Meyer and Kadane (1992) report a method for reconstructing the adjusted population (by age, race, and sex) for the half of the census blocks in Florida not made available to them. This paper studies the full adjusted data set, which is now available, to examine how well the original reconstruction was done. This is a rare opportunity to learn the exact value of quantities estimated.
The results show that the largest error made for legislative apportionment is 79 persons for one Congressional district. Thus the MK approximation could have been used instead of the unadjusted census, had the redistricting decision-makers so chosen.