CBS news poll weighted incorrectly...

[VirginiaDem]

I went into the pdf file of the CBS results from their Website located here:
http://www.cbsnews.com/htdocs/CBSNews_polls/campaign90904.pdf

At the end they list the weighted sample sizes. Out of 1059 respondents, they used 365 independents, 354 Dems and 340 Repubs. This works out to about 34.5% indep., 33.5 for Dems and 32 for Repubs. Shouldn't it be about 37-38 for Dems, 31-32 for Repubs and about 32 for independents? This isn't as badly weighted at the Time/Newsweek polls but it seems like it misses. Or am I missing something? Pollsters...? If I'm correct with my above assumptions then, just ballparking here, but the real CBS Bush lead is probably no more than 3 or 4 points.

[PROGRESSIVE1]

1. Dem 33%/Rep 30%/Ind 37%

[VirginiaDem]

3. what does this mean?

that those are the "legit" numbers? If so, then this poll is about right.

[King Coal]

2. Check this out.

http://www.emergingdemocraticmajorityweblog.com/donkeyrising/

[VirginiaDem]

6. according to the above link

But in 2000, according to the VNS national exit poll (which hits the overall percentages for Bush and Gore right on the nose), Democrats made up 40.3 percent of the electorate while Republicans made up only 36.5 percent of the electorate."

Which means, I suppose, that Dems should get 3.8 percent more than Repubs.

[rinsd]

9. I prefer registered voters to any exit poll numbers....

At least in terms of breakdonw of parties.

[cheshire]

15. Thanks for a link I don't have. It seems full of info.

[lancdem]

4. There should be more Dems and Repubs and fewer indys

The gap between Dems and Repubs should be 4 percent or so. This is not nearly as bad as Time and Newsweek.

[On the Road]

5. CORRECTED: In the *2000* Election

I thought the numbers were approximately equal for Dems and Reps. So any poll which attempts to reflect this year's election should show some kind of parity between the parties, no?

[VirginiaDem]

7. no, see post 6 <eom>

[VirginiaDem]

8. my bad

I see now you're talking 2002. I would assume that it would be better to use 2000. Incumbents and incumbent parties do poorly in midterms; I assume this is partially due to turnout.

[On the Road]

12. No, My Bad -- I Meant 2000 (not 2002)

I assume Reps had higher turnout than Dems in 2002. Come to think of it, maybe that's the justification for these poll weightings. Not that that will happen this year.

[knowbody0]

10. polls and unemployment statistics

unbelievable- when your unemployment runs out after 26 weeks you are no longer counted as unemployed. this statistic only counts unemployment claims. how about just a # for the unemployed?

[newyawker99]

11. Hi knowbody0!!

Welcome to DU!! :toast:

[knowbody0]

14. kerry history

way back when 300 body bags a week were coming home to the usa, kerry was compelled to expose the attrocities and insanity of viet nam. so now today as i watch cross fire the bald fish eyed dude says "he made it sound like viet nam vets were psychotic drug addicted maniacs" well they were. and now in iraq there have been 26 suicides. mayhaps it is important to revisit the dark days of 35 years ago.the pentagon investigation that was quelled in 1975, the legacy of the broken soldiers all these years later. it is official viet nam was not a just war, not engaged in accordance with geneva convention. history is repeating itself.another president who refuses to "loose a war" yuk..

[tritsofme]

13. Nope, not true

The unemployment rate and unemployment benefits are unrelated.

The unemployment rate is calculated by the BLS through the Household Survey.

Gee, I usually only have to do this at the beginning of every month, not in the middle.

[ParanoidPat]

16. You might want to check with the Ministry of Creative Mathematics.....

.....:evilgrin: Here's a good resource page for raw numbers.

http://www.bls.gov/data/home.htm

Here's the July 2004 report in PDF with some description of the methodology.

http://www.bls.gov/news.release/pdf/laus.pdf

<Snip>

Method of estimation. Effective January 1996, estimates for all
states, the District of Columbia, the Los Angeles-Long Beach
metropolitan area, and New York City are produced using estimating
equations based on regression techniques. This method utilizes data
from several sources, including the CPS, the CES, and state
unemployment insurance (UI) data. A detailed description of the
estimation procedures is available from BLS upon request. Effective
January 1998, estimates for census regions and divisions are obtained
by summing the model-based estimates for the component states and
then calculating the unemployment rate.

<Snip>

Additional information

More complete information on the technical procedures used to
develop these estimates and additional data appear in Employment and
Earnings, which is available by subscription from the Superintendent
of Documents, U.S. Government Printing Office, Washington, DC
20402 (telephone [phone]), and from the BLS Handbook of
Methods, Bulletin 2490, August 1997.

Estimates of labor force and unemployment for over 330
metropolitan areas are available in the news release, Metropolitan Area
Employment and Unemployment. Estimates of labor force,
employment, and unemployment for all states, metropolitan areas,
labor market areas, counties, cities with a population of 25,000 or
more, and other areas used in the administration of various federal
economic assistance programs are available on the BLS Web site at
(http://www.bls.gov/lau/). Employment data from the CES program
are available at (http://www.bls.gov/sae/).

Information in this release will be made available to sensory
impaired individuals upon request. Voice phone: [phone]; TDD
message referral phone: [phone].