are probably the best, but they aren’t available for this Congress yet.

  16. David, hi.  Our treatment of absences is nuanced.  We treat an absence on a closed vote as a “bad” vote, but if the vote is not closed, then an absence is treated as neither good nor bad.  

  17. … the last sentence should be “If you don’t plan to make a diary about this (and that many of these House members deserve serious primary challengers) on Daily Kos, may I have

    your permission to do so?”

  18. What I took away were the observations on how in fact MORE Democrats MAKE BETTER Democrats.

    Obviously being in majority helped more than anything. The majority controls the agenda, and the leaders can send difficult bills off to the Committee on Nevermore.

    But also there’s a simple group think, or peer pressure, that will continue more strongly even if we ‘merely’ increase our majority.

    This group voting was easily seen in the Senate. On key votes on the War of Occupation and F.I.S.A., our Plains-Mountains Senators left us — all of them: Tester MT and Salazar CO, Dorgan ND and Ben Nelson NE, Bingaman NM and Conrad ND, Baucus MT and Johnson SD. We’ll see if adding Udall CO and Udall NM can tip this caucus to the left. Here pray for Rice OK to join them.

    In Dixie and Border States, when we lose them, we usually lose almost all: Landrieu LA, Pryor AR, Lincoln AR, McCaskill MO, Bill Nelson FL, and sometimes even Webb VA, Byrd WVa, and Rockefeller WVa. Mark Warner can’t change this caucus, Neal NC a bit, not Kagan NC, Musgrove MS not at all. We just don’t have enough chances to pick up seats in the South to nudge the group.

    But this study brings great good news from the House.

    The biggest jumps … were among the Blue Dogs, who had … an opportunity to make up more ground by voting with the new Democratic majority.

    Also, there’s a lot of low-hanging fruit in the form of unpleasant right-wingers in light-red and swing districts; let’s get picking!

    Yes indeed. MORE Democrats make BETTER Democrats, so let’s go get more — by beating Republicans.


    In fact bigger jumps were shown by many of the other Dems subject to internal challenges or the threat thereof, often at the behest of the netroots: … Harmon, Tauscher, even Henry Cuellar [went] from 74.25 to 85.24.

    Yeah, slim pickings and slow going from primarying Dems. Count ’em, one, two, three, plus Wynn of course, four. We’ll see about Lipsyncka.

    And are we sure it was potential threat of primary that moderated Cuellar? No one’s name got serious mention in ’06 or ’08.

    Congressman Henry Cuellar is … also a Majority Senior Whip.

    Well, well. A minor leadership position for this turkey, maybe more peer pressure to vote with the majority, he’s a whip now, after all. But not disputing Cuellar knows which way the wind blows. In Hispanic South Texas, as Ciro’s win demonstrated, it’s been blowing Democratic.

    BTW When calculating the PVI of Cuellar’s TX-28, Ciro’s TX-23, Doggett’s TX-25, Smith’s TX-21, and Hinojosa’s TX-15, keep in mind that these five districts were redrawn by court order in August of 2006. I doubt that the PVI has been recalculated from the old borders in place in 2004 and 2000. Roughly speaking, Ciro Rodriguez’s district became a few points more Hispanic and thus more Democratic. Lloyd Doggett’s became much less Hispanic and slightly less Democratic. Cuellar’s TX-28 became probably slightly more Dem (and much less dominated by his home base of Laredo). Hinojosa, little change. Smith’s TX-21 became much more Repub, thus dooming the campaign of Netroots Favorite John Courage.

  19. planning to cross-post it, probably on Monday (after giving it a few days here to make sure that no glaring methodological problems get pointed out). But feel free to grab anything from this diary and work it into your own materials. I’ll probably do some more work on this topic when I get some more time, so keep your eye out. (One thing I’d like to do is incorporate some key votes into the equation somehow, to make up for the fact that Prog Punch and National Journal seem to give the same weight to, say, S-CHIP as they do to below-the-radar procedural stuff. I don’t know if there’s a way to build that into the formula, though. Maybe I’ll need to build a separate Bad Dem Bad Vote matrix.)

  20. I thought of the same sort of issue with this analysis.  I think the way I’d measure this is to find the mean and standard deviation of both PVI and PP/NJ for Democratic seats. Then compare the Z-score of each district with its congressman.

    In any event, this is a really neat post, and I’d love to play around with the data if the author has the data in a spreadsheet.

  21. example of what you’re talking about is probably PA-14 (Mike Doyle’s district), which is about D+22, but it’s Pittsburgh, so it’s dominated by older, Catholic, socially conservative Democrats (people who, by virtue of demographics, might well be Republicans anywhere else, but are Democrats because of heavy union concentration and, well, tradition). So his district is functionally one that’s less liberal than its PVI implies, and he acts like it.

    Also, you might point to places like Seattle or Portland as maybe being more ‘liberal’ (in the sense of the Pew typologies, if you’re familiar with that approach) than, say, African-American majority districts where there’s a larger concentration of Democrats (or at least a smaller concentration of Republicans), but where more of the people who vote Democratic are more socially conservative.

    (I’ll look for somewhere to stash the database online so that anyone can access it. If I can’t fit it, I’ll just grab your e-mail from your Kos profile and send you a copy. It’s an Excel document.)

  22. you may have noticed was on the Top 10 Lifetime Bad Dems list, but he’s reverted back to the middle of the pack in terms of out-of-whackness based on 2007 numbers. I don’t know if he’s feeling the heat (I would suspect not because he’s pretty machine-protected, right?), feeling more peer pressure now that he’s in the majority, or ramping up his liberal cred with an eye toward Lautenberg’s seat (although it looks like he’ll have to wait another 6 years).

    Bilbray’s lifetime score is a lot higher because he used to be in a mostly different district with almost no geographic overlap (the one now held by Susan Davis, which is now something like D+10). Strange that he veered so far right, though, since I think he ran a pretty rhetorically moderate campaign when he took over his current seat.

  23. Almanac of American Politics has the newly revised PVIs and vote counts for the redistricted districts in Texas and, more importantly, Georgia, where every district got thoroughly reshuffled.

    Anyway, the peer pressure point is a very important one, especially when you look at the flipside and see the breakdown in Republican discipline now that they’re in the majority (as seen with how the GOP moderates’ Progressive Punch scores shot up from the teens, usually to the 20s and 30s).

  24. I was worried, when I was doing this, that there would be a lot of amplification of relatively insignificant differences in Progressive Punch scores (note that there are a lot of people clustered in the 94-96% range, so a few variances on, say, some minor farm bill amendments could push you down enough ranking slots that it could make or break your appearance on the list. But mostly I’m concerned with the real outliers, which is hopefully who’s mostly on the list). I wouldn’t say, for instance, “Representative X got an OWI score of 20 but Representative Y got a score of 10, so Representative X is clearly much better,” since that sort of difference would be based on insignificant differences in PP score or in district lean (like a D+3.2 vs. a D+3.5).

    Anyway, I’ll be very curious to see if your approach leads to particularly different results. I’ll see if there’s a way I can let people access the Excel document online; if there isn’t, you can just go to my Kos profile (I’m also Crisitunity there) and find my e-mail there, and I’ll send you the document if you send me your e-mail.

  25. into hardcore nerdsville here, but I tried this approach and the results weren’t very satisfactory… it seems to overemphasize PVI and underemphasize Progressive Punch scores because the highest and lowest PVIs tend to be more standard deviations away from the mean PVI than is the case with Progressive Punch scores. (For instance, the most Democratic district, NY-16, at D+43, has a Z-score of 2.95, while the most liberal representative, Tammy Baldwin, at 99.26, has a Z-score of 1.05.)

    So the new ratings (I’m assuming ‘compare’ means subtract the PVI Z-score from the PP Z-score) tend to just very closely mimic if you just listed all districts from highest to lowest PVI, with Progressive Punch scores making very little of a dent on the ratings. If that makes sense (I’m not a statistician by trade, so I may be phrasing things inelegantly… in fact, I had to look up the formula for Z-score before doing this).

    For PVIs, the mean was 0.6514 (setting a baseline of 0 and treating R+ scores as negative) and the standard deviation was 14.0176. For Progressive Punch scores, the mean was 54.8531 and the standard deviation was 42.3564. So here are the top five underperformers:

    Rep PVI Z-score PP Z-score Differential
    Seranno 2.95 1.01 -1.94
    Rangel 2.85 0.94 -1.91
    Towns 2.72 0.98 -1.74
    Fattah 2.67 0.95 -1.72
    Clarke 2.61 0.99 -1.62

    And here are the top five overperformers:

    Rep PVI Z-score PP Z-score Differential
    Edwards -1.36 0.82 2.18
    Pomeroy -0.99 0.84 1.84
    Matheson -1.27 0.54 1.81
    Taylor -1.25 0.44 1.69
    Skelton -0.86 0.82 1.68

    I need some sort of an additional step to smooth out the apples & oranges problem between what the two Z-scores are measuring. Thoughts? (Also, the diary has been updated to include the link to the database via Google Documents, if you want to crunch numbers yourself.)

  26. Apparently there’s no penalty (or not as much penalty) for missed votes as there is with ADA ratings or Progressive Punch ratings. So if you’re only showing up for the really key votes because you’re spending most of your time on the campaign trail, you’re going to have a distortedly liberal record… or a distortedly conservative record, if not voting has the same effect on the calculation as would voting conservative. You can see this at work by comparing Obama’s 2007 Progressive Punch score (#40 in the Senate) vs. his National Journal score (#1 in the Senate). PP vs. NJ is pretty consistent for almost everyone else in Congress.

    Anyway, I included both to expand the sample size and smooth out distortions caused by any methodological quirks used by one or the other.

  27. are interesting, too.

    In some cities (e.g. New York) there are a lot of CDs that are entirely in the city; in others (e.g. Chicago) most contain suburbs.

    But is, say, NY-15, which is mostly Harlem and gets D+43, really much more liberal than, say IL-02, which includes part of the South Side of Chicago, but also suburbs, and gets D+35?

    Or MI-14, that includes a lot of Detroit, but also some suburbs, and gets D+33?

  28. Glad to know that the Almanac has the goods on those new district. And I was glad to see that Ciro Rodriguez was not on your lists, not too hot, not too cold, just about right for his district, at least until he gets settled in.

  29. in some of the comments above that some districts with a higher PVI are less ‘liberal’ than districts with lower PVIs. Pew, in their Typologies study from a few years ago, has tried to quantify the different flavors of Democrats (and Republicans and indies) at the national level; the distinctions they made echo what you’re saying (that San Francisco falls into the ‘liberal’ category, while the Black Belt would probably fit in the ‘Disadvantaged Dems’ category). The distinctions also kind of echo the Obama/Clinton split, where PVIs are useless in predicting voter behavior.

    Unfortunately, they’ve only polled this at the national level. Polling this at a statistically significant level for each district would be a gigantic undertaking (national sample size 25,000?), but if they ever did it, that would be a blockbuster. Here’s the three Dem typologies and their numbers:

    Liberals (secular and anti-war)  17% general population, 19% registered voter population

    Disadvantaged dems (‘social welfare loyalists’) 10% general and registered voters

    Conservative Dems (‘latter-day New Dealers’) 14% general, 15% registered

  30. Another example of the kind of district you’re talking about is my own: Ohio’s 17th. While the district’s western third (Akron and Kent) is fairly liberal although not exceedingly democratic, the eastern 2/3rds (Youngstown-Warren) is very democratic but not all that progressive on social issues (large population of Old-line Catholics and ethnics). That being said, Tim Ryan is probably the most progressive elected official we have.

  31. Mo Udall represented the Tuscon area from 1961 to his death in 1991. His son, Mark, represents the Boulder, CO area now.

  32. … and general state of mental confusion. Thanks for the correction.

    Anyway, I’m sure Mark Udall has been — and will be — called many worse things than “Mo”!

  33. to add a link to the database via Google Documents, if you were interested in doing a scatter-plot of the data or some other advanced tinkering. (I don’t have fancy stats software, just Excel.)

  34. Crisitunity-I love the work you are doing here using  We’ve created Progressive Punch to empower the activists community as well as average citizens.    Actually, using your own methodology, I’d suggest using our “Chips are Down” category because they’re the closest votes &/or votes where progressives were on the losing side. There’s quite a bit more gradation among the Democrats using these scores than there is using the overall progressive scores.  Go to and click on the “Chips are Down” title to get an explanation of our methodology in computing those scores.I strongly invite you to email me at to communicate about some relatively sophisticated efforts we’ve made comparing members’ voting records to their districts that we haven’t yet posted in a dynamically updateable fashion on the web due to costs of programming and our strapped level of funding.  

    Needless to say, everyone’s in love with their own product.  But I think National Journal’s ratings are bizarre because they’re solely rankings without any percentages attached to them. So if 40 Republicans score zero percent liberal and the 41st scores 1 percent liberal, then he/she is miraculously the 41st most conservative member of Congress.

    [See also the methodological note I posted above in response to David’s comments about how we treat absences on votes.  There’s also extensive info on our methodology from the “What is a Progressive Score” link on the lower-left corner on any page of]

  35. I don’t think he’d mind being called by his father’s name. His dad was a fine congressman, and I think he’d be proud to be confused with such a person. The person who confused them, though, might experience some embarrassment ;-).

Comments are closed.