The house where Jessica Stephenson lives in Milwaukee bustles with activity. On weekdays, it’s filled with chattering children in the daycare center she runs out of her home.
The U.S. Census Bureau says no one lives there.
What’s going on?
The government uses data from the every-decade census to help distribute funds, draw voting districts, and allot resources. However, information from the census—like financial data—can be sensitive, the kind people don’t normally share. So census takers attempt to protect respondents’ identity and privacy.
Studies show that simply removing names or other facts doesn’t guarantee privacy. Computer programs can easily compare multiple sources, like medical records and DMV reports, to decipher someone’s identity and details.
Digging for evil, spreading strife, devising mischief—sound familiar? These are activities Proverbs warns about. (See Proverbs 16:27-28.) And they’re reasons the Census Bureau has resorted to a method called “differential privacy” (DP).
Many people believe DP helps balance data and confidentiality. It allows analysts to publish information about groups while withholding information about individuals. This can be important when collected data involves something illegal or controversial, like drug use or immigration status.
To add a layer of privacy, DP adds what analysts call “noise,” or random intentional errors, to answers. Companies like Google and Apple have long used DP to gather information about their users. The Census Bureau used DP for the first time in the 2020 census.
Critics say DP’s errors could affect drawing political districts and dispensing funds. At least one analysis suggests DP could harm minority communities by undercounting racially mixed areas.
DP has produced some wacky and downright false results. For example, the official 2020 census results say 54 people live in Stephenson’s Milwaukee census block—but also that there are no occupied homes. In Florida, the 2020 census lists 15,000 neighborhood blocks as having 200,000 residents—but no occupied homes.
Similar unreliable data has created headaches for community planners. Minnesota demographer Eric Guthrie says city managers worry about state and federal funding. “I explain to them . . . that it’s not an error in the traditional sense,” he says. “The bug is there by design.”
The Census Bureau insists the data is as accurate as in past censuses. It maintains that the low-level errors don’t present a large problem.
But Acting Census Bureau director Ron Jarmin does warn that DP could produce some “fuzzy” figures at the neighborhood block level. He urges data users to combine blocks for accurate results.
Demographer Rich Doty agrees. “We were warned about [errors],” he says. “We just didn’t expect this many.”
Why? In a world plagued by sin, humans contrive methods to achieve goals in principle, even when the methods sometimes conflict with those principles. Christians should know what those methods are in order to be equipped to address the consequences.