Newsweek Poll Undersamples Republicans by 10 Points: Only 27% Republican (Election 2004 - 37% GOP)

PR Newswire via Yahoo ^ | November 12, 2005

[new yorker 77]

LAST QUESTION in the polling data:

D2. Now I have just a few more questions so we can describe the people who took part in our survey ... Regardless of how you might have voted in recent elections, in politics TODAY, do you consider yourself a Republican, Democrat, or Independent?

27 Republican

37 Democrat

32 Independent

2 No party/Not interested in politics (VOL.)

1 Other party (VOL.)

1 Don't know

100

[new yorker 77]

More Bias.

[rocksblues]

And this is surprising because? /sarcasm

[right right]

They would lie. No way.

Liberals would lie when it would benefit them to tell the truth.

[new yorker 77]

Just because it is consistently biased does not mean it is not relevant to expose it every time.

[Right Wing Assault]

That's not a problem, is it? /sarc

[italianquaker]

tagline says it all

[bray]

Of course when DNCNN reports his terrible poll numbers they will state this is a skewed poll. Wont they?? What absolute garbage!

Pray for W and Our Troops

[Digger]

This is not scientific however IMO this is how true conservatives see this administration: In the poll by Viguerie's ConservativeHQ.com:

69 percent of those surveyed said Bush is not governing as a conservative. 62 percent said they were "disappointed" or "angry" over President Bush and the Republican-led Congress. 70 percent said they would support "a principled conservative candidate running against an establishment Republican incumbent in a 2006 Republican primary." 25 percent said they would reduce their financial support for the GOP in 2006, and 27.4 percent said they would end it completely. Regarding the 2006 election, 13.36 percent agreed with the statement: "I'm so disappointed that I may sit this election out." Respondents gave Bush a grade of D for "controlling government spending" - and a D- for "reducing illegal immigration." "To save his presidency and prevent devastating Republican losses in '06, President Bush needs to take drastic action by replacing most of his White House policy personnel with effective, principled conservatives," said Viguerie, who pioneered ideological and political direct mail and helped elect Ronald Reagan in 1980.

[RWR8189]

Newsweek uses the Gallup model in their party ID figures, this is what Gallup said last year about the issue your bring up, when they were accused of "oversampling" Republicans:

26-Sep 2004 8:14 pm

Steven writes in with a question that reflects misinformation that is boomeranging around the Net. We have posted blogs below that go into some detail outlining the whole issue of party identification in a survey context. Our Gallup samples are rigorously executed and checked and weighted against a number of know U.S. Census Bureau parameters: age, gender, region of country, race, and education. Party ID ("In politics, as of today, do you consider yourself a Republican, a Democrat, or an independent?") is not a variable that is measured by the U.S. Census Bureau, is not fixed, and in fact is to a significant degree a measure that is quite likely to change based on the environment. After 9/11, polls showed many more people identifying with the Republican Party than the Democratic Party because citizens were rallying behind the president. This winter during the primary season, polls showed more people identifying with the Democratic Party than the Republican Party because the news coverage was focused almost exclusively on the Democratic primaries. Analyses shows that polls had more identification with the Democrats than Republicans after the Democratic convention this summer, and then more identification with the Republicans than Democrats after the Republican convention. The measure of partisanship we and other pollsters use is not measuring some lifelong fixed value like gender or race. It is an attitudinal identification with one or the other party at the time of the survey. So, if there are forces at work out in the environment that are favorable to the Democratic Party, for example, they will cause more people to identify with the Democratic Party in the survey, and also cause more people to say they will vote for the Democratic candidate.

Here are links to two recent summaries by other students of polling that go over this same concept.

http://mysterypollster.typepad.com.

http://people-press.org/commentary/display.php3?AnalysisID=97

This whole issue of partisan identification is one that pollsters and survey scientists have been discussing and dealing with for years in publications and scholarly conferences. It's not a new issue.

It's surprising that some people on the Net feel that they have suddenly "discovered" something about polling as if pollsters are not highly aware of the variables like party identification that we measure in each survey.

Gallup has a team of experienced editors who have been conducting polls for decades, and teams of statisticians and methodologists who work on every poll. All of this is not to say that there can't be legitimate scientific debate on this and other issues. There can be, just as heart surgeons have conferences and debate the value of different methods of conducting coronary artery bypass surgery. But I can assure all users of Gallup Poll data that the methods we use in pre-election polls are the results of about 70 years of experience in conducting them (since 1936) and intensive, ongoing study and examination of each element of the survey process.

[Alter Kaker]

The 'Rats cried fraud last summer at Gallup for exactly the same thing (if I remember, Moveon.org came out with an NY Times ad decrying the undercounting of Dems). What pollsters will tell you -- and, not being a pollster, I don't know if this is true -- is that party allegiance fluctuates wildly and that it makes no sense to control for it. If that's the case, both you and Moveon are wrong if the sample is truly random.

[ssaftler]

You're correct, it WAS biased. Usually Newsweak polls are 10% Republican, 28% Independent, 2% Raving Loony Party (and other such foolishness) and 60% Democrats, mostly from Boston, New York and San Francisco.























</sarcasm>

[NewMediaFan]

What's up with the MSM's blizzard of new polls? It's not like there's an election any time soon.

[Dog Gone]

What I don't understand is why they make their bias open to anyone who wants to look at the details. They could either withhold the information or simply lie about it.

Why let us know that the poll is fatally biased?

[new yorker 77]

What I will in a couple days is get the internal data from Gallup in detail in the PR Newswire of yahoo.

Then I can dissect the bias.

Remember, CNN pays Gallup's bills.

Many polling firms try to get their act together at the last minute to save face. Th e most biased consistently produce fake polls until the last minute.

Just ask the L.A. Slimes how it's done. They did it in 2003 with Arnold.

[bray]

ping me when you do. Most of us are wise to their tactics.

Pray for W and Our Troops

[RWR8189]

See post 9.

As Alter Kaker said, you criticism is just as silly as MoveOn.org's was last year.

[new yorker 77]

Listen if you want to be a liberal, go right ahead.

Pointing out the FACT the Newsweek polled 10% less Republicans than the 2004 election is not irrational.

You seem agitated like a liberal poser.

27% in a poll is not 37% at the polls.

That's MATH.

[altura]

Put most of those independents in the liberal column too.

[new yorker 77]

Your response is uneducated.

Is 27% the same as 37%?

If you want to defend Newsweek for clearly being biased and caught, then you are as conservative as Hillary.

[RWR8189]

Did you read the analysis given by Gallup?

Did you understand it?

You are making the same argument that MoveOn.org made in 2004 when Gallup was "oversampling" Republicans.

An exit poll is not a census, and party ID is not something that should be strictly controlled.

Calling me a liberal "poser" is childish and asinine, I am am merely being consistent.

Go back and look at the same arguments I made a year ago when Gallup was accused over "oversampling" Republicans.