The Hocus-Pocus in AP-Yahoo's Racism Poll

On September 20, Noel Sheppard of NewsBusters posted on a misleading Associated Press/Yahoo poll on racism. The poll asserted that if Barack Obama loses, it will be because of "[d]eep-seated racial misgivings" held by "one-third of white Democrats."

Later that day, NB's Michael Bates criticized the AP's report on the poll for its historically inaccurate claim that the US "enshrined slavery into its constitution."

NB's Lyndsi Thomas got into the neighborhood of the concern I'm about to note on Sunday, when she noted that the pollsters tried to ferret out racism by asking questions that could be seen as purely political and having nothing to do with race.

But it seems to me that the pollsters engaged in a bit of hocus pocus. These three paragraphs from a story explaining AP's methodology carried at the Minneapolis Star Tribune gave me that impression:

Since many people are uncomfortable discussing race with pollsters and others they do not know, the poll also used subtler techniques.

For one thing, the survey was conducted online, as have all AP-Yahoo News polls since they began last November. Studies have shown people are more willing to reveal potentially unpopular attitudes on a computer than in questioning by a live interviewer.

The poll also used a technique aimed at measuring what psychologists call "affect misattribution." This involved showing faces of people of different races quickly on a screen before displaying a neutral image that people were asked to rate as pleasant or unpleasant. Studies have shown that people consciously or unconsciously transfer their feelings about the photograph to the object they are rating.

I think that whether or not people are "more willing to reveal potentially unpopular attitudes on a computer" is highly debatable, especially in regards to race. Sure, you're afraid of a live interviewer thinking ill of you, but anything you type online, stays online. I'll bet that some "studies have shown" the opposite of what AP asserts, depending on topic.

As to the "affect misattribution," I'm supposed to believe that whites will rate a "neutral" image negative if they saw a bunch of faces quickly on a screen just before that. Even if they "transfer their feelings" to the "neutral" image, what in the world does that prove or disprove about voters' willingness to vote for or against a black presidential candidate?

Having read through AP-Yahoo-Stanford's full report, I can't help that the pollsters decided what answer they wanted, and designed polling techniques, including exotic and bizarre ones, to get that result.

Cross-posted at BizzyBlog.com.