The folks at NG are attempting to prove that 'progressive' cities -- large metropolitan areas marked by liberal politics and advanced municipally-funded city planning -- are not racially diverse, as the hypocrite lefties would have you believe, but rather nothing more than modern examples of white flight, perfected. See? For all their big talk about multiculturalism, racial diversity, and planned urbanity, the liberals live in bastions of whiteness, just like conservatives!
It's all very salutary, until you realize that their entire methodology is cooked like a Denny's Grand Slam breakfast. First of all, as Roy Edroso points out, author Aaron M. Renn insists that, for the purposes of his study, New York, Los Angeles, Chicago, Philadelphia, San Francisco and Washington must be excluded. Not to put too fine a point on it, but this tends to rather seriously undercut any argument that liberal urban centers are not ethnically diverse. It's like arguing that if you exclude Russia, Germany, France, England, Spain and Italy, there are no large countries in Europe.
But even beyond the chicanery Roy cites, the study is three-dollar-bill bogus. Renn excludes Asians altogether, and cites Hispanics only when their numbers are low, excluding them when the numbers are high. He does the same with historical patterns of black migration, making note of them when they seem to strengthen his argument and ignoring them when they don't. Perhaps most egregious of all, his population figures come not from urban statistics, but from county numbers. Everyone who's studied the issue at all knows that suburbs and county populations are always predominantly white; this is how, for example, Obama carried the four largest cities in Texas (Dallas, Austin, Houston, and San Antonio) while the state still went to McCain. If his argument wasn't already bogus to begin with, using the county date to prove an area's 'whiteness' sinks it like a stone.
Being wrong: the conservative national pastime.