"Helping You Make Informed Decisions." That Census Bureau motto was on prominent display throughout "America's Scorecard: The Historic Role of the Census in an Ever-Changing Nation," the March 4 and 5 conference sponsored jointly by the Bureau and the Division of U.S. Studies. The external use of Census data to make informed decisions was one of the topics addressed; others concerned the Bureau's internal decisions about the kinds of information it collects. As former Census director Kenneth Prewitt (Columbia University) noted, the Bureau's prime responsibility is to maintain data quality. That goal, he said, must illuminate the controversies about what data to gather and how and with whom to share it, and the controversies were indeed thrashed out during the two days of the conference.

Privacy and Confidentiality Conference participants agreed that current discussions about the uses of Census data must be viewed in the light of past experience. During World Wars I and II, for example, there were significant infringements on the confidentiality of the data. In 1917, as Margo Anderson (University of Wisconsin-Milwaukee) noted, the Census Bureau was instructed to give its age and sex information to the draft board, to ensure that all men of age had registered; in 1942, the War Powers Act required the sharing of Census data that constituted "vital war information." In both cases, she and commentator Eleanor Singer (University of Michigan) remarked, there were competing interests at stake, and after the national security threats passed the confidentiality protections that had been suspended were restored. Both cautioned, however, that infringements on confidentiality are more likely to reemerge in eras where there are perceived threats to national security—and that information privacy, once violated, is always difficult to reinstate. Ralph Rector (Heritage Foundation), commenting that other governmental agencies do not have the same statutory requirement of maintaining informational privacy, argued that ensuring information security is necessary in order to keep the American people willing to cooperate with the Census and thus uphold standards of data completeness and accuracy.

At the same time that some speakers asked for increasingly detailed information, others emphasized the need for privacy and confidentiality. Ramon Barquin (Barquin International) described privacy as reflecting an individual's desire not to share personal information, while confidentiality pertains to safeguards against the sharing of that personal information by the Census Bureau. Dr. Prewitt charged the Bureau with conflating the two: attempting to address privacy concerns by enhancing confidentiality protections. He insisted that the Bureau must remain aware that there is a point at which questions become so probing that respondents become defensive and unwilling to provide any information at all. He thereby took issue with Kevin Hassett (American Enterprise Institute) and Robert McGuckin (The Conference Board), who encouraged the Bureau to ask as many questions and to gather as much microdata as possible. Hassett and McGuckin focused on the importance of information to future researchers, while Prewitt argued that each additionally probing request for information increases the possibility of a respondent's opting out of the Census entirely. Even if probing questions prompt only "item non-response," Prewitt reminded the audience that non-responses can make all responses to a particular question potentially useless.

Stephen Ruggles (Minnesota Population Center) subdivided confidentiality concerns into the fear of "additional authorization" that will permit other federal agencies to gain access to Census Bureau data for purposes beyond information gathering and the fear of "unauthorized access" by data snoopers. Barry Steinhardt (American Civil Liberties Union) stressed that unauthorized access is less of a problem because the Census computer systems can easily be improved to guard against snoopers. "Additional authorization," however, should be particularly worrisome to and should be taken seriously by the American people, as government agencies often demand access to information already gathered by other government entities for entirely different purposes.

Non-Apportionment Use of Census Data Speaking of the need for Census data at the local level, Jackie Byers (National Association of Counties) described the way county officials rely on Census figures when assessing, e.g., the number of schools to build; the amount of trash that will have to be collected or the amount of water localities will require; the type, quantity, and routes of transportation services needed to get residents to and from work; and the number and language of ESL classes to staff. Byers stressed that Census data is irreplaceable for the 3066 United States counties seeking information about themselves as well as comparisons with other similar counties across the country. Bruce Cohen (Massachusetts Department of Public Health) similarly emphasized the importance of nationally standardized data in assessing trends in public health, pointing out as well the need for information about ethnic subgroups, based on country of origin, rather than data that amalgamates racial categories. He urged the Bureau to expand and release increasingly specific information about the American people in order to enable researchers to think about national trends along as many dimensions as possible.

Demographics and Racial Data Racial categorization has always been part of the Census. As Melissa Nobles (MIT) noted, the only categories that were constitutionally required for the first Census' primary purpose, apportionment of seats in the House of Representatives, were free people (counted as whole persons) and slaves (counted as 3/5 of a person). The first census-takers nonetheless inadvertently reflected the historic importance of race to the American enterprise by identifying respondents by race. In 1850, responding to the desire of scientists who wished to "prove" that racial mixing would weaken the country and who wanted documentation showing that mulattos were less fertile and lived fewer years than "pure" whites or blacks, the category of "mulatto" was added. At least one religion came close to being inserted as a racial category in the early 1900s, according to Joel Perlmann (Bard College): tracking the practice of immigration officials, Census officials proposed to define "Jewish" as a race, and were prevented from doing so only by fierce lobbying by Jewish elites.

Roderick Harrison (Joint Center for Political & Economic Studies) pointed out that the enactment of civil rights legislation in the 1960s meant that once-deplored racial categories, designed to maintain segregation and discrimination, have become a mechanism for collecting information used to rectify the effects of past and current discrimination—but at a time when the black/white categories have been called into question by the move toward multiethnic self-identification. Nathaniel Persily (University of Pennsylvania) discussed the 128 possible racial and ethnic combinations that could be checked off on the 2000 Census long form and noted that they reflected deference to the demand for self-identification, but that a desire not to permit the new categories to dilute the count of African-Americans—the original focus of the 1965 Voting Rights Act—led the Bureau to adopt what amounts to an updated version of the much-hated "one drop" rule: people checking off both "African American" and "white" on their forms are counted as African American. Many of the speakers in various panels emphasized, similarly, that the kinds of information gathered by the Census and the ways in which they are used are driven by political forces and cannot be understood except in light of the politics of specific eras.

Greg Rodriguez (New America Foundation) spoke about the fluidity of racial self-classifications in response to census questions. By comparing individuals' answers to racial questions in the Los Angeles censuses taken by the Spanish in 1781 and 1790, researchers have documented "trading up" in self-reported whiteness across the years. This phenomenon was commented upon by Census director Charles Louis Kincannon, who described the so-called "ancestry question" as eliciting information more similar to an "opinion poll" than to factual data and who said he has no intention of shifting it from the selectively administered long form to the universally distributed short form. David Kennedy (Stanford University) went one step further, suggesting that current American exogamy rates will make it impossible to discern a racial frontier in the foreseeable future. This degree of racial mingling, he argued, will obviate the need for a Census question regarding racial background, as there will be so much overlap in categories that the inquiry will produce no useful data.

Limitations of Census Categories One of the most fundamental critiques of the ethnicity, ancestry, and nation-of-origin questions included in the Census 2000 short form was the extent to which they fell short of the specificity many individuals would have preferred. While some speakers emphasized the need for the government to stay out of the business of validating individuals' desires for racial or ethnic identification, others stressed the importance of the Census' permitting the expression of individual choice. There was general agreement, however, that the Census categories affect research and assumptions about the nation that are made on the basis of its data. As Kennedy pointed out, categories that are not recorded on the Census and for which information is not compiled functionally do not exist. Frederick Jackson Turner's "frontier thesis," for example, arguing that the closing of the frontier in the early 1900s resulted in a shift in American consciousness, was based almost entirely on population density figures in the 1890 Census. Thus, Kennedy reminded Census officials, scholars, and critics, the questions that Americans are asked to answer ever ten years have currency far beyond apportionment and resource allocation: information from the Census has real and significant effects on how Americans see themselves and conceive of a shared national identity, as well as on the public policy decisions subsequently taken by their elected officials.

Drafted by Ann Chernicoff

Philippa Strum, Director of U.S. Studies (202) 691-4129