In my Sep 4 post, I said something about using PUMS data to address the question of whether the "creative" people gravitate to "high density" locations. This is a favorite theme of urban economists and many others.
The PUMS data are attractive because they report the education level of the migrants and many of the PUMS areas (PUMAs) are smaller than metro areas or counties. Average densities computed for large areas can mislead.
But colleague Sung-ho Ryu corrected me by noting that the sending PUMAs are spatial aggregations, so stick with the large sample (N=2077) of receiving PUMAs. I did say that I was late to the party.
So here we go again. In the 2006 survey, 83 percent did not report having moved, 3.5 percent moved within the same PUMA, and 13 percent moved to another PUMA. For the latter group, the correlation between the number of arrivals and the population density of the receiving area was 0.07, but for movers with a bachelor's degree, the correlation was 0.16, and for movers with an advanced degree, the correlation was 0.21. I'll stick to my story. Just population density (even for sub-metropolitan units) tells us little.