Open Thread, 11-01-2012

By Razib Khan | November 1, 2012 9:51 pm

Inform us!

CATEGORIZED UNDER: Uncategorized
MORE ABOUT: Open Thread
  • Mamikon

    I have been following this blog for a few months and still haven’t put my finger on one question that bugs in regards to Out of Africa.
    Are all hominids considered to be originally from Africa and not just modern humans, they just dispersed from there in multiple waves? Am I correct to assume that Neanderthals, Denisovans and perhaps other archaic hominids left Africa, populated the rest of the world and then later humans followed their course?

  • toto

    Apparently it’s Nate Silver hunting season on the internet – the trend has even crossed the Atlantic!

  • Chris_T_T

    A lot of the Nate Silver criticism is down right bizarre. How does developing a statistical model to aggregate as much data as he thinks is relevant to predict the winner and margin constitute attempting ‘to poll [his] way to victory’? It’s like someone developing a model to try to predict the likelihood of an NFL team to win each of its games and someone else claiming they’re trying to win their team the Superbowl rather than playing it out on the field!

  • https://plus.google.com/109962494182694679780/posts Razib Khan

    #3, i think they think that silver can impact turnout by asserting that the conclusion is forgone. it’s like he’s a hedge-fund analyst who pumps the stock ahead of shorting it, impacting a market in which he is a player. i think that’s a bullshit argument.

    btw, i get surprising few takers on pro-romeny polls-are-skewed bets. i’m going to finalize some i made this weekend in a post.

  • Karl Zimmerman

    Chris T -

    I know I’ve read the argument that the last 5% or so of voters to decide are literally waiting until they know who the winner is going to be, so they could jump on the winning bandwagon. I’m unaware if there’s any statistical proof of this however. Clearly, the Romney campaign believes this line however, because they tried for a few weeks to get the press to report their momentum was continuing, and not stalled out and possibly reversing slightly. I think much of the anger on the right is because Silver and other “math guys” are pulling back the curtain, and showing that at this point Obama looks to be an 80% favorite, although the margin remains close enough that something unexpected could still change the momentum to a narrow Romney win, when the press is supposed to continue to call it a tossup right up until election day.

    I think the anger from the “nonpartisan” punditocracy is different however. The typical way being a mainstream pundit works in the U.S. is you put forward hypotheses based upon little to no data, you make your analysis as evenhanded and vague as possible, and you receive no blowback for actually being wrong. I saw a very funny version of this yesterday, when Chris Cillizza and Aaron Blake moved Ohio from Lean D to Tossup in part because of “the absolute necessity for Romney to win the state if he wants to be president.” Data-driven nerds are a threat to political pundits because they show how shallow their understanding of the facts really is.

  • https://plus.google.com/109962494182694679780/posts Razib Khan

    #5, there isn’t evidence for the last second thing that’s robust from what i know. though there is some suggestions from some races. there is LOTS of evidence that a minority lie about who they voted for (or claim they’ve voted) after the fact, with a winner bias.

  • Chris_T_T

    Data-driven nerds are a threat to political pundits because they show how shallow their understanding of the facts really is.

    Heh, they can rest easy. Anyone who actually pays attention to something like statistical analysis already knows that pundits only have their jobs for how well they can craft a story. Sports journalists have been spinning stories out of random numbers for years without a problem even though statisticians have done far better analyses (and shown a lot of what people focus on is meaningless).

  • http://opines.mythusmage.org Alan Kellogg

    I’m writing with a query, has anybody any leads on a company to do a little genome scanning? I’d like to have a genome done (not mine) and I’m looking for people to do it reliably. By “genome” I mean a full genome, and not selected bits.

    It involves a primate, an unknown species, and the company would be working with cells from hair follicles. (The living ones, before they die and are incorporated into the hair shaft.) If you can help drop me a line at “mythusmage (at) gmail (dot) com”

  • Mike Steinberg
  • RafeK

    I am curious what some of the more numerically skilled in here think of this analysis of the election http://baseballcrank.com/archives2/2012/10/politics_why_i_2.php I tend to see this as a toss up but thats most a gestalt impression fueled by a very shallow set of priors.

  • atyl

    “I know I’ve read the argument that the last 5% or so of voters to decide are literally waiting until they know who the winner is going to be, so they could jump on the winning bandwagon. I’m unaware if there’s any statistical proof of this however. Clearly, the Romney campaign believes this line however, because they tried for a few weeks to get the press to report their momentum was continuing, and not stalled out and possibly reversing slightly.”

    The theory is, the undecideds always break for the challenger when the race with the incumbent is close. And in every election the challenger narrows the gap:

    http://www.weeklystandard.com/blogs/morning-jay-historically-speaking-obama-isnt-strong-shape_652813.html
    http://www.weeklystandard.com/blogs/morning-jay-race-has-just-begun_653404.html
    (Scroll down to the charts).

  • https://plus.google.com/109962494182694679780/posts Razib Khan

    atyl, do you want to bet on the election that romney loses? (if you are quoting jay cost i assume you are gullible and i would like to take your money!)

  • Icepick

    I just don’t see how statistical models can run that counter to intuition and past experience. Presidents that win a second term in office have almost always gotten bigger electoral wins the second time around. The only exception on absolute basis was Woodrow Wilson in 1916. He did much worse the second time around. That was at least partly because the Republican vote was so badly fragmented in 1912. That is, Wilson did better in 1912 because the opposition was split and with no split in 1916 he couldn’t possible hit his previous mark .

    In relative terms James Madison also did worse when he won a second term, though he got more electoral votes. (More electors had been added in the meantime because of population growth and the addition of one more state.) But Madison had the distinct disadvantage of running for re-election shortly after a war had started.

    Obama got 365 electoral votes last time. Anyone think he can even get close to that this time? If not, what is so special about these circumstances that he can do what many others have failed to do? He didn’t face divided opposition in 2008, and he’s not dealing with some monumental new crisis such as an invasion by our former imperial masters.

    Ultimately Presidents win second terms do so because people think things are going well, or at least progressing reasonably. (I’m thinking of 1864, 1936, 1940 and 1944 for “progressing reasonably”.) Who thinks things are going well?

  • https://plus.google.com/109962494182694679780/posts Razib Khan

    I just don’t see how statistical models can run that counter to intuition and past experience.

    but your sample size is small. the statistical models done well actually highlight how uncertain we should be. i think this becomes more clear when ‘economic determinism’ became the vogue ~2010 in regards to this election. it sounded plausible, until people pointed out that it was actually quite a tenuous proposition (e.g., economic variables predicted a 10 point gore win).

  • Icepick

    (e.g., economic variables predicted a 10 point gore win).

    There’s a large difference between 2000 and 2012: an incumbent President is running on his own record whereas Gore was running on the record of someone else.

    Basically I’m calling for judgement. One can say the sample size of chess games between Valery Salov and Kasparov is too small to make a statistically valid judgement of who was the better player. Or one can look at their games and determine that Kasparov was the superior player. Presidential elections aren’t that clear cut, of course, but I don’t see why judgement should be dropped from the discussion.

    the statistical models done well actually highlight how uncertain we should be.

    Yes, but is Silver’s model that really good or has he gotten lucky? (Zogby was once the flavor of the day for getting something right. Not so much any more.) How much track record does he have? And is there any reason I should favor his model over the Berry-Bickers model? Silver largely seems to get all the attention because the NYT promotes him and because a lot of people seem to like that Obama is a massive favorite to win in Silver’s predictions. Is that a statistically valid reason to give his model all the benefit of the doubt over any and all others?

  • https://plus.google.com/109962494182694679780/posts Razib Khan

    #15, like a good bayesian nate gives you a distribution of outcomes. that’s why you should prefer IMO. the older pollsters gave you a expected val + margin of error only.

    i would tend to go with the idea though that his model doesn’t add that much value over averaging the polls.

    also: Presidential elections aren’t that clear cut, of course, but I don’t see why judgement should be dropped from the discussion.

    if you read nate’s book he argues that judgement + statistics performs best. so you’re kind of making a caricature of him…though that’s not your fault. some of his acolytes do seem to not ‘get’ that his model is very provisional.

  • http://www.ahnenkult.com Ortu Kan

    Ainu SNP data available soon.

  • http://shinbounomatsuri.wordpress.com Spike Gomes

    Not overly surprised by the link that shows genetic kinship between Ryukyuan and Ainu. They both phenotypically resemble each other more than they do Naichi Japanese.

NEW ON DISCOVER
OPEN
CITIZEN SCIENCE
ADVERTISEMENT

Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Gene Expression

This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!

About Razib Khan

I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at http://www.razib.com

ADVERTISEMENT

See More

ADVERTISEMENT

RSS Razib’s Pinboard

Edifying books

Collapse bottom bar
+

Login to your Account

X
E-mail address:
Password:
Remember me
Forgot your password?
No problem. Click here to have it e-mailed to you.

Not Registered Yet?

Register now for FREE. Registration only takes a few minutes to complete. Register now »