your orgtheory pennsylvania primary post

Quick hits:

1. Prediction: 54-46 for Clinton, with about +10 or so in the delegate count. Logic: polls show Obama solidifying a 40-45% base and he’ll get about 1/3 of the undecideds. My gut feeling is that PA is kind of like NH – Hillary’s numbers are deflated by Hillary voters claiming to be undecided when they’re not. Ultimately, this changes little because an Obama win in NC and tie in IN brings Obama’s pledged delegate lead back up to about 150, with all small states remaining after that.

2. Around 6:07 you can see me shake hands with Obama.

3. In the comments to this post, Belle Lettre claims I’m a bigger nerd than Jeremy Freese. I bet you didn’t think it was possible.

4. The rest of the primaries: PA will be canceled by IN/NC. WV/KY will be offset by OR/MT/SD/Guam. Hillary will get a net gain from Puerto Rico. By the end, Hillary will reduce the delegate gap from 150 to about 130. Of course, the trickle of Obama add-on delegates and superdelegate switch/endorsements increases his lead. By June, he may only need about 90-100 of the currently non-endorsing supers to clinch it, which isn’t so hard.

Written by fabiorojas

April 22, 2008 at 1:28 am

5 Responses

Subscribe to comments with RSS.

  1. By June, he may only need about 90-100 of the currently non-endorsing supers to clinch it, which isn’t so hard.

    Fabio, your ability to power on with such certainty in light of past performance is impressive. I think a more substantive ‘Clinton is Toast’ title will look better for posterity…



    April 22, 2008 at 1:37 am

  2. Is this a test of meta-analysis? Most of the polls in the last five days or so favor Clinton by an average of 7-8 points. Only one poll favors Obama (46-43%) but that poll is the one with by far the largest N — 2300 compared with 600-750 for most of the others.


    Jay Livingston

    April 22, 2008 at 12:59 pm

  3. Peter: #1 rule of punditry – don’t let the past deter you! Also, I got a lot right about the primary so far: who won Iowa and New Hampshire, who won the big states, Hillary’s dogged base of support, the GOP race.

    The big problem of my earlier predictions was that I completely underestimated Obama’s mobilization capacity in small states, which brought him victory on Super Tuesday, and Hillary’s complete lack of preparation for the February round of primaries. That allowed Obama to turn a rather slim delegate lead on Feb 5 into a virtually unbreakable lead by March 1.

    Jay: 54-46 is within that 7-8% lead.



    April 22, 2008 at 2:33 pm

  4. Fabio, what I meant by the meta-analysis question was this: you have several small-n “studies” (pollls) that average out to an 8-point Clinton win — your 54-46 prediction if you distribute or eliminate the undecideds. But there’s one large-n study that has Obama winning by 3 points.

    I have no idea about the quality of any of these polls. But the analogy in my mind — using sample size analagous to “quality” for purposes of the discussion — was where you have a study you really think is first-rate but whose results are contradicted by a meta-analysis of a bunch of studies you don’t really know much about or that you think might be flawed.


    Jay Livingston

    April 22, 2008 at 11:57 pm

  5. Good point, Jay. I wonder what real meta-analysis of polls – versus averaging – would do to the results.


    Fabio Rojas

    April 23, 2008 at 12:55 am

Comments are closed.

%d bloggers like this: