Yesterday, I noticed that ABC News and the Washington Post released a poll on Bush's job approval that was significantly
out of whack with nearly every other polling organization. Originally, I had not planned to blog about this poll, since I believe that the majority of polling tells the most accurate story about the state of public opinion, rather than one or two outliers. However, earlier this morning I noticed that this poll had become one of the top stories in
the U.S. section of Google News, so I felt compelled to respond.
However, when I was conducting research for my response, I noticed something wonderful. It turns out that Richard Morin, the poll editor for the Washington Post, will hold a live chat about their latest poll at 1pm eastern. What better way to respond to the poll than by directly chatting with the poll editor himself! Truly, this is an early, nerdy Christmas present for all of us poll junkies out there who would like to see more responsible polling from commerical news organizations.
I have a lot to ask Mr. Morin, but I imagine that I will only be allowed one question. As such, I urge as many people as possible on MyDD to take part in this chat, so as to better insure that he is presented with all relevant and appropriate questions.
Since ABC News / WaPo was one of two original recipients of The Distorty Awards, my question will focus on their Iraq polling. Interestingly, much like Pew, in their latest effort since they received the "Distorty," ABC / WaPo backed off their long-standing withdrawal question, and instead released the following withdrawal question:
"Do you think the number of U.S. military forces in Iraq should be increased, decreased, or kept about the same?"
This is a noticeable improvement, since this question does not force those who support partial withdrawal to choose between keeping all troops in Iraq indefinitely and withdrawing all troops from Iraq immediately. It is not a perfect question, since it still groups highly divergent opinions into identical responses, but it is an improvement. I am going to ask if they plan on continuing to use this question, or if they will revert to categorizing all of those in favor of withdrawal as being in favor of total and immediate withdrawal.
There are, of course, many other questions that need to asked. Here are just a few topics that jump to mind:
- How come ABC News / WaPo poll has not yet polled on impeachment?
- What is the partisan breakdown in your latest poll?
- Why don't you make more internal numbers from your surveys available to the public?
- Are you disturbed that your poll is in such great conflict with other recent polls, such as Gallup, Opinion Dynamics, NBC, Pew, AP-Ipsos and Quinnipiac?
Stop by and ask a question. The chat starts at 1pm, but you can submit your question beforehand. These polls do not just report public opinion--they also help shape public opinion. Let's use this opportunity to get some answers from one prominent pollster.
Update: Mystery Pollster writes on the partisan breakdown of the conflicting ABC and Gallup polls:
One possible clue to these differing trends is this line from the ABC News summary: "Republican self-identification is up six points, to 33 percent of the public." As of this writing, the Washington Post has not yet released results for party and sample demographics (as it has for other polls in 2005). On its last two surveys, the ABC Post polls showed a four point Democratic edge in party ID (31% to 27%).
Gallup reported party identification for this latest survey at 31% Republican, 32% Democrat and 36% independent. That mix is roughly the same as their average party identification result for the last three months (32% Republican, 33% Democrat, 34% independent).
Considering that the Washington Post released partisan ID info for other polls in 2005, I find it odd that that the Washington Post has not yet released its partisan breakdown for this poll. Also, it is clear that the jump in the ABC-WaPo poll is in fact the result of a massive jump in Republican identification in the ABC-WaPo survey.
As I wrote often in 2004, I still maintain that such rapid movement for partisan self-identification is bogus. If there was indeed such rapid movement in partisan self-identification, then annual results on partisan self-identification produced by
Harris,
Pew and
NAES would show far more volatility. They don't, and the reason is because, barring a national realignment, partisan self-identification changes only slowly.
