04.22.08

Final Pennsylvania Polls: Undecideds break for Clinton? (Zogby C+10)

Posted in Pennsylvania, Presidential Primaries, poll result, presidential primary forecast tagged , , , , , , , , , , , , , , , , , at 7:59 am by xstryker

Latest Pennsylvania primary polls:

Pollster Date Obama Clinton
InsiderAdvantage 4/21 42 49
Zogby 4/20-4/21 41 51
Insider Advantage 4/20 39 49
Rasmussen 4/20 44 49
PPP 4/19-4/20 49 46
Quinnipiac 4/18-4/20 44 51
SUSA 4/18-4/20 44 50
Suffolk U. 4/19-4/20 42 52
Zogby 4/19-4/20 42 48
Strategic Vision 4/18-4/20 41 48
ARG 4/17-4/19 41 54
Zogby 4/18-4/19 43 46
Mason-Dixon/McClatchy 4/17-4/18 43 48
Zogby 4/17-4/18 42 47

To recap:
PPP: O+3
Rasmussen: C+5
Mason-Dixon: C+5
SUSA: C+6
Quinnipiac: C+7
Zogby: C+10
Suffolk U.: C+10
InsiderAdvantage: C+10
ARG: C+13

I’m very, very tempted to scrap my prediction from yesterday, decide that undecideds will break mostly for Clinton, and give her a 9-12% lead. But Zogby’s final day’s Ohio poll showed the race was a tie, which proved far less accurate than the polls from the day before (like SUSA’s dead-on 10 point margin). So, as I see it, there are 3 scenarios:

Scenario 1: SUSA is very accurate. This is the scenario I went with yesterday. Clinton wins by 4%-8%, with a likely median of 6%. SUSA nailed Ohio, California, and Massachusetts.

Scenario 2: Undecideds break for Clinton. This is the next most likely scenario. Clinton wins by 9%-15%, with a likely median of 12%. This is what Elliot predicts on his delegate-by-delegate breakdown. SUSA’s final Texas poll has Obama ahead by 1, when the final result was Clinton by 4%, neatly predicted by Zogby’s final day numbers. Zogby was also more accurate in Missouri, which SUSA botched.

Scenario 3: PPP is on to something. PPP nailed Ohio, Texas, and were closer than anyone on South Carolina. This scenario involved either an Obama win or a Clinton victory of 3% or less. Al Giordano of The Field predicts a Clinton win of 4.6%, which is close to this scenario. This scenario means that the other pollsters failed to properly model the surge in voter registration, and that tons of first-time Democratic voters will push Obama to outperform expectations.

After careful consideration, I am scrapping yesterday’s fling with 6%, and going back to what I predicted all of last week, which was “Clinton will win by 5%-15%, with a median of 9%” – and I will blend them. Clinton will win by 4%-13%, with a median of 8%, and gain 12 delegates. This is the median between scenarios 1 and 2, and assumes that SUSA was right, but then the undecideds break 2-1 for Clinton, giving a final result of Clinton 53.5%, Obama 45.5%. I think Dan Abrams predicted 8% last night as well.

UPDATE: Added InsiderAdvantage, which shows C+7, up 3 for Obama since yesterday. This doesn’t impact my prediction.

3 Comments »

  1. Rigg said,

    Somehow I am just bored with this thing.

    Both of these Democrats will be bad for out Country.

    Today I need humor,

    Check out this HIll-arious post at:
    EDIT BY XSTRYKER: No, you can’t have traffic for your right-wing nonsense, sorry.

  2. Pandora said,

    Way to go, X! I’m way too on edge to deal with right wing nonsense. Why don’t they find their own candidate to blast… oops! I forgot! They have McCain! I wish them luck.

  3. [...] other words, 9.2%. Still single digits, and still doesn’t round up to 10. So I was off by 1.2% and 1 delegate (so far). Not [...]


Leave a Comment