Articles

The Census Bureau’s proposed rule to rewrite how the next decennial census is run poses substantial challenges to the insights industry. The rule will restrict demographic questions, and change who counts and where they are counted.

On September 10, 2026, the U.S. Census Bureau published a proposed rule that would eliminate race and ethnicity questions from the 2030 Census, and change the residence rules for the 2030 Census, writing them into federal regulations at 15 CFR Part 60, in multiple ways that the Insights Association worries would make the 2030 Census harder for ordinary people to answer correctly, less consistent in where the census counts them, and less dependable as the benchmark on which all U.S. statistically-representative research relies.

IA will be filing public comments, which are due November 2, 2026 (following an extension of the original 30 day comment period). We also encourage all insights professionals to use the IA Action Center to warn their Representatives and Senators about the race/ethnicity question elimination.

Why the residence changes matter: When the rules for where people are counted become inconsistent or ambiguous, the errors do not stay in the census — they flow into every dataset calibrated to it.

Switching from Census Day to a 3-month "Census Enumeration Period": Since 1790, the census has counted people at their usual residence. In 2020 that meant the place where a person lives and sleeps most of the time, as of April 1. The proposed rule replaces that with a calendar math test: a person's usual residence would be wherever they "lawfully spent the greatest number of days" between January 3 and April 1. A day counts toward a residence if the person was there for as little as one hour, and a single day can count toward more than one residence. For someone who stays in multiple places (a snowbird splitting the winter between two states, a child in shared custody, a traveling nurse, a commuter with a weekday apartment), an accurate answer would require good knowledge of that person’s hourly schedule over 89 days, and then figuring out how overlapping days add up. Insights professionals are well aware that research subjects struggle to reliably recall such levels of detail days or months later and that complicated math will lead to inconsistent answers. Swapping a relatively-simple intuitive question for a complex question would dramatically increase respondent burden and error.

Counting college students at their parents' homes: Since 1950, students living away from home at higher education institutions were counted at those institutions, where they actually live and sleep, even if they happened to be home on April 1 on vacation. The proposed rule drops that specific instruction. Students would instead be treated as general dependents with their residence determined by tax records. Since many parents claim these students as dependents on their tax returns, and the one-hour rule means vacations and weekend visits home would all add days to the parents' address, this means college students are likely to be counted at home and local data on 18-to-24-year-olds, an audience whom the insights industry already struggles to reach and measure, will be statistically distorted.

Clear rules for group quarters deleted: The 2020 Census residence criteria addressed more than 20 distinct residence situations; the proposed rule consolidates them into eight. It maintains specific rules for prisons, military barracks, and transitory locations such as RV parks and hotels, but it gives no guidance for nursing homes, college dormitories, shelters, group homes, or residential treatment facilities. Absent explicit rules, residents of these places would have their residence determined by the day-counting and tax-record tests, which don’t suit, for example, a nursing home resident whose tax return lists her adult child's home address.

Tax records are not designed to capture residence: The proposed 2030 Census rule lays out that a person's usual residence should be consistent with and evidenced by federal tax records such as returns and W-2s. However, tax addresses frequently reflect mailing addresses, PO boxes, tax preparers, or employer payroll offices rather than where someone lives. Some people, such as lower-income households, older adults, children, and students, are not required to file taxes at all. Tax returns cover a full year rather than the January-to-April window, and filings made after April 1 (including extensions into October) would make determining someone’s residence even more messy. The Bureau acknowledged its reliance on these records will vary by situation, which underscores that they cannot serve as a consistent standard.

Unclear scope of counting only citizens and lawful permanent residents: The Insights Association has no stake in the apportionment process, but every stake in the overall headcount. The proposed rule states that foreign citizens who are not U.S. citizens or lawful permanent residents would be "not counted for apportionment." Yet the notice describes the criteria as determining whether and where to count people in the census generally, while also noting in a footnote that the Bureau will continue to publish statistics separate from apportionment. The proposal never states plainly whether people excluded from the apportionment base would still be enumerated and included in the total population, redistricting files, population estimates, and survey controls. For data users like us in the insights industry, the difference is enormous: excluding millions of residents from the base counts would leave every downstream benchmark short of the real population.