Transmittal LetterSeptember 30, 1999
The Hon. Albert Gore
United States Senate
Washington, DC 20510
The Hon. J. Dennis Hastert
United States House of Representatives
Washington, DC 20515
Dear Mr. President and Mr. Speaker:
As Congressional Members of the U.S. Census Monitoring Board, charged with overseeing the Census Bureau’s preparations for the 2000 Decennial Census, and particularly to review the degree to which those preparations “shall achieve maximum possible accuracy at every level of geography,” we write to alert you to a number of serious concerns over the proposed statistical adjustment raised by a new study of Census Bureau data.
This report, the third in a series required by our statute, shows that statistical adjustment, heralded as a kind of statistical remedy, will fail in its main charge: to prevent traditionally undercounted communities from receiving less than their fair share of representation and funding. In fact, the claimed remedy will do little if anything to correct the severe undercounts that often affect predominantly minority neighborhoods.
According to our extensive evaluation of 1990 Census data – only recently made available to the public – heavily undercounted areas will remain heavily undercounted despite statistical adjustment, and overcounts in many areas will actually be increased.
The purpose of this report is to warn that the benefits of statistical adjustment have been overstated. Our concern is that statistical adjustment will fall short of its goal – correcting severe undercounts in traditionally undercounted areas. Therefore, we believe it is imperative that Members of Congress, state and local officials and community leaders understand the inadequacies of adjustment and take appropriate steps locally to ensure their constituents are counted.
Relying on statistical adjustment to fix the problem of undercounts would be a tragic mistake that could have serious repercussions for the people living in severely undercounted areas. A critical failing is that statistical adjustment conveys the illusion of correcting the undercount, when in fact actual people’s needs are not met.
This report provides an explanation of the statistical problems with adjustment revealed by our study, and the methodology we employed.
Clearly, the only certain way for any neighborhood or local area to overcome the differential undercount is to get the best initial count during the census. As a result, we also include a series of steps that Members of Congress and local leaders, who are likely to face undercount problems in spite of statistical adjustment, can take to help ensure that their communities are counted properly.
Without the active involvement of informed local leaders in the census process, traditionally undercounted communities will once again be left with empty promises and unrealistic expectations.
All of us want the same outcome – the fairest and most accurate census possible, including a dramatic decrease in the differential undercount of African American, Latino, Asian and American Indian populations. No one denies the census will miss people. In a nation as large, diverse and mobile as ours, that is unavoidable. However, we believe it is unacceptable for the 2000 Census to systematically miss the same people in the same communities that depend so heavily on the census for a fair share of crucial services and political representation.
J. Kenneth Blackwell
Dr.David W. Murray
A. Mark Neuman