Here’s another silly AP-Ipsos poll that purports that people in this country are so fed up with the GOP that the Democrats will regain control of the House and Senate in November.  Unfortunately for the AP, and everyone who reads the trash that they put out,  the results of the poll are so nebulous that coming to that conclusion seems far more like wishful thinking than cogent analysis.
The first thing you should look for in any political poll is the political affilation of those who were polled. The political affiliations should mirror the political affiliations of the general public, and if they don’t the entire poll is questionable. Very often the political affiliations of the people polled by left-leaning media organizations such as the AP will be skewed towards Democrats, and will therefore show that Bush’s approval ratings are lower than they actually are etc.  In today’s poll, the AP doesn’t even bother to tell us what the political affiliations are! Here’s what the AP says about the poll’s demographics:
Results were weighted to represent the population by demographic factors such as age, sex, region, race and income.
So we have no idea how many Democrats, Republicans, and independents there are, and therefore can not make any logical conclusions about what we’re reading. The AP however, assures us that the poll is statistically valid.
No more than one time in 20 should chance variations in the sample cause the results to vary by more than plus or minus 3 percentage points from the answers that would be obtained if all people in the U.S. were polled. The margin of sampling error for registered voters was plus or minus 3.5 percentage points.
Excuse me if I fail to take their assurances seriously, since they’ve proven many times in the past to intentionally oversample Democrats so that they can get the poll results that they want to get. Without knowing the party affiliations, questions like these are totally meaningless:
1. Generally speaking, would you say things in this country are heading in the right direction, or are they off on the wrong track?
-Right direction, 30 percent (26)
-Wrong track, 67 percent (70)
-Not sure, 3 percent (4)
2. Overall, do you approve, disapprove or have mixed feelings about the way George W. Bush is handling his job as president?
-Approve, 36 percent (35)
-Disapprove, 63 percent (63)
-Mixed feelings, 1 percent (2)
-Not sure, X percent (2)
3. And when it comes to handling the economy, do you approve or disapprove or have mixed feelings about the way George W. Bush is handling that issue?
-Approve, 38 percent (38)
-Disapprove, 60 percent (60)
-Mixed feelings, 1 percent (2)
-Not sure, 1 percent (X)
4. And when it comes to domestic issues like health care, education and the environment, do you approve or disapprove or have mixed feelings about the way George W. Bush is handling that issue?
-Approve, 36 percent (38)
-Disapprove, 62 percent (60)
-Mixed feelings, 1 (2)
-Not sure, 1 percent (X)
Likewise, if you look at the section of the poll that reports on which political party the person will vote for, it’s not possible to make heads or tails out of it.
8. If the election for U.S. House of Representatives were held today, would you vote for the Democratic or Republican candidate in your congressional district?
-The Democratic candidate, 47 percent
-The Republican candidate, 36 percent
-Other candidate (Volunteered), 3 percent
-Undecided (Volunteered), 11 percent
-None (Volunteered), 1 percent
-Will not vote (Volunteered), 1 percent
-Not sure, 1 percent
(QUESTION 8A ASKED OF THOSE WHO SAID UNDECIDED, NONE OR NOT SURE)
8a. Do you lean more towards the Democratic or Republican candidate in your district?
-The Democratic candidate, 26 percent
-The Republican candidate, 30 percent
-Other candidate (Volunteered), 1 percent
-Undecided (Volunteered), 32 percent
-None (Volunteered), 5 percent
-Not sure, 6 percent
(SUMMARY TABLE, REGISTERED VOTERS)
-Democratic candidate, 47 percent
-Lean Democratic, 3 percent
-Lean Republican, 36 percent
-Republican candidate, 36 percent
-Other candidate (Volunteered), 3 percent
-Undecided (Volunteered), 3 percent
-None, Will not vote (Volunteered), 2 percent
-Not sure, 1 percent
-TOTAL DEMOCRATIC CANDIDATE, 51 percent
-TOTAL REPUBLICAN CANDIDATE, 40 percent
At the risk of being redundant, without knowing who is being polled, these questions are meaningless. We also have questions asked where there are so many people who say they might change their minds before the election that the poll results aren’t  worth the paper they’re written on.
In November, will you definitely vote for that candidate, probably vote for that candidate, or do you think you could change your mind before the election in November?
-Definitely will vote for Democratic candidate, 23 percent
-Probably will vote for Democratic candidate, 15 percent
-Could change mind (Democrat), 17 percent
-Definitely will vote for Republican candidate, 17 percent
-Probably will vote for Republican candidate, 12 percent
-Could change mind (Republican), 13 percent
-Not sure, 3 percent
So on top of all the other issues with this poll, the results could change significantly prior to Election Day. Overall, this poll is a complete pile of garbage. What the AP wants is to try to manipulate people into voting Democratic, with the idea being that certain people will support whoever they think is going to win. If a voter who’s unsure of whom to vote for sees this poll indicating a backlash against the GOP, they more be more willing to vote Democratic than they would’ve before. I doubt if these polls actually accomplish their goal, but they do point out how biased and untrustworthy the Associated Press is, and why anything they publish should be taken with a huge grain of salt.
