Sampling or not, the U.S. Census Bureau wants to get the most accurate enumeration possible for the upcoming 2000 census. But rather than rely on good old-fashioned patriotism to motivate the public to return census forms, the government is trying good old-fashioned paid advertising instead.
The $100-million campaign marks the first splurge on paid advertising for the census-in the past, the Bureau used only public-service announcements to encourage people to fill out the form. Not surprisingly, reaching minority populations is a major focus for the campaign. Adagency Young & Rubicam, which is handling the project, has tapped both divisions within its company and outside vendors to develop targeted ads for various ethnic groups. Kang & Lee, for example, will handle ads for Asian Americans, while The Bravo Group will cover the Hispanic population. According to a Bravo Group spokesperson, the agency hopes its upcoming campaign will answer the question, "What's in it for me and my community?" for Hispanics. Educating minorities about how the census impacts major decisions, such as funding for local schools and health care, may be one way to boost response rates in those communities.
Early indicators show that paid ads might help maximize response to the census. The Bureau ran ads during three dress rehearsals for Census 2000 in Columbia, South Carolina; Menominee, Wisconsin; and Sacramento. Ads appeared on television, radio, and other media sources. Results were positive. Although a direct link between reported ad exposure and likelihood of returning a census form was not found upon analysis of the dress rehearsal data, Roper Starch Worldwide concluded that advertising did have an indirect effect on the likelihood of returning a census form. If people are expecting the form in the mail, they are more likely to return it, and exposure to advertising makes people expect the form.
In all three dress rehearsal locations, non-whites still had lower levels of knowledge about the census than whites. However, Hispanics in Sacramento and non-whites in South Carolina reported "significantly higher levels of census exposure through paid advertising media compared to whites."
That's good news, because minorities are more likely to be undercounted in an actual enumeration. The overall undercount of the 1990 Census was 1.6 percent, but Asian Americans were undercounted 2.4 percent, blacks 4.4 percent, Hispanics 5 percent, and American Indians 12.2 percent. In the dress rehearsal, the estimated undercount in Sacramento for blacks was 8.4 percent and for Hispanics, 8.2 percent. Census officials note that high undercounts were expected as in most dress rehearsals because the main focus was on testing field techniques.
Concern about confidentiality of information supplied is a major obstacle to higher response rates. Barely a majority of the population believes that the Census Bureau would not let another government agency (including the FBI, IRS, and INS) see the information collected, and it's a particularly sensitive issue among Hispanics. "With Hispanics, it's critical to reassure them about confidentiality," says Jennifer Marks of the Census Bureau. "Numerous messages will be included in that campaign."