Posted on 01/13/2006 9:17:56 AM PST by rabair
Here are the most important parts... but my blog post goes deep in to what seems to have happened here, etc. http://www.rightwinged.com/2006/01/fox_news_poll_on_electronic_su.html
(from AP-Ipsos, a couple days ago)
56 percent of respondents in an AP-Ipsos poll said the government should be required to first get a court warrant to eavesdrop on the overseas calls and e-mails of U.S. citizens when those communications are believed to be tied to terrorism.Agreeing with the White House, some 42 percent of those surveyed do not believe the court approval is necessary.
(from Fox News-Opinion Dynamics poll out today)
By 58 percent to 36 percent, Americans think the president should have the power to authorize the National Security Agency (NSA) to monitor electronic communications of suspected terrorists without getting warrants, even if one end of the communication is in the United States. Furthermore, six in 10 say they are personally okay with the NSA monitoring their international telephone calls.
Again, get the breakdown here: http://www.rightwinged.com/2006/01/fox_news_poll_on_electronic_su.html
I suspect that's the problem, and that's what my blog post was all about. We know how the MSM has a habit of asking questions in a tricky way to get their desired result... I bet that is why there is such a huge discrepancy here.
And I would like to compare the phraseology, only I couldn't find the AP's... I may not have known where to look, but I did check a few places and didn't come across it... where as Fox had a link to their's posted right in the article.
I got an email that said the same thing, but I wouldn't post that on my blog unless I had some actually evidence... although it's most likely the case
Congressman Billybob
Latest column on Newsbusters.org: "AP Poll Biased: Anti-Bush, Anti-Republican"
You arrive at 12 percent, I think, by faulty math, that is, subtraction. I believe Democrats were oversampled by 30 percent.
There were 40 percent Republicans and 52 percent Democrats questioned.
I believe you divide the excess by the lower figure to get the correct percentage, that is, 12 divided by 40=30 percent.
awesome, thanks Congressman Billybob.
found it. here's where i originally saw it.
Thanx again
http://www.freerepublic.com/focus/f-news/1553329/posts
On the surface there appears to be a clear problem with sample and a presumption that Republicans are under-represented. However, statistically speaking there is not enough information to calculate a measure of the bias in this poll - if there is any. One correct calculation of over- or under-sampling of Democrats is given by:
(actual % Democrats sampled (52%)minus % of Democrats in total population)divided by % of Democrats in total population.
Clearly this calculation requires us to know the % of Democrats in the total population.
There are other ways to calculate the bias. However, there is also an issue of how Republican and Democrat are defined. Is it a "likely voter" poll, a "who did you vote for in 2004" poll, etc. The permutations are endless and as is the room to influence the poll through manipulating the sampling. Any professionally responsible polling company will make clear the sample definition. Of course, the journalists have to understand what it means and what bias the sample will create in the results.
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.