By the end of last year’s U.S. population, the Census Bureau did not have data on nearly one-fifth of U.S. occupied college dormitories, nursing homes, and prisons, so statistical agencies were institutionalized for collection. I had to make an 11 hour phone call. Provide information or use last resort statistical techniques to fill the gap.
43,000 of the 227,000 occupied dormitories, prisons, barracks, homeless shelters, group homes and nursing homes count until December, according to new documents and slide presentations recently released by the Census Office in the Information Disclosure Law proceedings. I wasn’t able to do it. According to the Republican constituency change advocacy group.
The document suggests a range of challenges faced by the Bureau in conducting large-scale counts in the midst of a pandemic. This effort was made even more difficult by wildfires, hurricanes, and attempts by the Trump administration to interfere with the census.
Pandemics have forced universities to close dormitories and send students home, and nursing homes and other facilities protect vulnerable residents from the virus.
Bureau officials are confident that they have filled the gap using statistical methods that they consider reliable, but they admit that the challenge was formidable.
Census official Barbara Lo Presti recently said that data collected from the group’s quarters accounts for most of the irregularities encountered by statistical agencies, but data processing “cannot be corrected by data collection.” Does not indicate a serious error. “
“Processing anomalies are not errors, but they can be errors if not evaluated and corrected,” LoPresti told a virtual conference of external experts assessing the quality of 2020 census data. rice field. “Therefore, our quality (checking) process was working.”
However, correcting the irregularity forced the Census Bureau to postpone the release of numbers used to divide parliamentary seats between states in a process called allocation. It also delayed the release of constituency change data used to redraw parliament and legislative districts by five months.
Although people living in group districts make up a small proportion of the total population (less than 3% of the 331 million people living in the United States), inaccurate information is provided to the populous areas and military of university towns and prisons. Base that can have a big impact. As a result, parliamentary representatives and the amount of federal funding they are eligible to receive may be reduced.
“In some regions, individual groups can have huge quarters,” Connie Citro, a senior scholar at the National Statistics Committee, said in a virtual meeting of outside experts.
The Republican advocacy group, the Fair Lines America Foundation, has appealed for information about how the Census Bureau’s quarterly counts were done. The allocation number will be announced by the Census Bureau in April, and the subdivision number used to depict Congress and the legislative district will be released next month.
Dormitories, elderly housings, and prisons while the Census Bureau is calculating numbers. This method has been used for some time to fill in the missing information about individual households.
“If the Census Bureau is allowed to make this kind of methodology change and implementation in a closed room … Election turmoil depends on the number of potentially flawed states in repartitioning. It could be due to being there, “Fairlines said in a court document.
In addition to the 43,000 group quarter addresses that were lacking data last December, another 3,500 addresses stated that the number of people was zero or too high, suggesting duplication. There was no count. Statisticians have removed duplicates such as college students counted in both dormitories and parents’ homes, the document said.
In the absence of information about dormitories, nursing homes, and prison residents, Census Bureau statisticians apply information they already know about the facility from either previous investigations, previous contacts, or administrative records. The count has been reached.
After substitution and deduplication, the revised numbers seemed to artificially inflate the group quarter count by 444,000 people. Almost 8.6 million people lived in place of the 8.1 million people expected to live in the group district. Group quarters of the revised data were significantly higher in California, New York, Florida, and Washington, the document and slide presentations show.
The Census Bureau said in a statement that the numbers in the document were not final numbers and that the 444,000 differences were addressed in a later calculation of the numbers. Statistical agencies did not state what the final numbers were or provide details on how the differences were handled.
“The Census Bureau has made some improvements to its methodology since the day these slides were created,” the statement said.
Follow Mike Schneider on Twitter. ttps: //twitter.com/MikeSchneiderAP