MURDER, POLITICS, AND THE END OF THE JAZZ AGE
by Michael Wolraich
Order today at Barnes & Noble / Amazon / Books-A-Million / Bookshop
MURDER, POLITICS, AND THE END OF THE JAZZ AGE by Michael Wolraich Order today at Barnes & Noble / Amazon / Books-A-Million / Bookshop |
The Gallup tracking polls on which I've been wasting good space here aren't worth the effort. IMHO successive reports appear to show important changes which are purely the result of Gallup's methodology
Consider Obama's rating in the polls issued on
10/21 10/23 and 10/26
45 47 46
How could Obama’s ratings change like this? It’s in the roundings. IMHO if Gallup published this info carried out to the first decimal point these variations would be seen to be trivial .Consider the following illustration in which I have created some imaginary daily data which calculates to the totals Gallup published.
Oct 21 Oct 23 and Oct 26th.
October 14th 45.0 % 16 45.0 % 19 45.5
15 45.0 17 45.0 20 46.0
16 45.0 18 45.5 21 50.0
17 45.0 19 45.5 22 49.2
18 45.5 20 46.0 23 45.0
19 45.5 21 50.0 24 44.0
20 46.0 22 49.2 25 44.0
Average 45.3 46.6 46.2
Gallup would use 45 47 46
Sorry to have used a sledge hammer to crack a nut but I've been unable to follow the wise advice of Articleman and just ignore Gallup.Now I will. You should too. Altho you probably have sensibly made that decision already.
Comments
Nate Silver appeared on "Real Time" on Friday. He viewed Gallup using the analogy of 4 out of 5 doctors suggest that lowering cholesterol is important in preventing heart disease. Gallup is the fifth doctor. I think Gallup is suffering from sampling error.
by AnonymousRm (not verified) on Sun, 10/28/2012 - 9:45am
I agree .But separate from that by rounding they exaggerate the apparent trends with its sample.
by Flavius on Sun, 10/28/2012 - 10:14am