This article originally apeared at Free Republic and was posted by Thinkin' Gal, who had granted permission for it to be copied here.

Topic: Media Bias

CBS POLL EXPOSED: 1175 Average Americans = 40% Dem, 35% Ind, 25% GOP?!!?!?!!?

07 January 1999 T.G.

Dear Freepers,

I hope you will not count this as a vanity post. After seeing the CBS poll that was posted by Marcellus this morning, I could not sit idle. First, here is the LINK - it will open in a new browser window.

The poll not only gives the overall percentages, it breaks down those percentages by political party. It also gives the number of participants: 1175. I'm very rusty on my algebra, but something told me I could solve this equation and figure out how many people of each party were polled. Using the information from the first question as an example:

Do you approve or dissapprove of the way Bill Clinton is handling his job as President?

  All GOP Dem Ind
Approve 69% 43% 87% 67%
Disapprove 27% 51% 12% 27%
Don't know 4% 6% 1% 6%
Totals 1175 x y z

I converted these to formulas, and went back and forth solving for x (GOP), y (Dem), and z (Ind). I'll spare you the details, but I started with something like this (question no. 1 used as an example):

.69(1175) = .43x + .87y + .67z
.27(1175) = .51x + .12y + .27z
.04(1175) = .06x + .01y + .06z
1175 = x + y + z

It was further complicated by the fact that x, y, and z, are whole numbers, and the percentages could represent a different number of people. That is, 69% of 1175 people could statistically be be between 806 (68.6%) and 816 (69.4%). So I stuck with the true percentages when calculating, figuring I'd get statistically close! After a little trial and error, I determined that the following numbers worked out well (the totals held up when I analyzed other poll questions):

GOP - 290
Dem - 472
Ind - 413

Apparantly, CBS has determined that a representative sample of Americans is 40% Democrat, 35% Independent, and 25% Republican. So that means you've got at least 75% of the poll-takers as liberal, confused (not sure exactly what "Independent" means here), brain-dead, or all of the above. Now you know why approval ratings stay so high!

Below is the breakdown of question no. 1, using my totals:

  All GOP (x) Dem (y) Ind (z)
Approve .69(1175) .43(290) .87(472) .67(413)
Disapprove .27(1175) .51(290) .12(472) .27(413)
Don't know .04(1175) .06(290) .01(472) .06(413)
Totals 1175 290 472 413

Again, the actual number of people may vary (per cell), but they statistically all match the percentage totals given on the poll question. And all total 1175. Here it is with the math worked out:

  All GOP (x) Dem (y) Ind (z)
Approve 813 125 411 277
Disapprove 317 148 57 112
Don't know 45 17 4 24
Totals 1175 290 472 413

Please note: I believe my numbers are very close if not the same as the actual breakdown. However, considering the problem with percentages - that they can represent different numbers of people, and also that I checked several, but by no means all of the poll questions, I could be slightly off in my calculations.

Of course, somebody around here can probably stick this data in some basic software, and come up with the answer in 30 seconds. However, since I have yet to see this type of analysis posted, and since the cry of "bogus poll" never ceases, I went and did it by hand. Hopefully you can see how I did it, and will jump right in and dissect more polls! LOL!

Geez, I hope I did this right... I'll probably hear about it otherwise!

Back To The Farm House.
Mail to:

drupal statistics