The Alan Grayson Page

The Anthony Weiner Page

Guest Contributors


  • BN-Politics' administrators respect, but do not necessarily endorse, views expressed by our contributors. Our goal is to get the ideas out there. After that, they're on their own.
Blog powered by Typepad
Member since 05/2007

Blog Catalog

  • Liberalism Political Blogs - Blog Catalog Blog Directory



« Ex-Aide Says Bush Misled America re: Iraq War | Main | NYT & WaPo Mis-Report on Florida & Michigan, Ignore DNC's Errors & Hypocrisy »

May 29, 2008


Deb Cupples


I found this Gallup just after I wrote the last comment to you. I had you in mind when I posted it, though I STILL question polls and predictions.


Obama is afraid -very afraid- he doesn't win this nomination. You can see it all over him like flies on dog piles - the corporate media is also very afraid

I saw this news story - "A train derailed on the Lower South side of Chicago. It was on an elevated track - when the wreck was caused by the main opperator's error. No one was hurt." - and thought it was great analogy for what is happening to Obama's campaign.

I really believe that Hillary will do more for Women and minorities. That she will do more than Obama to lift everyone up to a more equal level when she becomes President. I don't see Obama doing that - I can't even see him winning in November. I see only destruction if he does. But with Hillary I do see the brighter tomorrow. I do see why a woman becoming President will do a lot more for this country and the rest of the World than for another American Male.

Deb Cupples

Hi Danny,

Yes, I too think that Obama (and certain of his media supporters) are very afraid. That's why they don't want FL to fully count (though it should).

That's why those media don't squawk about the DNC's failure to punish Iowa, New Hampshire and South Carolina for having also moved their primary dates up in violation of the rules.


They should apply all the rules consistently. I'm really sick and tired of the rules of convenience. If Hillary doesn't win this nomination, I'm pulling the plug on my cable - I wouldn't be able to bear to watch another minute of the crap the corporations spew on this nation on a daily basis.


Four quick comments:

1) You don't particularly trust pollsters in general, I know, but Gallup has been one of the weaker ones over this cycle.

2) These aggregated numbers over limited sets of states are sort of meaningless. It makes a lot more sense to look at state-by-state polls since the key number is 50%+1 in any given state.

3) Those two caveats aside, there's nothing fundamentally surprising here. We know that over the last two weeks, Obama's supporters have started saying they would vote for Clinton in a hypothetical general election in greater numbers. There's nothing here that can't be seen in the state-by-state maps.

4) I disagree with the approach they use to define swing states. Not only is the approach sort of weak, but then they arbitrarily add two and only two states - Missouri and Arkansas. It's increasingly clear that Virginia is in play for Obama, West Virginia is in play for Clinton, and Indiana may be in play for either.

Deb Cupples


I suspect that either of us could competently find flaws in any of the polling outfits' methods or assumptions (if they clearly described them).

That's why I don't put enough stock in polls to actually bet money on them (not that I'm a gambler, I'm not).

But it is fun to look at them, no?


Although I'm not a fan of Gallup in general, my criticism is not really about their methods or their accuracy. It's their presentation of data. They've spent a lot of time grouping and displaying data in ways that just obscures their meaning. It's arbitrary and unscientific.

I would MUCH, MUCH prefer if they just released their state-by-state polling results, preferrably with sample sizes as well. That would be vastly more useful. And more fun, too. ;)

I'm a bit of a gambler, but I only bet when I think the odds are stacked in my favor. You won't catch me playing roulette.

Deb Cupples


The grouping of data has something to do with the analysis in this case.

Frankly, I've seen a lot of "unscientific" polling results. I don't even think of polling as truly scientific, because so many assumptions go into the results.

My big thing has been that the calculation of a z-score and margin of error rests on the assumption that the central limit theorem is valid where voter opinions are concerned.

I just never did buy it, because we're not talking about throughly mixed, colored gum balls in a jar.


Agreed that m.o.e. calculations are inherently unreliable from pollsters. They would only be correct if the sampling of voters was from a pool that matched the actual voters perfectly, AND if everyone was perfectly honest. Neither of these things are the case. As such, m.o.e. in a poll should always be read as "best-case error window".

That said, some pollsters are better than others at their sampling and polling techniques, so they get closer to that ideal.

I would much rather just get the state-by-state results. The "analysis" here is pretty weak - it makes a lot of arbitrary choices and groups things in a way that denies the reader the important information; i.e. who is ahead in which states.

Deb Cupples

Maybe so. I know I don't blindly accept any polls -- even the ons you cite.


Just FYI, the guy who does has unmasked himself, and is now getting some MSM attention:

It's a nice piece about the odd world of polling.

The comments to this entry are closed.