At 8.25am on 12 Jan Instapundit cited Ed Morrissey’s “fact check” of Nicolas Kristof’s 12 Jan (11 Jan online) column on infant mortality in the US. Morrissey finds Kristof guilty of “obscene” errors. (Morrissey’s post is here. Kristof’s original column is here.)
Morrissey takes issue with Kristof’s comparison of infant mortality rates in the US with those in China and Cuba. For example, 4.5 per thousand in Beijing against 6.5 in New York. Morrissey writes:
Kristof misleads his readers with that comparison. Using the CIA Factbook entry for China -- a source that Kristof uses for Cuba -- we find out that China has an astronomical infant-mortality rate of 25.28, a rate that has not been seen in the US since the 1960s. Perhaps the rate is better in Beijing, but it hardly matters if the rest of the country has that rate. It qualifies as cherry-picking of the worst order on Kristof's part.
Morrissey may have a case. I must say I wondered whether Kristof was right to rely on the Cuban and Chinese statistics. Totalitarian and would-be totalitarian governments don’t necessarily have a good track record in this regard. (Then again, how far should one rely on the CIA Factbook?)
But Morrissey does not address all of Kristof’s argument, and by not doing so may be indulging in some cherry picking himself.
Kristof also compares the US and other rich industrial nations, where records are reliable. And in this comparison Kristof’s case looks strong. Sweden, Japan and Iceland, for example, all have an infant mortality rate less than half of that in the US. Women are 70 percent more likely to die in childbirth in America than in Europe.
It is not only progressives, self-defined, who can be led astray or misuse by statistics from totalitarian regimes. It looks as if advocates of what become the 2003 invasion of Iraq were as keen as Saddam himself to believe the misleading, grossly exaggerated information fed to him by his own subjects about programmes for chemical and biological weapons.
My estimation of Kristof remains generally high.