Asian Americans were overcounted on the last census, according to a newly released report card detailing where the 2020 data fell short. But community experts say these numbers can be misleading, and, if taken out of context, can mask the experiences of lower-income Asians.
As community leaders across ethnic groups reflect on how accurately their people were tallied, it's obvious many Americans were left out. For the Black, Native American and Latino populations, the data is consistent: their communities were significantly undercounted.
Before the census was even distributed, Asian American advocates worried about encountering similar problems. Considering language and outreach barriers, Asians were among the least familiar with the census and least likely to say they planned to fill it out.
But the data paints an entirely different picture. The Census Bureau estimates that Asian Americans were overcounted by 2.62 percent in 2020, meaning many were counted twice.
“It definitely came as a surprise to me,” said Linying He, associate director of development at the Asian American Federation. “But this aggregated data is conveying a wrong message. There is a significant difference between the members of our community who are overcounted and those who are undercounted.”
Overcounting usually happens with college students or those with multiple homes who fill out the census in more than one location. It connotes an economic advantage, and experts say it could contribute to the false image of Asians Americans as wealthy “model minorities.”
In the ethnic group with the largest wealth gap in the country, He said an overcount can further shadow the experiences of vulnerable, undercounted Asian communities.
“Those who are most likely to be undercounted include economically disadvantaged groups, such as recent immigrants who are not aware of the census, populations with limited English proficiency or limited internet access, low-income families, young children and refugees,” she said.
The census numbers, along with other reports like the annual American Community Survey, can impact where resources and outreach are directed. It can also impact which language groups get translated government materials.
“It’s really important that we don’t view this overcount as meaning the Asian community is doing fine,” said Howard Shin, managing director at AAPI Data. “There’s significant portions of the Asian community that are harder to count, that are unaware of the importance of being counted or have language issues.”
Despite that, experts say local organizers deserve credit for making sure the number of Asians omitted from the census decreased in 2020. Rhetoric around citizenship and immigration turned off many in the community from participating, Shin said. So there was a ramped up effort among advocacy groups to spread the word.
“They did a lot of on-the-ground outreach work that was conducted in language and in culturally familiar programs to Asian American communities,” He said.
But with the Post-Enumeration Survey only being offered in English, filling in the missing pieces was impossible for some groups.
It’s a problem that exists across all government paperwork, she said, and without proper understanding, it can amount to disenfranchisement.
“The Asian American community has one of the highest Limited English Proficiency rates,” He said. “So they need translated materials, like for voting. If those materials were not provided to them, they would just be excluded in the process.”
Collecting data under the umbrella term “Asian American” — used to categorize everyone from Burmese Americans, who have a 39 percent poverty rate, to Indian Americans, who have an 8 percent poverty rate — is one of the things experts say is holding the community back.
Five states have passed laws requiring data be collected separately for major racial groups. Until the census catches up, He said, it will be hard to tell who is truly being left out.
“I don’t want people to read this overcount as equal to ‘oh, Asian Americans are doing better,” she said. “Only through disaggregated data can we tell who is overcounted and who is undercounted.”