Iraqis to U.S. soldiers:"Please don't go! We want to kill you!"Or why pollsters waste our time

In my opinion, public opinion polls are useless--kind of the mentally challenged younger sibling of the noble project of one-man-one-vote democracy. But when it comes to grinning idiocy, the antique media just can't get enough, so today we're going to hear a lot of nonsense about how many Iraqis feel nervous and how many don't, about how many can think of a family member hurt since the invasion and how many can't. Public opinion poll results are only slightly less malleable than Play-Doh. Reuters mentioned that the latest poll "oversampled" in Anbar, Sadr City, Basra, and Kirkuk. Oversampling is a bit of statistical funny business that seems hard to justify in this case, especially since quizzing residents of Anbar or Sadr City about the occupation is like stopping passersby in Berkeley and asking how they feel about Bush. Pollsters normally oversample to compensate for under-representation of minorities in random polls, and even in those cases I have my doubts over whether oversampling increases the poll's "accuracy" or whether it simply opens the door for manipulation by agenda-driven poll workers. Oversampling may have been partly to blame for the inaccurate exit polls that helped blow Kerry's chances in 2000.

As if to demonstrate the utter meaninglessness of opinion polls, this latest survey uncovered a startling bit of Iraqi cognitive dissonance.

Slightly more than half of Iraqis — 51 percent — now say that violence against U.S. forces is acceptable ...

About four in five Iraqis oppose the presence of U.S. troops but only a third want those U.S. troops to leave Iraq immediately.

That means that at least 18 percent of Iraqis think it's okay to shoot our soldiers and think our soldiers should stick around.

Now obviously this makes no sense, and only some pretty tricky (or pretty sloppy) polling could render results like that. The suspect figure is the 51 percent, which the pollsters say is a 300 percent jump from 2004. I'd like to know how the term "acceptable" was translated--whether a "yes" answer meant "I'd let my teenage son shoot at Americans" or something more akin to "shit happens"--a popular sentiment, incidentally, in Dar al Islam. I'd also like to know that the question was phrased the same way two years ago. But for some reason, mainstream media outlets never provide a full explanation of a poll's questionnaire and methodology. In this case, a little digging (about seven mouse clicks away) reveals that the poll workers' "oversampling" placed them in harm's way a number of times and subjected them to repeated questioning and harassment by Iraqi police. They were stopped from conducting interviews with female family members even when the poll protocol required it. And most importantly, they did not have the trust of the people they were questioning, a fact that pretty much invalidates the entire project.

The main reason behind this problem is the unstable security situation in this area, especially if we realize that the business of private research is new for Iraqis who are not familiar with such things. Many people believe that we are doing something against them and that we work for foreign interests.

If the interviewees cannot be certain they're not speaking with insurgent spies trying to root out coalition collaborators, they're not likely to give an honest opinion about the whole G.I. vs. jihadi popularity contest. And if the AP had more integrity, they would include these flaws in their own report on the poll, not bury them in hyperlinks or pass the buck to ABC (who sponsored the poll).

But we can't expect such integrity because it doesn't serve the mainstream media to reveal to us their biases, their agendas, their limitations or their manipulations. Like fast food, if we know too much about it, we just might stop swallowing it.