Two Sets of Census Numbers

June 16, 2000

The Census Bureau will produce two sets of numbers for the 2000 Census. The first are the normal numbers based upon the traditional head count. The second set are an adjusted set based on a huge post-census sample survey.

Now remember that the Supreme Court has already ruled that the adjusted numbers cannot be used to apportion congressional districts. But that hasn't stopped proponents of what could be called census sampling. And there seems to be some indication that the adjusted numbers will be used for redistricting and federal grant allocations.

Proponents of sampling argue that large numbers of minorities are undercounted using the traditional method mandated by the Constitution. They further argue that this results in a loss of federal funds. Well, we'll see when these two sets of numbers are published and then analyzed.

In fiscal year 1998, the General Accounting Office studied the allocation of $185 billion that relied on census figures and found that only one-third of one percent would have been distributed differently as a result of adjusting the 1990 census. The reason for the small difference shows population is only one factor in the allocation of federal funds.

There is now some evidence that adjusting census figures would not necessarily benefit all undercounted minorities. The 1990 post-census survey found proportionally more Hispanics in the barrios of the Southwest than blacks in the cities of the Northeast and Midwest.

By the way, the complexity of the sampling survey can also lead to errors. A post-census survey in 1990 added one million people to an undercount estimate, and the error went undiscovered for more than a year.

The Constitution authorizes the federal government to conduct a census every ten years to "enumerate" the number of citizens. It does not allow for sampling or adjusted figures. Just remember that when the Census Bureau posts its two sets of numbers.

I'm Kerby Anderson of Probe Ministries, and that's my opinion.