Gary Barnes' weighted-density calculations (pdf) for 31 U.S. urbanized areas are below.
Our numbers are very close for some cities. I got a perceived density of 33,000 ppsm for New York; he got 34,000 ppsm. Our numbers weren't so close for some of the others. I got 7,716 ppsm for Boston; he got 10,801.
He used 1990 U.S. Census data while I used 2000 data, which surely accounts for some of the variation. Also, as I noted last time, he used traffic zones rather than census tracts as his base unit. Traffic zones are generally smaller than census tracts, and weighted density rises as the as the geographical area is cut up into smaller units.