Iraq and conflicting polls-Edit: problems with images fixed

All polls in Iraq should be treated with a certain amount of skepticism, but neither should they be ignored, that is unless they hopelessly skew themselves from the get go, whereupon bit of interpretation will be needed. Amongst a number of interesting nuggets, Keith recently posted the results of a Times of London poll:

“49% of those questioned preferred life under Nouri al-Maliki, the prime minister, to living under Saddam. Only 26% said things had been better in Saddam’s era, while 16% said the two leaders were as bad as each other and the rest did not know or refused to answer.”

Meanwhile another poll was released which claimed:

And just 38% said the situation in the country was better than before the 2003 war, while 50% said it was worse.

Neither result surprises me, and both show that the invasion has not been in the eyes of all Iraqis an unmitigated disaster. Still, quite a difference. So which is closer to the truth? Mostly it seems it is due to one factor, from Dave Price:

The new poll is thirty-five percent Sunni Arabs (p37), who benefitted from Saddam’s regime and are generally much angrier and more pessimistic about the invasion and its results than the other groups. That’s right, thirty-five percent in a country where most estimates of Sunni Arab population are around 15% (note that the Sunni number given by the CIA Factbook includes Kurds, who are more than half the total Sunnis in Iraq). They’re oversampling Sunni Arabs by roughly two to one.

Unless I’m missing something here, somebody has some serious explaining to do.

Allah goes ahead and does some of that explaining for Dave:

The data comes from the new ABC/BBC poll of Iraqis published today. I’m going to highlight some of the key results here, but I recommend flipping through it all. One thing these questions won’t show is how bad the trend is since November 2005 — and how good the trend was then since 2004. The numbers were up across the board; then came the Samarra bombing last February and everything lunged. If we’d been able somehow to stop those rat bastards before they did it, the country might look different now.

Anyway, let’s start with the sample. What’s wrong with this picture?

It oversamples Sunni Arabs. Wildly. The weekend poll oversampled them too, but only a bit. This one more than doubles their actual strength in the country, which is only 10 to 15% according to best estimates. The numbers for the Kurds are about right, so this is all coming at the expense of Shiites. Bear that in mind as we go forward, because as you’ll see, the difference between Sunni and Shiite opinion on some key questions can reach 90%.

Here’s the big one.

Considering that Shiite Arabs and Sunni Kurds make up 85-90% of the population, that’s not a bad result at all. But look at the net totals, driven down by the unbelievably sour attitudes of Sunnis and the poll’s oversampling. 85% of the country approves of the invasion (to a greater or lesser extent) by 70+%, and yet somehow we’ve ended up with a clear majority who are opposed. Same thing here:

Support in excess of 68% for the Iraqi government among a super-supermajority of the population, and yet the net is negative.

Go ahead and read the rest of his analysis of the poll. There is a lot of good stuff, though not all of it is encouraging. There is good deal of evidence in the poll which suggests widespread support for the US leaving. At least one could come to that conclusion if one ignored this question:

Wrong:

Even now, the most support for an immediate pullout that the Shiites can muster — even with 60-65% of the population and control of the government and the Iraqi Army — is 28%. How that squares with the results of Questions 25 and (in particular) 28, I have no idea.

Reading the report in light of the oversampling is a portrait of twisting the news. Allah gives a nice analysis to compare it with.

[tags] Iraq, war, Sunni, Shiite, Kurds, polls, ABC [/tags]