Trendspotting: 2020 Census Data Gets a Green Light

NewsWire: 9/21/21

• The 2020 census results have gotten a cautious thumbs-up. A task force has stated that there is no reason to believe that the results were less accurate than those of the 2010 census. (The New York Times)
• NH: The 2020 census was plagued with problems. The pandemic shut down much of the count early on in April 2020, forcing a two-month extension. Extreme weather events, including wildfires and hurricanes, kept census takers from knocking on doors. And near the end, the Trump administration halted the count several months early.  
• All these challenges sparked fears that the 2020 data would be unusable and delegitimize its primary political purpose: reapportioning House seats. But a task force of expert statisticians has now given the results a cautious vote of confidence, saying there’s no evidence that either the nationwide or the state-by-state counts were less accurate than they were in 2010.
• Self-response (what people write on their census forms) is considered the most accurate source of census data. The 2020 self-response rate was 67%, which is slightly higher than 2010’s 66.5%. If the Census doesn’t receive a completed form, workers use administrative records, visit the address in-person, or get information from a neighbor. As a last resort, they use imputation, which is basically an educated guess of the number and characteristics of people living at an address.
• In 2010, 0.4% of households were counted by imputation. In 2020, despite fears that a troubled count would generate many more nonresponding households, this imputation share actually fell slightly.