The Other McCain

"One should either write ruthlessly what one believes to be the truth, or else shut up." — Arthur Koestler

Nate Silver Asks: Whose Shark Is This, and Why Do I Feel a Need to Jump It?

Posted on | October 19, 2012 | 35 Comments

It seems so long ago, but it was only late September, when Jonathan Chait dismissed as “poll denialists” those who expressed skepticism about the validity of polls showing Barack Obama sailing to an easy victory in Ohio and elsewhere. We were, said Chait, “totally crazy” for questioning the samples of these polls.

Three weeks later, Romney having zoomed upward in the polls, now it is the turn of liberals to complain, and we find that the previously respectable Nate Silver of the New York Times is now a poll denialist, complaining that the Gallup national tracking poll that showed Romney leading by 7 points is “deeply inconsistent with the results that other polling firms are showing in the presidential race, and the Gallup poll has a history of performing very poorly when that is the case.”

Whatever.

We’re now 18 days away from Election Day, and we’ll know soon enough whether Gallup or other polls are accurate. Is it likely that Romney will win the popular vote by a 7-point margin? I dunno. Probably not. This would be a landslide of historical proportions.

On the other hand, without going into Nate’s entire argument, I’ll say this: Polls are a lagging indicator. By the time a decisive shift begins to appear in the polls, the factors that caused the shift are usually several days or even weeks in the past. Instapundit wishes to make clear that he did not coin the term “preference cascade,” but if this is what we’ve been watching for the past three weeks . . .

Well, it is what it is. Donald Douglas says “wonder boy Nate Silver’s practically blown a gasket over the latest Gallup tracking numbers,” but the real question is, “Was Nate every really a wonder boy?” Or is it possible that, like Obama, he was overrated all along?

I’d make more fun of Nate’s newfound skepticism, except that for all I know, he’s right. This is my luxury: I’m a mere journalist, with a blog, and nobody’s paying me to be the Omniscient Expert, so I don’t have to pretend to know everything. Lucky me.

Bookmark and Share

Comments

  • Anamika

    On the other hand, without going into Nate’s entire argument, I’ll say this:

    But of course if you do go…into “Nate’s entire argument” rather cherry pick a sentence or two from his excellent blogpost, you would know that he makes perfect sense.

  • yetanotherjohn

    There is an underlying assumption in all the polls about how many dems and how many GOP are going to show up. If you normalize all the polls to one assumption (e.g. 2008, 2004 or 2010 turnout model) then they suddenly become much more in line. Is Gallup showing a bigger gap than the other polls? Absolutely. Is Gallup using a larger sample size than the other polls? Yes. Was Gallup closer to getting the 2010 GOP wave right than Nate? You bet. Nate is very good at damping down over blown enthusiasm on the left when the left is doing well. He isn’t as good at recognizing his own bias affecting his work when the right is on top.

  • RichmondG30

    As Obama plummets in the polls, Nate has his odds of him winning INCREASING. Currently up to 70.4%. What a pathetic little guy he is. Many tears will be shed on November 7.

  • http://2012.ak4mc.us/ McGehee

    Especially by Anamika.

  • http://twitter.com/yugan yugun

    ‘I’m a mere journalist’ ROFL. that was real funny. haven’t laughed this hard this early in the morning in quite a while. thanks.

  • Adobe_Walls

    Given that only 9% of those contacted will talk to the pollsters while 50 plus percent tell them to f**koff, I’d say all polls are suspect even if admitted to be “just a snapshot in time”. I’d say even the trends are merely informative not definitive. It will be interesting to see if even Rasmussen can maintain his reputation for accuracy when the votes are counted.

  • JeffS

    This from a known expert on trolling, goal post moving, and cherry picking. Tut tut!

  • JeffS

    I’m looking forward to the ‘sploding heads, myself. My stockpile of popcorn, butter, and salt is quite impressive.

    And ammo. Just in case.

  • http://twitter.com/ilovegrover Thane_Eichenauer

    FYFHO indeed.

  • elaine

    If you read his whole argument, he doesn’t even agree with the guidelines he sets out for reading all the polls: who’s ahead and what are the trends for each candidate? If he followed his own guidelines, he’d have Obama’s chances decreasing, not increasing.
    So, no, his analysis doesn’t make “perfect sense,” because he throws out his guidelines in order to flog an unsupported premise — that Obama is still winning and trending higher.

  • muffler

    It’s Math based on facts…. kind of like a science. Sorry if the facts are not supporting the bunk going around these days.

  • http://pulse.yahoo.com/_MJGP4QXZ5PRW2MFA5E25CV2WNU rosalie

    Anamika, You surprise me. I thought for sure that you WOULD go “into Nate’s entire argument”. I’ve learned that a lot of things that make sense to Liberals don’t make sense to anyone else. You are indeed a very special group.

  • DonaldDouglas
  • news4you

    No comparison between the right’s denialism of inconvenient facts and Silver analyzing the state of the race.

    A few weeks ago the right was obsessed with saying the polls were *all* wrong, It was nothing more sophisticated than a child jamming their fingers in their ears and screaming “LA LA LA LA LA” when told something they didn’t like.

    On the other hand Silver is looking at the state of all the polls and recognizing that Gallup seems to be an outlier. Furthermore he’s gone back to analyze historic data to show that Gallup has had some significant misses in recent years. That’s, you know, his job.

    And for the person complaining that Obama is improving recently in the 538 forecast- there’s a reason for that, it’s because after a steep fall following the abysmal first debate a number of indicators have started to recover a bit for Obama. His chance of winning has improved compared to the immediate post debate period but is still way down from the pre-debate period. You can easily see this in all the graphs on the right hand side of the 538 site.

    You just have to open your eyes to see it.

  • Wombat_socho

    Yeah, we’re all just imagining the rampant oversampling of Democrats that characterized all the polls we criticized. Get hired!

  • Wombat_socho

    Polling is not a science. Statistics is. You should learn the difference before exposing your idiocy in public.

  • muffler

    My pleasure if only to get you upset!

  • paulzummo

    “A few weeks ago the right was obsessed with saying the polls were *all* wrong, It was nothing more sophisticated than a child jamming their fingers in their ears and screaming “LA LA LA LA LA” when told something they didn’t like.”

    Wow, facts are clearly not your thing. It wasn’t as though Republicans merely denied what the polls said. We saw the internals of these various polls and laughed at their obviously ridiculous samples. When poll after poll has D-advantages that outpace what they even did in 2008, and when even these polls had Romney winning handily among Independents yet somehow still losing, then our skepticism was warranted. It’s Silver who is basically just shaking his head and going “nuh uh.”

  • http://wizbangblog.com/ Adjoran

    The hilarious thing is that when we pointed out the samples in many polls were running a higher % of Democrats than voted in their wave years of ’06 & ’08, we were told the polls just report what they find and we have to take them as they come.

    Now when the results aren’t to the left’s liking, they find reasons to criticize the polls.

    You have to be a complete, 100% lying hypocrite to be a leftist or Obama supporter.

  • http://profiles.google.com/rob5136 Rob Crawford

    How much does Biden pay you to comment?

  • http://profiles.google.com/rob5136 Rob Crawford

    “Special” in the “rides the short bus” sense.

  • Patrick DeBurgh

    Actually polling is a branch of statistics known as “sampling”. Go to wikipedia and look up “sampling”, integrate the information and perhaps your idiocy won’t be so exposed in public.

  • Wombat_socho

    I thought that sounded familiar. Hail, fellow F. Paul Wilson fan!

  • Wombat_socho

    I’ve taken statistics, thanks, I don’t need to rely on Wikipedia.

  • Wombat_socho

    Not enough to keep him from getting banned. So long, college boy, and I hope the Ducks get roasted this weekend.

  • JeffS

    Wombat is correct. Polling is not a science.

    Polling (or “sampling”, if you prefer) is an application of science to the real world by human beings, and hence subject to errors and manipulation.

    Now, go look up “integrate” and “information”, and perhaps your idiocy won’t be so exposed in public.

  • Pingback: Romney on the Right [IMAGE] « Lower The Boom

  • http://wizbangblog.com/ Adjoran

    Nah, not at all. Most of those kids are nice people. Today’s liberals rarely are.

  • http://wizbangblog.com/ Adjoran

    There is no science here, it is closer to pseudoscience, like soothsaying or reading the future from goat entrails. It is not statistically possible to “normalize” any results across a broad but exclusive array of studies with varying methodology and samples.

    And anyone remotely familiar with the scientific aspect of polling will tell you straight off they cannot predict the future. Yet Nate’s whole schtick is pretending he can.

    But since every race is between different candidates in different situations and with different voting populations, basing a prediction on prior years is a simply fallacy. Can’t be done, means NOTHING. It’s like the investment prospectuses are required to warn: past performance is no guarantee of future results.

  • http://wizbangblog.com/ Adjoran

    The response rate in the mid-80s when there were very few polling firms was nearly 75%. According to Pew Research, the average rate had declined to 37% by 1997, and they are also the source for the 9% figure. Their numbers come from the polling outfits themselves, who would never release their individual response rates because it is “proprietary information” (= “if our competitors find out we are doing worse than them, it will hurt business so we won’t risk it”).

    It is theoretically possible to still get an equally valid sample with that sort of low response rate, but in the real world it is far more expensive (more total calls have to be made to construct the same size sample) and there is no way to ascertain whether or not the refusers have different characteristics and opinions from the responders.

  • http://wizbangblog.com/ Adjoran

    Well, if the composition of the electorate in 2 1/2 weeks is D+7 or greater, then you and those polls will have been justified. Come back and gloat.

    Or if not, come back for sympathy. We’re a caring bunch.

  • Pingback: Larwyn’s Linx: Has the preference cascade begun? | Preppers Universe

  • Kevin Tierney

    Silver may or may not be right. But there’s something called “statistical noise.” He says “don’t trust gallup when Gallup is more out in front than the other polls.”
    Where is the magic line where it becomes “too out in front.” Which polls must it be in accord with? How many times did this situation occur where you can draw meaningful analysis, otherwise, it is statistical noise, and would be a great case in his new book on why people’s predictions are wrong.
    His entire argument does indeed boil down to a big fat “nuh-uh!” In order to make his argument, he relies on polls with relatively unproven methodology that has limitations even he acknowledges (RAND) or ones that nobody outside of Dem underground takes seriously (yougov online or google consumers) and he was singing a far different tune when Gallup was showing a pretty strong Obama lead. It is akin to some conservatives who were citing “unskewed polls” during August and September. They could be right, but not based on the sevidence they present!

  • Pingback: The Democrat Graveyard Whistling Choir Ignores ‘Preference Cascade’ Scenario : The Other McCain

  • Pingback: Signs and Omens: Obama’s Fading Hope and the Graveyard Whistling Choir : The Other McCain