Barringer, Felicity. "The Census, in One Not-So-Easy Lesson," New York Times, 22 April
The U.S. is officially only several weeks into the process of counting the population of the
country. Some problems have emerged already: fewer households have returned the
questionnaire than expected. Something has gone wrong as the rate of return is significantly
lower than the return rate of 1980. One household in three has not returned its report. The Census
Bureau will overspend its budget by a large amount trying to compensate for the underreporting:
by hiring more ennumerators to follow-up on the questionnaire's non-respondants.
The extra money is only a part of the problem. There are some politically powerful groups who
feel they have been undercounted. The census will be under intense scrutiny this year as courts
are asked to decide certain questions about the undercount. The idea that the Bureau can can
"everyone" has got to be abandoned and "adjustments" made for the biased undercounts. And
how the "statistics" will be manipulated is open in interpretation. One can use certain
assumptions and "correct" the count but those assumptions are questionable. There are
hard-to-count people: those who do not wish to be counted (illegal immigrants), who are
"invisible" (the homeless), who slip through the net (who have recently moved), and some of
these hard-to-count people are growing at a very rapid rate.
Stastisticians can suggest methods of improving the error term or reducing it to some level but it
seems that statisticians will not dispose of the problem. There are powerful political forces
operating in the counting process and the "science" is enmeshed in politics.