Census ponders producing less granular data in next release
U.S. Census Bureau officials said Friday that they were pondering whether to produce less granular data in the next round of 2020 census data, dealing with housing and family relationships, in a decision that could upset researchers, city planners and others who rely on neighborhood-level information.
The Census Bureau described the proposal as a trade-off between producing more accurate information while protecting the privacy of participants in the nation’s head count. This year, the Census Bureau introduced a new privacy mechanism into the census data that injects controlled errors into numbers at small geographies, such as neighborhood blocks, so that people can’t be identified.
By limiting some of the next round of 2020 census information to just the level of census tracts, which are made by combining blocks and have on average 4,000 people, the Census Bureau can make the data more accurate because it won’t have to fudge the numbers at smaller levels, Jason Devine, a bureau official, told members of the Census Scientific Advisory Committee during a virtual meeting Friday.
Some outside researchers said Friday that they worried the proposal would negatively affect their research.
“The lack of census block data will inhibit our ability to analyze neighborhood change over time,” said David Van Riper, director of spatial analysis at the Minnesota Population Center at the University of Minnesota.
Under the proposal, the next round of 2020 census numbers could be released sooner because the privacy technique wouldn’t have to be applied to block-level data, Devine said.
The next round of census data will have more detailed information on race and ethnic background than in previous releases, as well as detailed figures on housing and household relationships. They aren’t expected to be released for another year, due to complications in applying the privacy technique, as well as delays caused by the pandemic.
The Census Bureau already has released two sets of 2020 census data. The first set of data was state population counts, which determined how many congressional seats each state gets, and the second set was redistricting data used for drawing congressional and legislative districts.
Some members of the scientific advisory committee — made up of statisticians, geographers and academics — expressed concerns with the proposal, noting that researchers in various fields rely on data at the smallest levels.
“This has huge and broad implications,” said Mario Marazzi, the former executive director of the Puerto Rico Institute of Statistics. “I want you guys to really think this through.”
Census Bureau officials said they welcomed feedback from researchers and other users of the census data before they make any final decision. University of Minnesota demographer Steven Ruggles, who is not on the advisory committee, said he expected significant opposition to the proposal.
“My guess is they will get a lot of pushback and won’t end up doing it,” Ruggles said. “Too many people depend on it for planning.”
Follow Mike Schneider on Twitter at https://twitter.com/MikeSchneiderAP.