Privacy Concerns Weigh Down Long-Form Response RatesThe Census Bureau yesterday said people are refusing to complete and return the Census 2000 long form because of confusing messages in the media about privacy issues and the census.
In a press conference, Kenneth Prewitt, director of the Census Bureau, said final results of the 2000 census mail-back phase are in, and 66.6 percent of all households that received the short version of the form had returned it, compared with 54.1 percent of those that received the longer version.
That 12.5 percentage point gap differs markedly from the results of the 1990 census, when the gap was only 6 points. In 1990, after census enumerators visited households that did not return the forms, the gap closed to 4.5 points.
Prewitt said one reason for the low response rate may be that "a garbled message was sent that somehow [census] statistical information threatens privacy."
Indeed, in March, many Republicans, including Senate majority leader Trent Lott, R-MS, told people not to answer any long-form questions that they thought invaded their privacy. In addition, conservative members of Congress said they received hundreds of angry calls from voters who received the long forms and thought the questions were too intrusive.
Prewitt also mentioned a survey conducted for the Census Bureau which found that in the first two weeks after the forms were mailed out, 10 percent of those polled said they thought the census was too intrusive. In the fourth week, as news accounts detailed concerns of some Republicans about the questions, the number of people who said the census was too intrusive jumped to 18 percent and in the fifth week to 22 percent.
"The Census Bureau has probably the best record on privacy of any government agency. ... The privacy of the people who are filling out [the long forms] is completely protected," said Richard Barton, senior vice president of congressional relations at the Direct Marketing Association, Washington. "The data mined from it [are] totally anonymous, aggregate data."
By law, individual answers from each household are strictly confidential. The Census Bureau may not report or share with any other government agency information that can be tied to individuals or households.
The 53-question long form is sent to one in six, or approximately 20 million households. The proportion is higher in rural areas. The short form contains eight questions.
All 120 million census questionnaires were sent out in March. About 500,000 census enumerators will visit households that did not return forms by April 18. The enumerators will begin making the rounds April 27 and will collect information through July 7.
This phase of the census is called nonresponse follow-up. Enumerators will make six attempts to contact nonresponding households -- three personal visits and three phone calls.
While the short form asks basic household information -- such as whether a person owns or rents the home and the person's age and race -- the long form includes questions such as income level, what time the person leaves for work and how many bathrooms are in the house.
Data providers purchase aggregated data collected on the long form so they can offer statistical pictures of specific neighborhoods. In particular, they can track demographic estimates and the number of housing units in small areas such as census tracts and block groups -- information that is difficult to retrieve from any other source. Direct marketers, for example, often use census information because it fills in blanks in large consumer databases from companies such as infoUSA and Experian. Data from the long form are also used to develop key economic indicators, including the Consumer Price Index.
Direct marketers are worried about the low response rates and are concerned that all of the questions on the long form may not be completely answered.
"The long form mail-back rate is a concern, but I am even more worried about the completeness of the long-form responses we do get," said Ken Hodges, director of demography at Ithaca, NY-based Claritas Inc., a consumer marketing information firm that uses census data in many of its products and services. "If too many people leave a question blank, data quality would be impaired, and in a worst-case scenario, we might not get the data we need. Imagine if the census provided no new data for items such as income, education, commuting patterns and housing value?"
Hodges said, "The impact on the quality of private-sector data products is potentially severe. In the absence of long-form data, we at the private suppliers would do what we could with alternative resources, but even the best alternatives [such as consumer databases] do not come close to replacing what we are used to getting from the census. The greatest impact would be not on the suppliers but on the businesses that rely on the quality of the products we provide."
However, Hodges, like may in the direct marketing and marketing research industry, said it is too soon to tell what the results will really mean.
"I think we need to wait until everything gets in before we get worried about it," said Barton.
Michael Morreale, executive vice president at Donnelley Marketing, Omaha, NE, added, "It is too early to predict the final response rate. [But] we are optimistic that the Census Bureau-sponsored marketing campaign to promote participation, as well as their door-to-door efforts, will help remedy the anticipated decrease in responses. Until we have a better understanding of the final outcome, we are not in a position to estimate the impact on the direct marketing industry."
Bureau officials are calculating which households need to be visited to collect the long and short forms, as well as information from households that did not completely fill out the long forms. The bureau will hold a press conference April 25 to discuss its operations in more detail.
Despite the disappointing long-form numbers, Prewitt said, "Sixty-five percent of the households in America have returned their census forms. ... I believe this American achievement represents a significant reversal in the trend toward civic disengagement. We did better than we planned [61 percent] and we matched what we did in 1990."