A census supervisor in Alabama told workers to fake household data in order to check off as many houses as possible, according to text messages obtained by The Associated Press.
The supervisor allegedly wanted census workers to finalize cases without interviewing households. According to the messages, if two attempts to interview a member of the household failed, and two unsuccessful attempts were made to interview the landlord or neighbors, then the census workers were to put down that only one person lived at that residence.
“You are to clear the case indicating occupied by 1,” a text from a census supervisor in Dothan, Ala., read, the AP reported.
An enumerator who traveled to Alabama from Florida to help it catch up with the census count shared the texts with the AP and said she refused to follow the orders, believing she would be falsifying data if she did. The enumerator requested to remain anonymous and did not provide the full name of the supervisor who sent out the instructions.
In photos sent out to census workers, the supervisor allegedly shared handwritten 15-step instructions on how to mark a residence as having only one person, advising them to do so two to three hours after visiting the address in order to avoid suspicion from higher-ups who can track census workers through their bureau-issued iPhones.
Michael Cook, a spokesman for the Census Bureau, said the agency was investigating the Alabama case, saying, “We take falsification allegations very seriously.” According to Cook, data problems can be addressed by revisiting households and improving accuracy.
The 2020 census was allowed to end about two weeks early this year after the Supreme Court allowed the Trump administration to do so. According to the AP, dozens of census workers have contacted the news agency to share accounts of being unduly rushed in order to meet the deadline.
In October, a lawsuit was filed by advocacy groups and local governments that alleged census workers were pressured to falsify data in order to claim they had reached 99.9 percent of households.
Alternative methods of counting, including using IRS records, interviewing landlords and neighbors and getting headcounts without gender, age and racial data are all believed to have been used in order to fill in information gaps.