Posted on 07/07/2004 6:42:56 PM PDT by GAGOPSWEEPTOVICTORY
CBS and the New York Times are at it again -- distorting polling methods to create "news" showing a Kerry lead in the polls to shift the perception of the American public.
We all remember the posts of a week ago, when a CBS News/New York Times poll showed Kerry with a one-point lead over President Bush (45-44%). What the Times didn't tell you was that the poll itself oversampled Democrats by 10%.
After the selection of Edwards, CBS News and the Times conducted another poll.
The result: a predictable Kerry/Edwards lead of 49% to 44%.
Ah, but the devil is in the details and CBS and the Times really outdid themselves this time and of course, the details are listed way at the bottom of their article.
This poll was conducted among a nationwide random sample of 462 registered voters interviewed July 6, 2004. These respondents had originally been interviewed by CBS News and "The New York Times" June 23-27, 2004.
So what did CBS and the Times do? They used the same tainted sample of "registered" voters, and contacted them again last night about the Kerry/Edwards ticket.
Obviously, they could not contact everyone in that first tainted sample, and were able to get in touch with less than half of them (462 as opposed to 1,000 plus).
Curiously, CBS didn't publish any additional information on the polls methodology -- particularly about party affiliation, which tainted the poll the first time. (Perhaps the Times will shortly.)
The bottom line is that CBS and the Times continue to abandon valid polling techniques (e.g. likely vs. registered voters, proper R/D/I split) to put polls out there that skew their polls to favor their political agenda. Don't believe anything they tell you.
Overnight polls are garbage designed to make to news, not cover it.
Nothing more to say, other than my tagline.
Edwards's idea of fighting terrorism invloves hiding behind a liberal judge and suing Osama.
Hidden polling methodology? Hmmmmmmmm...
Wow, that's pretty bad if the same Democrat-oversampled group they hit on last month STILL can't push Kerry over 50%, only a 4 pt boost on a vp pick...LOL.
Not that it matters much, because they are in the same ideological boat,but wasn't the discredited poll to which you refer a product of the Los Angeles Times rather than the New York Times?
No, I did a search here and looked at the earlier CBS/NYT poll to confirm. If I remember correctly, the LA Times used similar techniques with their poll earlier in June as well.
It bothers me to no end that because of the Times' alleged credibility, this "poll" will be picked up by AP and the other lamestream news sources and treated as "news" for the next cycle.
we've got a problem here. Whether it's the media disorting the news and covering for socialists, or one of my states US reps calling for UN oversight for OUR elections.
This should scare the hell out of everyone.
Pray for W and The Truth
"We all remember the posts of a week ago, when a CBS News/New York Times poll showed Kerry with a one-point lead over President Bush (45-44%). What the Times didn't tell you was that the poll itself oversampled Democrats by 10%."
While the New York Times/CBS News poll you refer to was skewed, it did not oversampled Democrats by 10%.
CBS News/New York Times Poll
Campaign 2004: The Race for the White House Tightens
June 23-27, 2004
PDF Page 2/10:
q5 If the 2004 presidential election were being held today and the candidates were John
Kerry, the Democrat, and George W. Bush, the Republican, would you vote for John Kerry or
George W. Bush? [CANDIDATE NAMES AND ANSWER CODES WILL ROTATE]
*** REGISTERED VOTERS ***
*** Party ID ***
Total Rep Dem Ind May04b
% % % % %
John Kerry 45 4 82 44 49
George W.Bush 44 90 10 37 41
Other (Vol.) 1 0 0 2 1
Won't vote (Vol.) 1 0 2 2 1
Depends (Vol.) 2 2 2 2 2
DK/NA 7 4 4 13 6
Some have commented that they believe that the NYT/CBS poll used an unfairly skewed "50% Democrat/40% Republican/10% Independent" party affiliation breakdown. Not so, because we actually have the New York Times/CBS News poll with a nation-wide public party affiliation profile of 30.29% Republicans, 33.03% Democrats and 36.69% Independents in the 875 Registered voters who were asked the above listed poll question #5. This is not good, because there is no way that there are 36.69% of the nation-wide American voting population who are self-identified as "Independent" voters. Independents are much closer to being around 10% of said population. This is the largest skew factor in the poll, as the NYT/CBS "Independent" poll respondents favored Kerry over Bush by 44% to 37%. The fact that the NYT/CBS poll has 2.74% more Democrats than Republicans is of course a factor, but a actually a lesser one that the over-representation in the number of Independents poll respondents.
The Gallup polling organization has the nation-wide political party affiliation breakdown as of January 2004 as 45.5% Republican, 45.2% Democrat and only 9.3% Independents. If we use the more accurate Gallup nation-wide party affiliation percentage numbers with the raw data from the New York Times/CBS News poll, what would the revised poll result look like? Look below:
Original NYT/CBS Poll Data Revised NYT/CBS Poll "Results"
================================== =====================================
John Kerry 45% | 43.3% |
George W. Bush 44% | Rep: 30.29% 48.8% | Rep: 45.50%
Other (Vol.) 1% |- based on: Dem: 33.03% 1.0% |- based on: Dem: 45.20%
Won't vote (Vol.) 1% | Ind: 36.69% 0.8% | Ind: 9.30%
Depends (Vol.) 2% | Total: 100.00% 1.2% | Total: 100.00%
DK/NA 7% | 4.9% |
------ --------
100% Stated MOE: 3% 100.0% MOE: 3.3%
Source: CBS News/New York Times Poll Campaign 2004: The Race for the White House Tightens. June 23-27, 2004. PDF page 2/10, 10/10
Without the internals, we have no way of knowing what the subset of the previous voter survey looks like, so the just published subset poll cannot be trusted...
dvwjr
If the vanity post is true then this is a good sign. It means the Libs are into fantasy when it comes to Pres. Bush and their skewed polls are a means of assuaging themselves.
This refusal to deal with reality means they'll take less action fair or foul -- legal or illegal during the campaign.
Hopefully the Times will post the same subset information when they go to print later. I couldn't find it anywhere on CBS's site and the Times hasn't posted the story yet.
Is a subset poll even random a random sampling anymore if some are home and some are not when you call?
While we may not agree as to the oversampling percentage, we do agree that there is a liberal dim skew, correct?
What the Times didn't tell you was that the poll itself oversampled Democrats by 10%.
We don't know what numbers that CBS/NYT used yet in the subset poll, but look at dvwjr's post as to the numbers in the first poll.
Also keep in mind that this is a "registered" and not "likely" voter poll to begin with, which generally skews things in the Democrat's favor by a handful of points.
Thanks..... Just curious on what bases they define what the proper number is to use among the various groups... The 2000 election voting pattern was something along the lines of 39% democrat, 34% republican and 27% independent....... Then you can take it even farther to race, income, age, education, liberal, conservative, moderate, etc. etc... Sometimes they adjust their numbers to get back to a dispersion pattern they want and you don't know if they did unless they give you the details of the poll's internals.... I guess the question are what is the correct dispersion patterns?
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.