Princeton Election Consortium

Innovations in democracy since 2004

Outcome: Biden 306 EV (D+1.2% from toss-up), Senate 50 D (D+1.0%)
Nov 3 polls: Biden 342 EV (D+5.3%), Senate 50-55 D (D+3.9%), House control D+4.6%
Moneyball states: President AZ NE-2 NV, Senate MT ME AK, Legislatures KS TX NC

The House – new, with less democracy!

November 9th, 2012, 2:00pm by Sam Wang

Today, Speaker John Boehner stated that his party’s leverage comes from the fact that it retained control of the House. Yet they lost the popular vote. How can this be?

Before the election, I predicted that even if more people voted for Democratic House candidates, Republicans could still retain control. The reason I gave was redistricting since 2010, which has tilted the playing field significantly. The prediction was correct – though if anything, I underestimated the effect.

I estimated that Democrats would have to win the national popular vote by 2.5% in order to have a 50-50 chance of gaining control. I also predicted that the House popular vote margin would be D+0.0%, for Democratic gains of 2-22 seats . As of now, counting the leader in each undecided race, the new House will be 235 R, 200 D, a gain of only 7 seats. ThinkProgress reports a popular-vote tally of 50.3% D to 49.7%, a margin of D+0.6%. Both results are within range of my prediction.

However, this is quite notable. The popular vote was a swing of more than 6% from the 2010 election, which was 53.5% R, 46.5% D. Yet the composition of the House hardly changed – and the party that got more votes is not in control. This discrepancy between popular votes and seat counts is the largest since 1950.

Did I underestimate the tilt of the playing field? Based on how far the red data point is from the black prediction line, the “structural unfairness” may be higher – as much as 5% of the popular vote. That is incredible. Clearly nonpartisan redistricting reform would be in our democracy’s best interests.

Incidentally, some readers have suggested to me a reform in the Electoral College so that each Congressional district votes for its elector directly. As you can see, such a rule change would allow redistricting to influence the fairness of the Electoral College. Winner-take-all state races occasionally cause a problem, but which party gains is somewhat variable. It seems that there are worse things than the status quo.

Tags: 2012 Election · House · Redistricting

136 Comments so far ↓

  • 538 Refugee

    I was going to ask if you still stood by the 2.5% tilt after all the numbers were in. I read a site that said the Republicans are gleeful that the civil rights act gives them some cover in drawing districts that keep minority vote ‘segregated’ so to speak.

  • JasonB

    No argument about redistricting, but do the raw numbers really tell us anything? Could the vote be concentrated such that balanced districts would not have changed anything?

    • Richard Vance

      No way to tell unless you had some preferred balanced district lines to bounce off of. The number of combinations of possibilities would be huge. I guess the best you could do would be to use the 2008 district lines and compare their votes in 2012 but how do you do that without knowing those added to and taken from the districts voted in 2012?

    • Ira S

      The new-Jim Crow gerrymandering is worst in the South – and how it passed the test of the Civil Rights Act is beyond me. Take South Carolina, which voted overall 55% R and 45% D. With 7 House districts, that should break out about 4 R and 3 D. But the Republicans have created one district which captures a huge proportion of the black/Democratic population, so that one district votes 90% Democrat, while the other 6 around 55% Republican, so instead of a 4/3 split, you get a 6/1 rout outrageously out of proportion to the popular vote. With the net effect that one white Republican vote is worth 3 black Democratic votes. How did this pass the scrutiny of the Civil Rights Act voting provisions? And though the South is the worst example, Ohio and Pennsylvania are just as bad.

  • Paul G

    ThinkProgress has 2 caveats to their calculation:

    1) Votes are still being counted, and many more votes have been added since 2 days ago (though if anything this should help the Democratic margin).

    2) Some district had 2 candidates of the same party, and they counted both of those candidates’ votes, but did not count the votes of candidates who ran unopposed. This does not really seem justifiable to me.

    I think more interesting statistics would be things like how many congressional districts Romney won despite losing the popular vote by 3% and how many districts have Republican PVI versus Democratic PVI, as well as how many districts have strong Republican PVI versus strong Democratic PVI (however you define strong).

    Certainly the fact that the Republicans claimed 21 out of 29 US house seats in the blue states of Virginia and Pennsylvania is an anti-democratic travesty.

    • Richard Vance

      Pennsylvania was a well oiled Republican dirty tricks machine.. without the PA supreme court intervening they could well have taken PA for Romney.

  • Dell Martin

    On another subject…since you have only a few comments, perhaps you read and respond to them…how do you account for Nat Silver’s rather large errors in the Montana and North Dakota Senate races? If he had insufficient data to make a good prediction, perhaps he should have a category for “who knows?”

    • Michael S

      Nate puts fingers on the scale. He used the wrong fingers.

    • Joel

      @ Dell

      Sam probably won’t answer your question, because it’s been addressed elsewhere. But here’s my take on it: Nate Silver introduces additional (potentially redundant) variables to his electoral estimates. These variables are based on a state’s past electoral history, economic factors, etc. This type of modeling is very useful where there’s very little polling data available, and in the case of North Dakota, that might have been the case (however, a median analysis would have nailed that race anyways). Montana was a different story, because there was plenty of polling pointing to a Tester victory. But Silver figured in what he calls “state fundamentals” anyways, which switched a lean Tester race (based on the polling) to a lean Rehberg race. This was a mistake, IMHO.

    • Dell Martin

      I have been posting about this, and have not seen any good explanation, or even an attempt to explain. I could see that what threw Nate off so badly were the “state fundamentals”…seems that he should revise his use of this variable, and he should explain to his incredibly wide following what went wrong. He nailed so many races right on the money that when he misses he needs to account for the miss.

    • Dell Martin

      Oops…I had not read all of Sam’s “Feed Carl Rove” article…yeah, he does address these errors. And, certainly, Huff Post’s predictions (based entirely on polling) had both races as Dem wins.

    • Neal J. King

      The other one that unexpectedly went down to the wire: Nevada Senator’s race, Berkley (D) vs. Heller (R).

      I wish someone had high-lighted that one: I had given up Berkley for a no-hoper, but she came very close. A bit more funding might have put her over the top – and made Adelson’s largesse a batting average of 000 flat.

    • Sam Wang

      I am kicking myself over that one.

  • Michael

    Quoting from the link:

    “The second caveat is that these numbers include several California districts where two members of the same party ran against each other, and they do not include districts where a single candidate ran unopposed. ”

    I hope that we can eventually get a look at the real total vote count. Not just because of the caveats cited, but also because I’m very interested in the rate of undervoting in house races. It seems like a lot of people don’t “turn out” in these elections even when they’re already standing in the voting booth. Of course, it’s hard to blame them when an incumbent is running unopposed or with only token opposition in a rigged district.

    • Michael

      Paul must type faster than I do.

    • Paul G

      I think that is indeed a big issue with counting the “house popular vote” – these rigged districts affect candidate recruitment and quality, and some candidates actually run unopposed. That’s why I’m really interested in the PVI’s – if we knew in how many districts Romney performed better than his national percentage that would give some clue as to the Republicans’ built-in advantage.

      In any case, it’s really offensive to hear Republicans bragging about how Americans voted for a Republican House right after abusing their power to deliberately discount the votes of Democratic constituencies. In fact I think the Constitutional arrangements for geographic representation both in the House and the Senate are extremely anachronistic and anti-Democratic to begin with, and gerrymanderin only makes it worse.

    • George M

      I don’t vote for any unopposed candidate, Republican or Democrat. If they didn’t win a primary they don’t get a vote.

  • Lanny

    What’s more, the Republicans in the gerrymandered districts have every incentive to remain obstructionist and obstinate, and little incentive to compromise. In their highly monochromatic districts, they worry a lot more about challengers from their right than do they about challengers from their left.

  • Elliot

    Please continue following this story, which is truly important, and completely ignored by the conventional media.

  • Michael S

    There’s a master’s thesis out there on how gerrymandering towards an advantage can over time become one’s own trap.

    • Kerr

      Can we get hold of this thesis? I would be interested in reading it.

    • Peter

      I’ve always understood 2006 to have been so dramatic in part because of overreach in gerrymandering by the Republicans.

    • JJ

      @Kerr Just look at California Republicans over the last 30 years. Gerrymandering insulated the nativist extremists from losing elections while changes in demographics tipped everything else to the Democrats because of those extremists policy position. This phenomenon is particularly pernicious because it is a positive feedback loop where extremism causes moderate losses making the parties elected representatives on average more extreme which leads to more moderate losses and so on. In California the Democrats now hold all state wide elected offices as well as 2/3rds majorities in both state houses. It’s a common right wing meme that California has been a disaster over the last 20 years because of Democrat policies but I think it is actually an example of how bad things get during the period when a party has become irrationally extremist but still has the power to block policy. Now that Democrats can raise taxes without Republican votes things will likely improve dramatically, though the lack of a credible governing alternative is likely to mean escalated levels of corruption and near corruption that single party scenarios usually entail. The national Republican party has entered this same feedback loop and is unlikely to escape if it doesn’t in the next four years. If the Republicans don’t escape this loop then the obstruction of the last four years will be a permanent feature of US politics for a generation. In this scenario the most important thing to allow government to continue to function is ensuring simple majority rule in both houses. In other words, filibuster reform and then nonpartisan redistricting.

    • Matthew M

      Interesting point, JJ. If I can offer a tl;dr version, it’s that we’re likely to see unintended (and unexpected) consequences to the Repubs’ gerrymandering frenzy over the next several years.

  • dsm

    Any plans to look at this on a state by state basis?

  • Nick W

    North Carolina is a shining example of the effects of republican redistricting. Pop statewide congressional vote was around 2.194 million D to 2.188 million R, and assuming McIntyre keeps his lead it will be 4 Dem seats to 9 Repub seats. This is a change from 7 D/6R.

    • Helen Bedd

      Ohio is even worse…from cycle to cycle the vote never goes much more than 54-46 to 46-54, yet in a state Obama just won, the GOP took 12 of the 16 house seats.

  • Rui Da Luz

    How about this: Congress passes a law which says that beginning with the next Census , and within some time certain thereafter, new districts should be drawn, and their shape should randomly generated (taking into account the one person, one vote principle, of course) through a pre-approved (by Congress) algorithm.

    • Ralph Reinhold

      It would require a constitutional amendment because that is one of the things left to the sates.

      How about a district’s geographic boundaries have to align with existing local government boarders or be whole contained within a government’s borders. That is, any line drawn from east to west cannot be less than 1/2 the maximum extent from east to west. The same applies to north south segments.

  • Pat

    Here is an interesting link discussing where the totals might land.

  • don in fl

    these guys are still talking momentum and sandy.maybe you and silver and others should defuse it.or should we just let them ramble on?

  • E L

    I have some sympathy for Nate on Montana. I live there. Rehberg had won election to our 1 House seat by wide margins in 12 elections. Tester had beaten a washed up and increasingly deranged (alcohol) Conrad Burns of the “rag head” slur on the floor of the Senate. The polls were sparse and poor quality in our state. They showed Romney +9. Karl Rove SuperPac dropped a lot of money in this state per capita. However, about a week before the election, PPP conducted a poll which found Rehberg’s approval rating in the mid-30s as I recall. That’s when I realized that Tester message about Rehberg had sunk in and he would win, probably easily.

    • Sam Wang

      That is all very interesting, but direct polling data is king. All that demographic and campaign whatnot is second-order, as they say in math and physics. If polling numbers are A>B, then I will predict that A beats B. The rest is things like determining what time window for calculating the median. I enjoy this activity…but it is not rocket science!

    • dsm

      E L

      I am a native Montanan (no longer live there) but much less sympathetic to Nate’s call on MT. Montana has a complicated political past and present, and outside the presidential trend I am not sure that “state fundamentals” have much of any meaning one way or another. To me, though, this is all the more reason to leave these extras out of the mix and defer to the polls. Trying to assign a “state fundamentals” value to a state like Montana reeks of the usual punditry noise to me. If the universe of polls is robust enough (a big “if” for a state like MT) then any “state fundamentals” or other extenuating factors ought to be built into the polls by election day.

      Of course a potential problem arises is when there are no reputable polls and/or only a few similarly biased polls. That to me is what is so impressive about Professor Wang’s North Dakota Senate call. I hope at some point Prof. Wang explains in detail the difference between his approach and 538’s approach regarding that particular race.

    • E L

      @ Sam Wang & dsm In Montana, only three polls were conducted on Tester/Rehberg in the last week by Mason Dixon, Rasmussen, PPP. The median was Tester +1 using the three polls in the last week. Only 5 polls total were conducted in the last 4 weeks. The median was Tester +1. Tester won by 3.9.

  • E L

    @ Sam Wang If you or Andrew could write a post every two weeks or so, I think the commenter community would stay together. We need you to jump start the conversation. Please.

  • Bret Lyon

    Hi Sam,

    Congratulations. I do have a question. You say you picked correctly on the senate, state by state, but I have never been able to find your predictions for the senate except as a group – Dems vs. Rep. Where did/do you post your predictions for senate races? How do I find them? I kept looking and it was disappointing not to find them.

  • K. W.

    There’s an article at about the victory of science in the election but the author means the victory of poll aggregation. People here looking for more discussion should jump in there. PEC commenters can add much to the conversation there.

  • Soothsayer123

    Just wanted to thank you for keeping me sane after the first debate. By the time Election Day rolled around, thanks to your research I was confident that Obama would win, even though everyone around me was predicting a Romney victory. They thought I was nuts. They don’t think so now!

  • Olav Grinde

    Dr Wang, fellow PEC readers, this may interest you. It’s a very different statistic, but it does have some relevance.

  • ThatSeattleGuy

    The New Yorker did an excellent – and prescient – feature on gerrymandering and its pernicious effects on democracy. Unfortunately, it was perhaps ten years ago, and I don’t have the reference handy. Someone with more Google-fu than I can probably scare it up. It’s definitely worth the read; what was said there is still – sadly – completely relevant to the current political scene.

  • Osso

    WE need multi-party system in the US.

  • Olav Grinde

    Someone may have answered this, but:
    • How many provisional votes were cast in 2012?
    • Is there a state-by-state list?
    • When will the final count be finished?

    • ThatSeattleGuy

      Good questions all. Don’t know the answer, but in my state (WA) the very slow, very underfunded, but thankfully very accurate vote counting is still going on – I think we’re still only at 84% tallied as of Friday night (*). And some of the bluest-voting counties in the state – like mine – still have 100K+ votes to go that aren’t in the totals yet.

      So the national popular vote margin – which as I write this is up to 2.6% in Obama’s favor – is going to still rise further as we go into next week. Hitting 3% wouldn’t be out of the question by the time every state certifies.

      (*) There are other reasons, especially WA state law which has a “postmark by election day” instead of a “received by election day” standard for ballots – meaning they’re still coming in the door of the county offices today!

    • Michael

      There are still more than 3 million votes to count in California. That’s mostly mail-in, but includes provisional as well. I believe that counties have until December 4 to report final tallies and the results will be certified by December 7.

      Have not been able to find comprehensive lists of either provisional ballots or total uncounted ballots. Have been checking a few state election board websites. One other thing to keep in mind is that most news outlets report % of precincts reporting which is not the same as % of votes counted. CNN has estimated % of votes counted on its results.

    • Michael

      Correction: CA counties have until 12/7 to report final tallies and the SOS will certify by 12/14.

      Latest report on unprocessed ballots:

    • Pat

      Surprisingly, for Washington, the CNN website has been mentioning “estimated 55% vote in” for many days…
      Even if it’s an all-mail voting state, it’s hard to understand why the vote count is so slow and have evolved much since Wednesday. Is CNN simply slow to update its count, or is Washington state not reporting regularly on its new counts?

    • securecare

      Go here to track Washington state vote counting progress, w/maps and numbers.

    • Michael C.

      Looks like CNN needs to update. There are about 360,000 votes still to count in Washington, or around 11%. Could be a few more votes out there, because the rule is postmarked by election day (not received by election day), but I doubt it’s very many.

    • securecare

      And why Washington state is so “slow”.

  • Ethan

    “This discrepancy between popular votes and seat counts is the largest since 1950.”

    Please explain this. I was wondering about this and so I’ve run some numbers (I’ve only had time to do since 1978), but it appears that the disproportionate share of House seats given the popular vote was regularly larger (to the Democrats’ advantage) between 1978 and 1992 than the the Republicans’ advantage is now. (Of course control of the House is a different matter.)

  • mediaglyphic

    @dr. wang, no-one else is talking about this as clearly as you. This analysis will reduce the legitimacy of the republican congress. We need to be as loud as possible about this. Thank you. I have told this story (pop vote vs. house seats) to a dozen people, most of them very aware, no-one knows this. WE NEED TO BE LOUD!!!

    • 538 Refugee

      It would be interesting to see Obama us it in the bully pulpit as he negotiates over averting the ‘fiscal cliff’. After all, he got more votes than all of them combined if I read things right.

    • Olav Grinde

      Yes, this needs to be the subject of many msm news stories. Oh, and Dr Wang, please call Bill Maher, Stephen Colbert and Jon Stewart. And Fox News… :)

  • David

    The current system or total popular vote.

    Or, divide each state’s EV up by percentage of vote candidate recieves with all fractions being rounded up in favor of the overall winner. Closer to popular vote but it still matters a bit to win a state.

    Popular vote is easiest.

    • Ken Dogson

      No on the dividing the EC votes by district.

      Gerrymandered districts would deliver a R to the WH. OH in 2012, for example, would have delivered more EV votes to Romney than Obama.

      Popular vote is the way to go. One citizen, one vote.

    • Craigo

      That’s not what he suggested, Ken. He’s suggesting that if a state has 10 electoral votes, and X beats Y by 55% to 45%, then X should receive 6 electoral votes.

      I’m working this morning or I’d do it myself, but I have to see how closely this method would track the popular vote to see whether I could endorse it.

    • Craigo

      I should make it clear that David’s method mitigates against the strongest (only?) argument against straight popular vote: The chaos of a nationwide recount after a close election. In this case, only states where there is a possibility of an electoral vote tipping would need to recount.

    • Craigo

      OK, this is a fun idea – what about when there are two overhang seats, and only two candidates have met the threshhold for electoral votes?

      For example, this system would give 21.7 electoral votes in Texas to Romney, and 15.7 to Obama. So Obama has earned 15, and Romney gets rounded up for being the overall winner. But that’s only 37 votes, while Texas has 38. Obama should get the extra one, no? 22-16?

    • Craigo

      Okay, the David method, in 2012, would have awarded 280 EVs to Obama, 258 to Romney.

      DC, Vermont, and Wyoming were the only places where a candidate would have swept the elecotral votes. Obama even got a vote in Utah.

    • Pat

      To solve the TX problem, I guess a solution would be to only award EV among the 2 top votes getters in the state. In that state, with Romney 57.2%, Obama 41.4%, we’d have Romney: 57.2/(57.2+41.4)*38 = 22.04 and Obama 41.4/(57.2+41.4)*38 = 15.95.
      From the system proposed, It would be Obama 15 – Romney 23.
      I like this idea. Would be interesting to calculate what it would give for the 2000 and 2004 elections!

  • Mark F.

    Even if you cut out the blatantly partisan redistricting (for which Dmocrats are just as guilty as Republicans), the fact remains that the Senate is deliberately undemocratic by design and nothing can change that. Senators representing 25% of the population can outvote Senators representing 75%. And that is not a bad thing as otherwise small population states could be steamrolled over by the high population ones.

    • Craigo

      “the Senate is deliberately undemocratic by design…And that is not a bad thing.”

      Yes, it is. Unfortunately, it will likely never change.

      Your argument essentially boils down to “small groups of people could be outvoted by large groups of poeple.” The horror!

  • Some Body

    On a tangent – what are the chances Puerto Rico is joining the union soon, and how would that affect the house?

    • Matt McIrvin

      My impression is that it’s extremely unclear how the Puerto Rico situation will shake out. The pro-and anti-statehood sides are both trying to spin the ambiguity about what the vote really means, and also nobody is sure how Congress will respond. It may be a while.

      At the current size of the House, Puerto Rico would get five seats (so seven electoral votes). Whether the House gets expanded or votes get taken away from other states is actually up to Congress, by statute. If the votes were reapportioned immediately, I believe California, Texas, Florida, Minnesota and Washington would each lose a seat.

      When Alaska and Hawaii were admitted, to avoid a snap reapportionment, the House was expanded temporarily until the next census, when it went back down to 435 members.

  • Michael

    Can you plot the historical results against a neutral line somehow? For example, we would know that in a tie vote (a 0 differential) we would ideally see a 0 seat differential. I am not sure how to calculate the line above and below there. It would be easy to do it as a percentage of the seats, but as you move away from 0 the seat impact should accelerate some, since a shift nationally is likely to be seen in all districts, and even neutrally drawn districts will tend to switch together if they are nearly balanced to start with.

    What I am going for here is to find out if recent redistricting is truly a historical shift in favor of Republicans, or as I am sure they would argue, is it a movement back to a neutral line.

    • Ethan

      This is a good point. I don’t really know what a neutral line would look like. If there is no acceleration further from 0, then it’s clear that Democrats have benefitted much more historically (I’ve looked since 1942) than Republicans from the mismatch between House seats and total national House votes (despite what I was hoping to find).

      I’d even say that you’d have to assume fairly significant acceleration away from zero before you could even begin to make the argument that what Republicans have in 2012 is unusual in the recent history of House voting (and again Democrats have almost always been the beneficiaries).

    • Michael C.

      First of all, there are two of us Michaels here, so I added an initial.

      Now, to the point — it is my understanding that prior to the passage of the voting rights act, redistricting basically only happened if there was reapportionment. Unless a state gained or lost seats as a result of the new census, the districts were likely to be left along. Of course, this is not to say that gerrymandering was invented in 1970. But I would think there’s been a lot more of it since then.

    • Matt McIrvin

      At any rate, it looks (by eyeball) as if the center of that cloud of points is a line that passes above the origin, which would indeed suggest that Democrats have benefited more from this historically, though not by as much as the Republicans did this year.

  • mediaglyphic

    Dr. Wang, is there an easy way to determine which states/districts have potentially illegetimate representatives and to publish this list. You could use your fame to bring light to this issue and ensure that these representatives don’t block legistlation that the majority of people support.

    • Craigo

      Nobody is in an “illegitimate” seat. They were all elected by the voters of their district, and to the extent possible while maintaining representation for the smaller states, each district has more or less the same number of inhabitants.

    • Michael C.

      “to the extent possible while maintaining representation for the smaller states, each district has more or less the same number of inhabitants.”

      This really isn’t as true as commonly assumed. Montana has 74% more people than Wyoming, but both have a single CD. California has 66 time the population of Wyoming, but only 53 times as many representatives. This anti-democratic constraint was imposed legislatively, not constitutionally.

    • Craigo

      That’s correct, though I was proceeding from an assumption of 435 seats.

    • mediaglyphic


      Godels incompletness theorem. Whats legitimate at one level is illegitimate at another.

      The makeup of the house is illegitimate and we need to SHOUT about it. The US collects less than 30% of GDP as tax revenue, much much less than other developed nations, as a result inequality is increasing massively here. Don;t drink the KOOL AID

  • E L

    The IUS Supreme Court has accept a case challenging the constitutionality of Section 5, the enforcement section, of the Voting Rights Act of 1964. If SCOTUS rules 5-4, as expected, to invalidate Section 5, a number of state with long histories of racial voter discrimination would be freed from federal supervision. The best source for information concerning this case is:

    Another case filed by the US Chamber of Commerce in Federal District Court would permit unlimited, secret, direct corporate contributions to candidates’ political campaigns. I will try to find the cite to this case on PACER.

    • Matt McIrvin

      Given that so much racially-tinged vote suppression happened this year in swing states that were not covered by Section 5, and given the gigantic blowback that resulted, I’m almost tempted to say “bring it on”. But that’s an easy, privileged thing for me to say since I wouldn’t be one of the people directly affected.

    • E L

      @ Matt The efforts this year at voter suppression were rather tame, but undoubtedly annoying. The easiest and most effective tool for voter suppression is a stiff poll tax. Poll taxes, along with literacy tests, were the favored weapons of the white supremacy, pre Voting Rights Act of 1964 South.

  • HowardA

    Using a Suffolk University poll of UNLIKELY voters I calculated what the vote would have been if ALL eligible voters voted. Obama would have won by a REAL landslide: 48% to 35.3% (note a large portion of those not voting say they would have support an unspecified Third Party candidate or were undecided) Obama’s margin in the popular vote would have increased from 3 million to at least 27 million. Raw data is here:

    • HowardA

      If the House breakdown followed this result, the Democrats would control the House 250 to 185. This further shows how unfair the Republican redestricting is. Following the same logic, the Democrats would control the Senate, 57 to 43, which is pretty close to the actual margin.

    • Paul K2

      Wow… This explains why the Obama campaign focused their (winning) strategy mostly on registering and GOTV campaigns. He wants to bring the real “silent majority” into the body politic.

  • Amitabh Lath

    Until recently I did not know who my state senator or assemblymen were in Trenton. I still don’t know who the freeholders are and what their role is. But the state level bodies are the ones that determine the districts. They took Cherry Hill out of NJ03, turning it from D to R.

    Gerrymandering is a problem, because Republicans control state houses at the time of the census results.

    They control state houses because Democrats do are famously not interested in local elections.

    Before we start agitating for federal level legislation let us admit our own culpability in this.
    I for one am going to get involved in county/state level politics (hurricane Sandy showed me where the real political power lies).

    • Paul K2

      Regarding the emphasis on local politics, this election shouts loudly “Yes!”. Spend money on local politics, because you get a better bang for the buck. Find people, and get them into the process. We need everyone.

  • Ivan White

    Maybe I am missing something here as I’ve been at ‘war’ with math since I was in the 5th grade but polls show something like a 9% approval rating of Congress. If this is this the case, then why was there as little change in both houses of Congress as there were?

    • Michael C.

      1) Gerrymandering.
      2) Pork.
      3) There has always been a substantial disconnect between approval ratings for Congress and approval ratings for individual Congresspersons.

  • terri

    Just. Keep. Voting. Democratic. It might take some time but eventually we’ll get rid of all the Republicans. Vote in every election, straight Democratic and don’t be shy about it. The Republicans are the fringe now, the extremists, and we Democrats are the normal, sane ones. Let’s build on that advantage.

  • Captain Ron

    I favor the Maine and Nebraska system for each state in the Presidential election but only if they go to the non partisan method in Iowa of redistricting.

  • Osso


    This just in!

    “ST. PETERSBURG, Fla. — President Barack Obama was declared the winner of Florida’s 29 electoral votes Saturday, ending a four-day count with a razor-thin margin that narrowly avoided an automatic recount that would have brought back memories of 2000.

    No matter the outcome, Obama had already clinched re-election and now has 332 electoral votes to Romney’s 206.

    The Florida Secretary of State’s Office said that with almost 100 percent of the vote counted, Obama led Republican challenger Mitt Romney 50 percent to 49.1 percent, a difference of about 74,000 votes. That was over the half-percent margin where a computer recount would have been automatically ordered unless Romney had waived it”.

  • JohnH

    Dr Wang, do you support proportional representation in the House of Representatives?

    • Pat

      Yes, I am wondering how it would work if a proportional system by state was established, like many countries.
      For example, in Colorado (7 representatives), people would vote for 7 candidates and the top 7 vote getters would be elected.

    • Olav Grinde

      Interesting idea! Not least because that would surely give third-party candidates a far better shot at getting elected.

      On the other hand, if the parties got most voters to vote along party lines, you might have Texas fielding a Republican-only delegation, while California might be Democrat-only. Moreover, you would stand to lose the close bond (to the extent one exists) between the candidate and his/her district.

    • Pat

      No no: assuming that the 2-party House vote split 61% Rep – 39% Dem in Texas, then the 36 seats would be apportioned proportionately, i.e. 22 Rep – 14 Dem.

    • Pat

      Here is a nice explanation, with different variants. Also shows where in the world such a system is used:

    • Olav Grinde

      @Pat, my post was in response to your sentence: “For example, in Colorado (7 representatives), people would vote for 7 candidates and the top 7 vote getters would be elected.”

      That’s not exactly proportional representation, but something else entirely.

  • Olav Grinde

    Here is a nice article on the BBC’s website about about Dr Wang and other successful predictors.

    • Amitabh Lath

      Drew Linzer and his Bayesian priors were sitting at 332 EV for months now. That’s either amazing, or a lucky guess (Obama wins all his 2008 states except IN, NC is a pretty good prior. You don’t need to be a rocket scientist).

      But I hope Linzer gets some kudos for his steadfastness.

    • Matt McIrvin

      Florida just declared today: Obama indeed got 332.

  • Ashbel green

    I am wondering how the result would have turned out if electoral votes were awarded by congressional district. So in Oregon, where I live, Obama won 4 out of 5 congressional districts and then would get the 2 other votes for winning the state popular vote. Wouldn’t this cause the candidates to spend more time in different states? Oregon has three competitive congressional districts. Or are most congressional districts both R and D so hopelessly gerrymandered so it would change candidate behavior. Maybe someone has already calculated this and I missed it. If so can someone give me a link? Thanks

    • Pat

      I read somewhere that in that case, Romney would have received more EV in states like Ohio and Florida even though he lost the popular vote. But I wonder if district-by-district vote counts are already available somewhere in a clean and usable way? I remember it was not easily accessible in 2008 and people ended up building a shared spreadsheet, with estimates often based on county-by-county data.

    • Ashbel Green

      I recall reading quite a while ago that Bush won a majority of congressional districts in 2000. I don’t recall whether the 2 additional EVs per state were part of the calculation.

    • Ashbel Green

      Okay, I think I’ve answered my own question. This web site is not perfect because it says Obama and McCain tied in 6 Congressional district (, but a rough calculation gives Obama 241 Congressional district, McCain 188 and 6 tied. Split them and Obama gets 244 to McCain’s 191. Obama adds 3 for DC to get to 247. He won 27 states, for another 54 EVs and a total of 301. He wins, but loses 64 EVs from his total on the winner-take-all basis (Give or take 1 or 2 because of Nebraska’s split). I wonder whether he would have actually won in 2012 under such a model.

  • mediaglyphic

    amazing to hear the talking heads on tv.

    1) no-one is talking about the popular votevs seats in the house
    2) everyone is acknowledging the importance of turnout and the obama campaigns superior big data operation, but no-one thinks that this operation could have an impact in the 2014 midterms. If i were obama, i would focus this operation on congressional seats democrats could win in 2014 and try and influence this

    • Michael C.

      I agree that Democrats should make a concerted and targeted effort to turn out their voters in the next mid-terms, but that’s going to be a tough row to hoe. Millions of people didn’t vote in their Congressional races even when they were already standing in the booth voting for Obama at the top of the ticket.

    • wheelers cat

      thats fine wid me.

      OFA never sleeps.
      expect us.

      Democracy 2014.
      Occupy the electorate.

    • mediaglyphic

      Michael, its a matter of focus, The president only recorded one robocall for a congressman this time around. I think turning the big data machine on congress will yield some results

      But the real power may be the fear of big data in the republican congress.

  • mediaglyphic

    … influence republican congressman to co-operate.

    Technology is power and the democrats have technology on our side. (it wasn’t always this way, remember william jennings bryan and the scopes monkey trial!!)

  • Foster Brown

    Thank you for the fine extensive accurate work of you and your team. The four or so meta-sites have been a great comfort throughout this long ordeal. See you, hopefully, in four years.

  • Andrew

    On Gerrymandering.

    Some readers have cited the example of Pennsylvania. However, the problem is really not so much gerrymandering as the classic donut hole vs. pie slice division of the electorate when one portion of the electorate is highly concentrated.

    In Philadelphia, the Democratic vote is very highly concentrated in Philadelphia and its immediately adjacent suburbs and in Pittsburgh and its immediately adjacent suburbs. Romney lost the state by 280,000 votes, but outside these areas he actually won by 500,000 votes. Obviously, almost all of the districts outside those two areas are going to lean Republican, and not surprisingly, Republicans won 13 of 14 districts in the rest of the state simply because the rest of the state voted 56-44 Republican.

    “Good” redistricting proponents typically cite the need to maintain individual political entities like major cities and counties in as few districts as possible insead of dividing them up. They have also stood for the correct principle that the Voting Rights Act requires the creation of minority majority districts in order to ensure political representation by minority groups.

    In Pennsylvania, that means Pittsburgh is kept in a single district, and Philadelphia is primarily divided into two districts. To make many more Democratic districts would mean slicing Pittsburgh into two districts and slicing Philadelphia into 4 or more districts. The Democrats tried this in 1990, and the result was Rick Santorum winning one of the Pittsburgh districts and the Democrats losing the Montgomery County district most of the decade when they probably should have been winning it. Even if it worked, this time around, it would only give the Democrats 2 or 3 more seats, still leaving them at an 11-7 or 10-8 disadvantage. Their major problem is not gerrymandering, its that in most of the state there simply aren’t any large concentrations of Democratic voters that would give them reliable majorities.

    Just like Republicans have an urban voter problem, Democrats have a small town/rural voter problem, and the small town/rural problem is a debilitating issue in a system using uniform districts, because a party that appeals in most places by 60-40 is going to dominate another party that appeals in limited large places 80-20. This is why Democrats also are doing so poorly in state legislatures, which have much smaller districts, but still see Republicans in a majority in a majority of potential districts.

    • mediaglyphic

      its like a parametric equalizer, the sound man can control which frequencies to supress and which to amplify. but the characteristics of the room (lets call it echo) is also important. We are the echo.

  • Robert de Neufville

    Republicans did win more seats than they would have been expected to on the basis of vote totals alone. There were four states—Michigan, North Carolina, Pennsylvania, and Wisconsin—where they won less than half the vote but won more than half the seats—36 to the Democrats’ 17. But there is a way the vote total is misleading: more Republicans ran unopposed, presumably because Democrats didn’t think they could challenge them. Essentially the Democrats started Election Day around 5 seats down, and even with less than half total vote, Republicans could be expected to win a majority in the House.

    • mediaglyphic

      wouldn’t this show up in the voting totals? If republicans ran unopposed, then they would win more of the vote.

    • Michael C.

      Some folks have been using AP national house vote totals that exclude races where a candidate ran unopposed.

  • Michael C.

    Virginia House Vote with 2575/2588 precincts reporting:

    R: 1,834,857 50.60%
    D: 1,736,014 47.88%
    O: 55,011 1.52%

    Republicans won 8 of 11 seats.

  • Richard Vance

    Lessons learned.
    1. Democrats MUST concentrate on building state house majorities by 2018.
    2. 1. Democrats MUST concentrate on building state house majorities by 2018.
    3. Democrats MUST concentrate on building state house majorities by 2018.

    • Olav Grinde

      Richard, I totally agree, but let me add a few thoughts.

      1.) Democrats need to maintain Obama’s impressive GOTV organization. This needs to be used in the 2014 Senate and House elections, future gubernatorial elections — and in state-level elections.

      2.) Democrats need to identify and increase the visibility of strong candidates that can challenge Republican incumbents. Except where absolutely necessary, one should avoid DINO candidates (Democrats in name only).

      3.) Democrats need to push for fas simpler voter registration. In my opinion, the model should be California’s online registration, supplemented Election Day voter registration.

      Currently eight states plus DC have the latter (Idaho, Iowa, Maine, Minnesota, Montana, New Hampshire, Wisconsin, Wyoming, and Washington DC.)

      California’s online voter registration? Take a look at the following: “More than a million Californians took advantage of this additional convenience to either register for the first time or update their current registration. Of those, nearly 60 percent were under age 35. And nearly half of all online registrants registered as Democrats, while only around a fifth registered as Republicans.”

      4.) Democrats need to develop and maintain a spine! This means staring down the Republicans, and not succumbing to the GOP demand that compromise means having to implement Republican policies!

      5.) Democrats need to keep their promises and pass legislation, including: i) tax increases for the 1%, ii) closing unreasonable tax loopholes for the very wealthy, iii) preserve tax cuts for the middle class, iv) compel corporations with “offshore headquarters” to pay a fair share, v) eliminate subsidies to oil companies and agribusiness, vi) thus reducing or eliminating the federal deficit, vii) pass a good immigration bill, viii) fine-tune the Affordable Health Care Act.

      If the Democratic Party does this, they will have secured election victories for many years to come!

    • Amitabh Lath

      Richard, I have been guilty of not paying attention to state and county level govt.

      Because of hurricane Sandy, I have become quite aware of local government. Olav, you make really good points about voter registration and legislation, and this has to include fighting for good local level laws (a zoning change for my section of NJ Route 27 affects us more than most national legislation).

      The next census is 2020, so we need to get the state houses back by 2018. We don’t have that much time.

  • Paul G

    Politico has a 2 page story today “How GOP freshmen won re-election” that does not even mention re-districting or gerrymandering.

    The journalistic malpractice is staggering. I mean, forget about the fact that aggressive partisan gerrymandering was crucial to the GOP success in the house. When comparing 2010 to 2012 how about at least mentioning that many Congressmen weren’t even running in the same districts this time???

  • 538 Refugee

    As we wind down and look to the next election:

    “Sylvia Manzano, an analyst at the polling firm Latino Decisions, said that Texas could be in play by as early as 2016 — if Democrats are willing to work for it and its ever-growing Latino population.”

    • Andrew

      Texas going Democratic is quite the fantasy.

      The PVI of Texas is R +9 to +10 range and has been near that since the 2000 election. That is UP from a PVI of R +3.5 in the 1980’s and 1990’s.

      By way of comparison, that is as Republican and Massachusetts and Maryland and New York are Democratic.

      Democrats would do well to focus on actually flipping Nevada (D+2), New Hampshire (D+1), Colorado (D+1) and Virginia (even) for good like they previously flipped New Jersey and New Mexico, and making sure they don’t accidentally let Pennsylvania (D+1 and down from D+4 in the 80’s) or Minnesota (D+2 and down from D+8 in the 80’s) slip away like they have West Virginia, Arkansas, and Missouri since 2000.

      Most states partisan make-up has been surprisingly stable since 1964 as to which side of the aisle they are on, although the amount of the lean has changed. As Virginia still leans slightly Republican, the Democrats only have a modest 272-266 advantage in the electoral college right now.

    • 538 Refugee

      I haven’t read anything as optimistic as Texas going in the next cycle but I have read other indications that look favorable for eventual Democratic leaning IF the parties stay unchanged. At this point I think the midterms can see the Republicans in serious disarray. War between the tea party purists and the moderates. I can see quite a few 3rd party candidates if the tea party doesn’t get their way and a lot of shift from Republican to Democrat if they do. A quick google turns up these articles. Eligible Latino voter registration lags big time in Texas. Some feel effective registration efforts can turn the tide sooner rather than later. I’m skeptical too that the numbers can turn around that quick but who knows? These articles have some numbers attached along with Democratic day dreaming.

Leave a Comment