WASHINGTON – The troubled 2020 census won a cautious vote of confidence on Tuesday from experts who said the tally seemed accurate enough for its overarching constitutional purpose: to reallocate the 435 seats in the House of Representatives.
But the experts, prominent statisticians with access to the interior of the census process, limited their findings to the aggregate national count and counts in all 50 states and the District of Columbia. Many more studies will be needed to assess the reliability of local population totals and characteristics such as race and ethnicity that are essential components of every census, they said.
The results were presented in a 59-page analysis of the 2020 tally commissioned and reviewed by the American Statistical Association. A task force set up by the group recommended the survey in October – and the Census Bureau agreed – amid growing concerns about the accuracy of the count among businesses, state and local governments and others who use the census to plan for the future.
In all respects, the census has encountered unprecedented hurdles: The coronavirus pandemic halted much of the tally just as it began in April 2020, forcing the bureau to extend its work by almost two months. Last fall, wildfires in the west and coastal hurricanes disrupted the work of the Census Bureau just as door knockers unfolded to survey millions of households that had not filled out forms.
On top of that, the Trump administration has brought forward the deadline to finish counting, throwing a political key into an already hampered count. The cascade of problems has led many experts, including some senior Census Bureau officials, to fear that the final tally is so flawed as to be unusable.
The report released on Tuesday hinted, at least, that those fears were overblown. Despite being worded in the cautionary language of scientists, he nonetheless said there was no convincing evidence that the national count or individual state counts were less reliable than the results of the 2010 census. , widely hailed as one of the most precise ever.
“Career staff took this census under what could be the most difficult conditions ever, and they did their best,” said John H. Thompson, former director of the Census Bureau who was part of the group. work of the statistical association which oversaw the review.
The analysis looked at 10 critical operations in the 2020 enumeration, such as compiling a national list of addresses to be surveyed and establishing statistical estimates on people living in households who refused to complete the forms. census.
The study found wide variations between states in the results of these operations. For example, one in 20 households in Louisiana was counted not by responses to written surveys, but by data in federal records that showed who lived there. In Hawaii, the corresponding figure was one in 60.
Logic says Hawaii’s tally, with more personal answers, would be more trustworthy than Louisiana’s. But the report cautions against rash judgments: Using more records to count households, for example, may have reduced the need to make even less precise statistical assumptions about them.
These statistical assumptions, called imputations, were among the most watched and controversial aspects of the 2020 census. In the previous census in 2010, 0.4% of households were counted by imputation. Many experts were anticipating a sharp increase in that share – and a less accurate census – as the troubled 2020 tally forced the bureau to guess the makeup of more homes.
In fact, the results were more mixed, according to the study. Overall, the share of population counts that were imputed declined slightly. On the flip side, the office was forced to guess much more often who lived in collective quarters like college dorms, long-term care facilities, and prisons. Part of the reason is that the pandemic has forced many college students to return home, making it more difficult to count them in dorms or apartments where they would normally have been counted.