UW News

April 8, 2010

Celebrating the census: Population numbers a powerful research tool for demographers

News and Information

Last Thursday, the UW staged what was probably the largest celebration of Census Day in the state. April 1 is designated as the day for the official count.

“Some people may regard a census party as a strange, geeky event,” said Shelly Lundberg, director of the Center for Studies in Demography and Ecology (CSDE), which hosted the party, complete with cake and refreshments.

Lundberg pointed out that, although most center affiliates don’t actually analyze census data, virtually everyone affiliated with the center uses data from human populations through administrative records, surveys, unstructured interviews, observation — “and we all care that our data are accurate, comprehensive, and representative of the population we are studying. We are aware of the hazards of basing our conclusions on samples of the easily-sampled, on the behavior of volunteers. In that sense, the attempt to conduct a census of the entire population and the efforts to make it as complete as is humanly possible are emblematic of our mission as demographers. It is that ideal of the census that we are celebrating.”

Linda Clark with the regional census center talked briefly about how the Census Bureau attempts to count people in a mobile society. It’s relatively easy for people who live in just one residence all the time to complete a form, she said. But that doesn’t describe all of America. People may be away from home at the time of the census. Some may be living with people to whom they are not related. Some split where they live among multiple residences. There are students in residence halls and people living in group facilities, including prisons. There are military personnel in barracks and deployed overseas. There are people living in campgrounds, on boats, in motels. There are even people living in housing at racetracks or in circuses. There are migrant workers.

And there are the homeless: living in emergency or transitional shelters, eating at soup kitchens, and sleeping outdoors, in cars, in doorways and abandoned buildings.

The Census Bureau has rules for all these circumstances and intends to count every person that it can (Clark said that counting all homeless individuals presents formidable challenges and that they are inevitably undercounted). Not only does the bureau send out the ubiquitous short forms, but it conducts extensive outreach to people who at first blush may not wish to participate:

  • First, the bureau tries to woo participants with the promise of three community benefits — power, justice and money.
  • Second, it tries to reassure people that all information is confidential under federal law and that these laws have been scrupulously observed for over 60 years.
  • Third, the bureau goes to great lengths to overcome any language barriers; it employs outreach staff who collectively can communicate in 101 languages and as a last resort can use flash cards with pictures to get the questions asked and answered.
  • Finally, to overcome fears that information might be shared with law enforcement officials or other federal agencies, the bureau seeks endorsement of the census from trusted clergy and other community leaders, as well as employing outreach staff who are familiar with the language, culture and customs in which they work as enumerators.

Charlie Hirschman, professor of sociology and CSDE affiliate, said “my favorite years end in zero. One of the highlights of my life was in 1980 when I had to fill out the Long Form.” He noted that in recent years the most controversial item on the form has been one that depends on “who you think you are” for the answer: your race or ethnicity.

Beginning with the 2000 census, the form has allowed participants to check one or more boxes that in combination might accurately describe them. The result has been 63 possible combinations, potentially a demographer’s nightmare. Hirschman, who studies identity formation, said in the 2000 census just 2.4 percent of those completing the form identified themselves as multiracial. “Some believe that this will actually be the high-water mark, but I think it will go up more this time.”

Marieka Klawitter, an associate professor in the Evans School and a CSDE affiliate, started using census data in the mid-90s for her research on same-sex couples, and specifically on the effects of state and local antidiscrimination policies on earnings. Prior to 1990 no information on this group was gathered, and even the initial effort was incomplete — how else to explain an increase of 300 percent between 1990 and 2000?

Stewart Tolnay, professor of sociology, uses historical census records, because by law he can look at individual forms that are more than 72 years old. One of his studies involved creating an inventory of lynchings that occurred in the South between 1880 and 1930. Some of the research that interests him will be impossible to do in the future because the census form now gathers much less information than a century ago. Earlier questionnaires asked about how people lived, where they worked, their family situation, military service, and disabilities — but no longer. “The work I do won’t be passing on to future generations of researchers,” he said.

While individual census records are protected by strict privacy laws, it is possible to gain access to them for research, said Mark Ellis, professor of geography, who has used census data for 30 years. He goes to the UCLA data center for his research on mixed couples, which requires finer slices of data than is available at the census tract level. Researchers are required to understand, and must swear to obey, the privacy laws and not reveal individual information to anyone. Researchers who make requests for access to what is called “microdata” must undergo extensive background reviews and are subject to to stringent oversight.

“I’m working to have a research data center here in Seattle,” he said. “It would give researchers convenient access to census data as well as other economic information that is valuable for research.” The idea of creating a research data center on campus has received broad support and has been included in the UW’s Fiscal Year 2011 federal agenda that was submitted recently to the Washington congressional delegation.