Daly Thoughts has an excellent article about the New AP poll that just goes to show you their bias once again:
The latest AP/Ipsos survey finds Bush?s job approval at 42% with
56% disapproving. Those are the same numbers that were found by the polling organization in mid-July, the last time they conducted a survey.55% disapproving. The approval rating is unchanged since the last time they conducted a survey in mid-July. The disapproval rating was one point higher then.
… First is that the survey?s sample has 10 points more Democrats than Republicans among registered voters. By comparison, surveys conducted this year by Pew during 2005 among all adults have found the gap between the parties ranging from 0-4 points (and from 4-8 points when leaners are considered), with their most recent showing parity (or a four point Democrat advantage with leaners considered). The most recent CBS survey also had the gap at three points among all adults. There is something with the AP/Ipsos methodology that consistently yields a more Democratic sample than other surveys. This is a subject which I have blogged about before.
Um, Democrats had 10 more points sampled? No one see’s anything wrong with that?
AnkleBiting Pundits sure did:
Like all of you I woke up to the news of the latest Ipsos-AP poll and groaned. Bush?s approval on Iraq is down to 38%. His overall approval is a sorry 42%. And a ?growing? number of Americans view the president as ?dishonest.?
Well, those are the results you get if you ask a bunch of Democrats about a Republican president, I suspect.
Take a gander at Questions 2 a. ? d. in the toplines of the Ipsos poll (PDF). Of the 79% of respondents who are registered to vote, almost half (49%) describe themselves as ?strongly Democrat? or ?moderately Democrat,? whereas only 39% describe themselves as ?strongly Republican? or ?moderately Republican.?
As for the 21% of respondents who aren?t even registered to vote? Well, not to sound rude, but who really cares what they think?
Question D1c. demonstrates that Ipsos surveyed almost twice as many 18-34 year olds (31%) as they did 65-and-olders (17%). That?s not right.
Question D14 is curious, too. 14% of respondents make less that $15,000 per year?
Why didn?t they just survey the staff of the DNC?
As did The New Editor:
Can anyone say with a straight face that this is an accurate party breakdown of the American voting public? In what universe?
In the end, I just don’t get worked up over polls. Never have, even when Bush was way up high because I don’t believe them either way. Look at the election and the polls then. A few thousand people will never prove to me how the country is feeling about something.
See author page