Barringer, Felicity. "Making Repairs on a Faulty Census Count," New York Times, 21 April
1991, p. E7.
Here is the Times' report of the post-census survey conducted by the Bureau to evaluate the
overall 1990 count. The conclusion: "Whatever else it is good for, the broad-based post- census
survey of 165,000 households nationwide has confirmed what statisticians, politicians and simple
citizens had long suspected: the 1990 census gets low grades for accuracy. By the new
calculations, between 4 million and 6 million of the nation's 252.9 million to 254.9 million
residents were left out."
If one looks at the specific undercounts, the largest was for Hispanic men 6.2. Indeed, the figures
for minority undercounts were higher across the board than they were in 1980.
Most of these data were "by states." However, one city, the District of Columbia, was also
reported on and the undercount there was nearly 7 per cent.
The post-census survey may not be detailed enough to permit the kind of statistical correction for
which it was designed.