Census 2010 in the United States is now in preparation, with 12,000 permanent employees and 1.4 million temporary workers joining the Census Bureau’s team for the 2010 Census will take place in spring 2010. The final cost of the Census for 2010 is estimated to be $15 billion, the most expensive ever; about $3 billion is allocated for information technology and laptop computers. Experts have long said the Census Bureau must do more to reduce a persistent undercount among minorities, as well as to modernize what is basically a paper mailing operation that has been in place for decades. Critics complain the Bureau has a bad reputation in terms of antiquated procedures, failing technology and incompetent bureaucrats.
President Obama has nominated Robert M. Groves, a foremost statistician and an expert on non-response, as the Census director. The basic design work for the Census is already in place.
There are at least five important issues at stake in the Census 2010:
- at least four electoral votes, and perhaps more, could shift from Democratic to Republican States for the 2012, 2016 and 2020 presidential elections
- the apportionment and gerrymandering of congressional seats within States with respect to Democratic (urban) versus Republican (rural) areas
- the allocation of federal dollars to regions of the nation based on the population counts
- the credibility of the Obama Administration in taking the count
- the need to minimize the number of people who are not counted. (The law says that illegal aliens have to be counted as well as everyone else.)
Quoting Article I of the Constitution, the Utah v. Evans court said, "The final part of the sentence says that the “actual Enumeration” shall take place “in such Manner as” Congress itself “shall by Law direct,” thereby suggesting the breadth of congressional methodological authority, rather than its limitation." The text of the constitution and the courts interpretation of that text both state clearly that the census is the responsibility of Congress. Congress has by law assigned the design of the Census to the Secretary of Commerce.
The White House has made clear that the Secretary of Commerce, will have final authority (and not the White House).
The Supreme Court later ruled in 1999 that the use of statistical sampling cannot be used to apportion House seats, but indicated that adjustments could be made to the population count when redrawing congressional boundaries.
Commerce Secretary Gary Locke has made clear under oath to the Senate that sampling will not be used for apportionment. He stated during his confirmation hearing that there are no plans to use sampling for redistricting, while indicating that sampling could be used to measure census accuracy or collect a wider range of demographic data.
Design Changes New to This Census
"The 2010 census has an innovative design, resulting in a census that will differ from its predecessor to a very substantial degree." So many changes could easily open a new loophole for manipulation, and it's important that we analyze these changes carefully. There are four major design differences between the 2000 and 2010 censuses.
First, the 2010 census will be a "Short-Form Only Census". "For the first time since 1930, there will be no "long form." The lengthier survey previously has gone to one in every six households and asked about everything from property taxes and indoor plumbing to education, ancestry and commuting patterns." "Every household in 2010 will get a shorter Census form, as required by the Constitution. This "short form" asks all members of every household their gender, age, race, ethnicity, relationship to the head of household and whether the home is owned or rented."
Second, the 2010 system will use an updated version of the MAF/TIGER system. As described by the Census Bureau, "The MAF, or Master Address File, is designed to be a complete and current list of all addresses and locations where people live or work, covering an estimated 115 million residences, as well as 60 million businesses and other structures in the U.S. The TIGER® portion of the project is a digital database that identifies the type, location and name of streets, rivers, railroads and other geographic features, and geospatially defines their relationships to each other, to the MAF addresses, and to numerous other entities." The improvements to the databases will include:
"(1) the realignment of every street and boundary in the TIGER database;
(2) development of a new MAF/TIGER processing environment and the integration of the two previously separate resources into a common technical platform;
(3) collection of global positioning system coordinates for structures on MAF;
(4) expansion of geographic partnership programs with state, local, and tribal governments, other federal agencies, the U.S. Postal Service, and the private sector;
(5) implementation of a program to use ACS enumerators to generate address updates, primarily in rural areas; and
(6) the use of periodic evaluation activities to provide quality metrics to guide corrective actions"
Third, the 2010 Census will greatly increase the number of Coverage Follow-up Interviews.
Fourth, duplicate enumerations will be removed.
Sampling and Hot-Deck Imputation
When he served in the census bureaus from 1990 to 1992, Robert M. Groves favored sampling, rather than Actual Enumeration. In 1999, Doc v. United States House of Representatives, 525 U.S. 316 (1999) ruled that sampling cannot be used in the count used for redistricting the House, but can be used in all other Census reports.
"Hot deck" imputation is a way to solve the problem that people do not return their census questionnaires. The hot-deck imputation technique replaces blanks with an estimate based on the actual response made by similar people. The Census has always used imputation--for example, by asking one person to describe everyone else in the building and not actually talking to the other people. The Hot Deck adjustment technique has been explicitly approved by the Supreme Court even though it can have an effect on the enumeration used to apportion the House. The Supreme Court has ruled it is not sampling, but a way to reduce errors in actual enumerations.
Statisticians now estimate that the 1990 Census missed about 4.7 million people who should have been counted. They estimate that the 2000 Census missed 6.4 million people, and double counted 3.1 million others, for a net undercount of 3.3 million. For example, only 64% of the people return the mailed forms, representing 200 million people. To reach the other 36%--100 million or so-- the Census tries multiple followup visits in person, or asks the neighbors. Even then many people refuse to answer the door. Many people do not want the government knowing there whereabouts, even and ignore Census promises that it will not reveal that information to any other government agency. It will cost $30,000 dollars to reach another 1000 of the missing 36%, or about $3 billion to reach all 100 million of them.
People who moved around are sometimes counted twice, as are some students and prisoners who can be counted at home and at their institutions. Children of divorced parents are often listed by both parents. Some households are so complicated or chaotic that the person filling out the forms is unclear who is living there today.
- Cork, Daniel L. et al. eds. Reengineering the 2010 census: risks and challenges (National Research Council Committee on National Statistics, 2004) online edition
- Freedman, David A. and Kenneth W. Wachter. "Methods for Census 2000 and Statistical Adjustments" (2004) online edition
- U.S. Census Bureau, online encyclopedia reference
- See AP report April 2, 2009
- See AP report April 2, 2009
- See John Stiller and Donald R. Dalzell, "Hot-deck Imputation with SAS Arrays and Macros for Large Surveys," (1997) online
- Utah v. Evans, 536 U.S. 452 (2002)online
- See Freedman and Wachter (2004) pp. 6-8