Saturday, September 20, 2008

Errata

I made a mistake in yesterday’s post. Big enough that it needs its own post to address, but I also note that the critics themselves made a goof. What happened was, I took the most recent published overall results from Gallup, and applied them to the most recent party-specific results presented by Gallup, and walked each back by weeks to show the progression of results. Since the party-centric data was not aligned with the overall data, I did in fact submit an erroneous comparison. Here’s how that data should be presented (LD = liberal democrat, MD = moderate democrat, CD = conservative democrat, IN = independent, LMR = liberal/moderate republican, CR = conservative republican) :

BARACK OBAMA
August 10: 47% Overall, 91% LD, 79% MD, 67% CD, 22% IN, 19% LMR, 6% CR
August 17: 46% Overall, 88% LD, 78% MD, 68% CD, 24% IN, 16% LMR, 6% CR
(loss of 1% overall against gains in 2 categories, losses in 3 categories)
August 24: 45% Overall, 91% LD, 78% MD, 63% CD, 29% IN, 13% LMR, 5% CR
(loss of 1% overall against gains in 2 categories of 3% or more, losses in 3 categories, 2 of 3% or more)
August 31: 49% Overall, 93% LD, 81% MD, 77% CD, 23% IN, 14% LMR, 4% CR
(gain of 4% overall against gains in 4 categories, 2 of 3% or more, losses in 2 categories, 1 of 6% or more)
September 7: 44% Overall, 93% LD, 81% MD, 70% CD, 29% IN, 16% LMR, 3% CR
(loss of 5% overall against gains in 2 categories, 1 of 6% or more, losses in 2 categories, 1 of 7% or more)
September 14: 45% Overall, 93% LD, 81% MD, 66% CD, 27% IN, 10% LMR, 3% CR
(gain of 1% overall against gains in no categories, losses in 3 categories, 1 of 4% or more)

This is a milder shift, but as I said before, this still shows that the Gallup Organization shifted party affiliation weights on a regular basis, invalidating the statistical impetus of voter responses. The last week in particular, where Obama’s overall support is alleged to have increased when he stayed steady or lost ground in every category. Looking now to McCain:

JOHN McCAIN
August 10: 42% Overall, 5% LD, 13% MD, 20% CD, 33% IN, 70% LMR, 90% CR
August 17: 44% Overall, 6% LD, 14% MD, 23% CD, 34% IN, 75% LMR, 89% CR
(gain of 2% overall against gains in 6 categories, 2 by 3% or more, losses in 1 category)
August 24: 45% Overall, 6% LD, 13% MD, 26% CD, 31% IN, 77% LMR, 91% CR (gain of 1% overall against gains in 4 categories, 1 by 3% or more, losses in 2 categories, 1 by 3% or more)
August 31: 43% Overall, 4% LD, 11% MD, 15% CD, 29% IN, 78% LMR, 94% CR
(loss of 2% overall against gains in 2 categories, 1 by 3% or more, losses in 4 categories, 1 by 11% or more)
September 7: 49% Overall, 4% LD, 12% MD, 21% CD, 28% IN, 78% LMR, 94% CR
(gain of 6% overall against gains in 2 categories, 1 by 6% or more, losses in 1 category)
September 14: 47% Overall, 5% LD, 12% MD, 24% CD, 32% IN, 85% LMR, 95% CR
(loss of 2% overall against gains in 5 categories, 3 by 3% or more, losses in no categories)

Again, look closely at the last week, where McCain’s overall support is alleged to have decreased when he stayed steady or gained ground in every category. There’s no way an unbiased reader can call that a clean poll, with all due respect to Gallup. This was spin, and no question about it. The main point of my poll report remains correct, therefore, though I expect the trolls to whine all the same. It’s what they do.

(Overall support as reported by Gallup, at
http://www.gallup.com/poll/107674/Interactive-Graph-Follow-General-Election.aspx)

(Party-specific support as reported by Gallup, at
http://www.gallup.com/poll/108049/Candidate-Support-Political-Party-Ideology.aspx)

No comments: