Princeton Election Consortium

A first draft of electoral history. Since 2004

There is a great deal of activity in the field

August 25th, 2015, 6:39am by Sam Wang


I’m on “vacation,” working on projects that sometimes remind me of this quote from Richard Feynman. Replace the green block with an area that comes to mind.

Tags: Uncategorized

43 Comments so far ↓

  • bks

    This is not a parody:Our new poll finds that Trump is benefiting from a GOP electorate that thinks Barack Obama is a Muslim and was born in another country, and that immigrant children should be deported. 66% of Trump’s supporters believe that Obama is a Muslim to just 12% that grant he’s a Christian. 61% think Obama was not born in the United States to only 21% who accept that he was. And 63% want to amend the Constitution to eliminate birthright citizenship, to only 20% who want to keep things the way they are.

    Trump’s beliefs represent the consensus among the GOP electorate. 51% overall want to eliminate birthright citizenship. 54% think President Obama is a Muslim. And only 29% grant that President Obama was born in the United States. That’s less than the 40% who think Canadian born Ted Cruz was born in the United States.
    http://www.publicpolicypolling.com/main/2015/08/trump-supporters-think-obama-is-a-muslim-born-in-another-country.html
    –bks

    • pechmerle

      After Henry Kissinger was awarded a share of the Nobel Peace Prize in 1973, song writer Tom Lehrer announced: “Satire is no longer possible.” And indeed he never wrote another song.

  • 538 Refugee

    Carson has pulled even with Trump in Iowa. Sanders gaining on Clinton. This is getting interesting at least. I wonder if knowing they are under national scrutiny influences the answers the peopel in Iowa give to polls? I read that ‘back in the day’ Nielson ‘family members’ would be less than honest about what they watched. They would give what the thought were the ‘expected’ answers. My understaniding is that in the causcus your choice is public. So anger and anti-establishement sentiments aside, your decision will be public.

  • bks

    At what point do we say that the Trump polling is not like the Cain/Gingrich/Perry/Bachmann polling of 2011? –bks

  • 538 Refugee

    “The Des Moines Register-Bloomberg Politics poll, considered the gold standard of Iowa surveys, found Clinton with the support of 37 percent of likely Democratic caucus-goers, followed by Sanders at 30 percent.”

    “Sanders’s favorability rating among likely Democratic caucus-goers jumped from 57 percent to 73 percent over the past two months, the poll found. The survey showed him attracting a coalition similar to Obama’s in 2008: young people, liberals and first-time caucus-goers.”

    http://www.washingtonpost.com/news/post-politics/wp/2015/08/29/iowa-poll-sanders-surges-to-within-7-points-of-clinton/

    I wonder if I can get shares in popcorn? This is getting interesting to watch.

    • Amitabh Lath

      And on the other side:

      By capturing 23 percent of the vote while 61 percent have favorable feelings toward him, “Trump definitely has upside potential,” Selzer said. “All the talk about a ceiling was a little premature.”

      I couldn’t have put that better, Ms. Selzer.

  • pechmerle

    I would like to see a draft of the law review article.
    Once upon a time I was on the board of editors of a law review.

  • Sam Wang

    Obliquely relevant to this post, I am working on a law review article proposing a standard for partisan gerrymandering, for use by the Supreme Court. It started from my NYTimes piece and other items I’ve written here. Anyone interested in seeing a draft?

    • Amitabh Lath

      Sure, love to read it.
      But will they publish an article from a neurobiologist?

    • bks

      Is there any other kind of gerrymandering?

    • Sam Wang

      There is single-district gerrymandering to protect a single incumbent or to isolate one community.

    • Amitabh Lath

      Impressive, I’ve never had a single author paper published in a peer reviewed journal. In my field you are doing well if you can keep it to a few kiloauthors.

      What state are you going to use for an example of extreme gerrymandering? We could dip into graph theory and come up with a metric for convolutedness of district boundaries.

    • 538 Refugee

      On the surface it would seem easy enough. If 50% vote for a representitive from Party A and 50% for a representitive from Party B, then the number of elected officials should reflect that.

    • Sam Wang

      Not exactly. First off, there’s a bit of uncertainty and variability, so that pairing a 49.5% popular-vote share with a 50.5% seat share would not be a miscarriage of democracy. Secondly, Democrats are packed a bit into cities, putting them at a slight disadvantage (though as it turns out, the effect of post-2010 gerrymandering is larger than the packed-Democrat effect).

      Secondly, your criterion doesn’t help for popular-vote shares that are any fraction other than 50%. Turns out the “fair” number of seats as a function of votes is an S-shaped curve. It basically looks a bit like the integral of a bell-shaped curve, because districts are distributed a little bit like a bell-shaped curve. Or they used to be…packing into cities and partisan gerrymandering have changed that. Anyway, the point is that if districts are normally distributed (i.e. to a bell-shaped curve), then 51% of the popular vote gives more than 51% of the seats.

    • Amitabh Lath

      It’s hard to prove a gerrymander just because the ratio of representatives does not match the ratio of votes. The gerrymanderers could argue it’s not their fault that all the liberals want to cluster in urban areas and waste their votes.

    • 538 Refugee

      Well, a ‘representitive government’ should be representitive and it isn’t like there aren’t enough historical records to guide you in the matter.

    • Amitabh Lath

      Historical housing practices and redlining have created areas such as the south side of Chicago that concentrate democratic votes.

      The nearby area consisting of Lake, McHenry, DuPage counties are purple-reddish. If african-americans had not been prevented from moving into these suburban counties by the extremely racist housing and lending practices, then definitely IL would have more democratic representation.

      Not gerrymandering in the strict sense of the word, but the effect is the same.

    • 538 Refugee

      That is why I prefaced my initial reply with ‘on the surface”. It’s always easy for the person not doing the work to see it as a simple fix. Like the 90/10 rule in clubs and organizations. 10% do all the work and the other 90% sit around griping that the 10% is ignoring their input.

  • 538 Refugee

    Speaking of statistical models:

    Clarkson, chief statistician for the university’s National Institute for Aviation Research, filed the open records lawsuit as part of her personal quest to find the answer to an unexplained pattern that transcends elections and states.

    She wants the hard-copies to check the error rate on electronic voting machines that were used in a voting station in Sedgwick County to establish a statistical model.

    Clarkson, a certified quality engineer with a Ph.D. in statistics, has analyzed election returns in Kansas and elsewhere over several elections that indicate “a statistically significant” pattern where the percentage of Republican votes increase the larger the size of the precinct. The pattern could be voter fraud or a demographic trend that has not been picked up by extensive polling.

    http://ksn.com/2015/08/24/kansas-seeks-to-block-release-of-voting-machine-paper-tapes/

    Obviously this ‘open’ stuff has gone too far when you question your state officials it seems. No mention of Diebold.

  • Amitabh Lath

    Sean Trende has a widget!

    http://www.realclearpolitics.com/articles/2015/08/26/demographics_and_the_2016_election_scenarios.html

    You may remember Trende as the author of that 2013 article that convinced many Republicans that they didn’t have to chase hispanic votes, all they had to do was boost the (xenophobic?) white vote.

    Anyway, with this widget you can dial in the percentage of various ethnic voting populations, and their R/D ratios. It spits out popular vote and electoral vote.

    Trende-code is not public, but it’s neat to fiddle with the knobs and see from the EV jump where CA turns red. The GOP dream scenario. Let the skewing begin!

    • bks

      Trende also has a theory that Bush is a bigger problem for the GOP than is Trump, because Bush captures all the corporate money without building a base:
      http://hotair.com/archives/2015/08/27/blast-off-shaky-jeb-bush-campaign-ready-to-soar-thanks-to-coveted-eric-cantor-endorsement/
      –bks

    • Sam Wang

      It is certainly true that among “major” GOP contenders, Jeb Bush has unusually high unfavorable ratings among Republicans, as well as with the general electorate. If that’s what “without building a base” means, then I think I agree.

      Republicans would be better off with Rubio or Carson, who have the largest favorable-unfavorable spread. By this measure, Walker also does OK…but he is flaming out.

    • Amitabh Lath

      Trende is one of the major enablers of Republican magical thinking. Coming out of 2012 they had convinced themselves that immigration reform needed to be passed, there was movement from the Senate (Rubio esp) but that Trende analysis of the white voter who sat at home seemed to puncture that effort.

      If an immigration bill had passed, would Trump be an item?

    • Matt McIrvin

      I just got alarmed by averages on the Huffington Post site showing both Trump and Ben Carson coming very close to beating Hillary Clinton in head-to-head polling… but when I looked more closely, both numbers relied heavily on one literally unbelievable outlier poll from Gravis Marketing (because the overall polling of these matchups is fairly sparse).

      Expect a lot more of that…

    • Matt McIrvin

      Of course, Breitbart is touting the Gravis poll to the heavens.

  • bks

    Returning to the conference topic: In Computer Science getting a paper in the conference proceedings is approximately equivalent to getting a paper in an archival journal in Biology. That makes for CS conferences that are replete with redundancy and maximally soporific. Biology conferences are much better as presentations can be a bit more adventurous. Plus biologists are more visual so the slides are, on the whole, more interesting. –bks

    • Amitabh Lath

      Focus groups? Really? Luntz picks out a couple of dozen avowed Trump supporters, puts them in a room, and then is shocked, shocked to find that they support Trump.

      Tell me again what I am supposed to learn from this exercise?

  • bks

    You should only attend conferences for two reasons:
    1. You have an important result to report
    2. It’s in a great location and you can skip out

    Meanwhile, interesting result in New Hampshire,

    Who would you prefer as the Republican candidate if you had to choose between just Jeb Bush and Donald Trump?

    Bush 33%, Trump 56%
    http://www.publicpolicypolling.com/pdf/2015/PPP_Release_NH_82515.pdf
    –bks

    • 538 Refugee

      3. You are paid to do so.

      Strange topic. Shirley you’re joking Dr. Wang?

    • Sam Wang

      The conference advice is excellent for established researchers.

      The survey result suggests that it is time for the GOP to exercise some strategery. In that same survey, Trump is only 47/39 over Ben Carson. Carson 2016!?

  • Amitabh Lath

    Sam, you must have been at Caltech as an undergrad when he was there. Was he still teaching?

    • Sam Wang

      I was indeed at Caltech from 1982 to 1986, and Feynman was still teaching, on his own terms. On Tuesdays at 5pm he gave “Phys X,” in which he entertained any questions from freshmen and sophomores. I recall once he brought in a 1/64″ wide motor and we got into the question of what such a motor would be good for. It was the precursor to nanotechnology.

      He was also a presence in student houses and in student life. In my house, one “alley” challenged another to a game of Feynman Says, which is exactly what it sounds like, Simon Says but with RF presiding. He said “Feynman says put your hands on your shoulders.” “What is the integral from zero to infinity of e to the minus x squared.” “uh uh, didn’t say Feynman says!” He also appeared in the school musical every spring. He led a parade in South Pacific, shirtless and in a grass skirt.

      Once, while studying late at night in Lauritsen, we passed one another in the stairwell. He emitted the loudest borborgymus i have ever heard, before or since. He grinned at me.

    • Amitabh Lath

      You are a lucky man to have personal memories of him. My advisors had stories about how in the late 60′s he would drive up to SLAC and go up into the control room and look over the data as it was coming in and basically make a genial pest of himself. He coined the word “partons” to describe what they were seeing. He figured out before the experimenters themselves that there was substructure to the proton. Amazing intellect.

  • Matt McIrvin

    I was just trying to remember what the word was originally, but I think I do now.

    • Sam Wang

      Rhymes with “depravity.”

    • Matt McIrvin

      Yeah, and the thing that reminded me is that it’s reminiscent of some of Sabine Hossenfelder’s grumpier complaints about the state of the field today… though it’s not nearly this bad.