Listening to Rep. John Murtha's arguments against the American troop surge in Iraq reminds me of a scene in "Bananas."
Facing an insurgency, the Latin American dictator in that Woody Allen classic reaches out for American aid. But he mistakenly calls in not the CIA, but the UJA — the United Jewish Appeal. Black-hatted rabbis, holding little charity boxes, are soon wandering through the chaotic battle zone.
Like the dictator, Murtha is confused about what does and doesn't work during a sectarian bloodbath.
"The latest polls show that 91 percent of Sunni Iraqis and 74 percent of Shia Iraqis want the U.S. forces out of Iraq," Murtha told the Senate Foreign Relations Committee last month. "In January 2006, 47 percent of Iraqis approved of attacks on U.S.-led forces. When the same polling question was asked just 8 months later, 61 percent of Iraqis approved of attacks on U.S-led forces."
The congressman has repeated these statistics again and again in his effort to convince his colleagues to oppose the surge. Look at how many Sunnis want U.S. forces out, how many Shiites approve of killing Americans, how many Iraqis disapprove of their own leaders . . I've got the numbers right here, scientific as a Con Ed meter reading and twice as authoritative.
This is the sheerest nonsense. In today's (or yesterday's) Iraq, independent pollsters have as much chance of gathering genuine data as rabbis have of collecting donations from Iraq's Sunnis and Shiites.
The problem isn't unique to Iraq. It holds for opinion surveys in any society ruled by kinship, secrecy and fear of outsiders — in other words, almost every country in the Middle East.
In these places, it is a cardinal principle, founded in the folk wisdom of self-preservation, that you don't share honest opinions on controversial matters with inquisitive strangers. In Iraq, where Saddam Hussein's regime habituated citizens to speak in whispers even among their own families, the problem is especially acute.
For these reasons, most U.S. polling organizations don't even try to work in Iraq.
The survey Murtha quotes was published last September by the Center on Policy Attitudes, a small think tank affiliated with the University of Maryland, "partnered" by the Brookings Institution's Saban Center.
Those are prestigious names - but neither Brookings nor the University of Maryland actually did any polling. They contracted the job out to a U.S. firm, D3 Systems — which subcontracted it to KA Research. Matt Warshaw, a D3 spokesman, says KA Research is owned by Iraqis and Turks, but isn't prepared to name them.
KA Research's Web site says it uses face-to-face interviews, computer-assisted phone queries, Web and postal interviews, and focus-group discussions to come up with its statistics. That claim invites skepticism. Do strangers really go around freely in Baghdad asking for political opinions? Conduct phone surveys among people with no phones, ask the folks in Anbar Province to return candid questionnaires via a non-existent postal service, go through the neighborhoods of Najaf requesting a few minutes with the lady of the house?
Warshaw claims they do; maybe so. But KA Research, that mystery subcontractor, also does its own quality control — with no real outside checks on its data. I was unable to find (and Washaw didn't know of) any contemporaneous American polls with which to compare results.
If this company has really been able to raise a nationwide army of courageous, reliable, honest and neutral pollsters, it should be put in charge of the Iraqi security forces.
I don't know how many Iraqis really want the United States out of their country. Neither does Jack Murtha. His guess is as good as mine, but with six-month-old statistics gathered under "Bananas" conditions, that's all it is — a guess.
And that guess doesn't belong in a serious debate over national-security policy. Some arguments are too transparently flimsy even for a congressman.