Huge win for the stat geeks

Let’s not forget one of the other big winners of the night in the geek vs. pundit showdown: Nate Silver. I’m sure it was tough for some of his followers to maintain belief that the presidential race wouldn’t be close, Electoral College-wise, when so many media pundits were breathlessly telling us that the race was tied. I was trying to calm family and friends with the forecasts, but people were having trouble accepting evidence over emotion even when the evidence was good news.

It also turned out that our Senate race was not too close to call. Even the most optimistic of projections seem to have undercounted Warren’s support and overestimated Brown’s. However, Silver did peg this race as safe Democratic when most of the media were telling us it was essentially tied.

It was good that many of us refused to believe that the seat was “safe” — not after what happened two years ago, even thought the situations were so very different — and worked as if the race was tied. But it wasn’t close to being tied, even if a couple of outlier polls claimed Brown had a lead due to likely voter models that were obviously off.

So thanks to Silver, the Princeton Election Consortium and others who remind us even while the media does not, that the appearance of one poll showing different results than other polls does not mean the entire race has shifted.

Another lesson: A great candidate with an excellent ground game can move the turnout mix in a favorable direction. And we need to beware likely voter models that take voting history into account in the face of a great candidate with an excellent ground game.


3 Comments . Leave a comment below.
  1. Thank you Nate Silver

    As a devotee of Nate Silver, there were times when I was watching the EV sitting higher for Romney, but knowing the numbers for the entire West Coast was not present yet. It definitely kept me from freaking out and made the Karl Rove meltdown all the more funny/pathetic.

    Hint: if you want some major Shadenfraude, check out RMC, where someone is opining that Scott Brown lost because he wasn’t Republican enough. Of course, you will have to skim through the normal psycho accusations.

  2. "Unskewed Polls"

    If I wanted to create a masterstroke of internet trolling against low information conservatives/tea party types, I would do exactly what the Unskewed Polls website did. I would have been Romney-leaning and just credible enough to that demographic for months and then make a wild, landslide prediction at the end to really get their hopes up.

    I heard a lot from Conservatives interviewed by outlets such as NPR talking about how Unskewed Polls was reflected the “real” data; boy would I love to see a montage of Unskewed loyalist downs as they came to grips with what actually transpired last night.

    • just such an event...

      boy would I love to see a montage of Unskewed loyalist downs as they came to grips with what actually transpired last night.

      … is scheduled for sometime in early 2015: there will be a 5 minute window wherein deep deep soul searching will occur over the events of ’12 after which both Jeb Bush and Marco Rubio will break wind simultaneously and all will take this as a divine sign for the new new Jesus ticket. Then we do it all over again.

« Blue Mass Group Front Page

Add Your Comments

You must be logged in to post a comment.

Tue 25 Apr 12:56 AM