CHN submitted the following comments to the Independent Census Scientific Committee for its fall 2026 meeting. You can find their proceedings and recommendations to the Census Bureau here. The meeting focused on two critical issues: the new privacy rule and the proposed residence rule for the 2030 Census.
They specifically said “I-CSAC asserts that the scientific evidence does not support the proposed changes to privacy protection in the Commerce Department Administrative Order (DAO) 216-26 order. The proposed methods will substantially reduce the utility of a wide range Census data products across many sectors and geographies.” They made a number of recommendations about how to mitigate the harm of this order.
For the residence rule, they recommended that the Bureau not include questions about citizenship or immigration status in the 2030 decennial census and include in its 2030 Census tabulations all persons who meet the residence rules applied in 2020 (regardless of whether or not they are included in the counts used for apportionment, which will be defined by the courts). They also recommended asking about race and ethnicity, in accordance with the 2024 update of the race and ethnicity standards.
They did not address a number of our comments and recommendations on these topics, and our comments also included recommendations on other points. You may find it helpful to read both their recommendations and our comments.
CHN Comments to the Independent Census Scientific Advisory Committee
The Coalition on Human Needs is submitting these comments with gratitude for the opportunity to provide advice to the Census Bureau on critical topics affecting the count of young children.
We have the following comments.
We urge the ICSAC to oppose the proposed residency rule outlined in the federal register notice of September 10 because it will fail to count millions of people who should be counted, including immigrants, adults with no tax records, and children whose parents do not file returns. Unless the parent files a tax return, even when the IRS has tax records on the parent they do not have records on their dependents who are mostly children. We believe that it is essential to collect race and ethnicity data and note that children are the most diverse age group. We also note that the proposed rule violates the statutory residence time period since the statute says people should be counted on April 1—not over a three month period.
We add our voice to the many people who have grave concerns about a revised resident rule that will not count immigrants unless they are naturalized citizens or hold green cards. We note that one in four children has at least one immigrant parent, so this proposal will likely result in missing millions of children if their parents are excluded even when the children are citizens. The plan to avoid collecting data on race and ethnicity is hard to accept. Data on race and ethnicity has been a staple of the Census since the first census in 1790 and provides critical information for a wide range of purposes. We note that it is particularly important for children since they are the most diverse age group.
We also note that following the new proposed rule requirement that addresses be evidenced by tax records would miss millions of low-income people who have no tax records, and millions of children who will not appear on their parents’ tax records. When the Census Bureau conducted an administrative census, it discovered that the administrative census which used 31 different sources of administrative records, missed three million native born people who were counted in the 2020 census, https://www2.census.gov/programs-surveys/decennial/2020/program-management/evaluate-docs/EAE-2020-admin-records-experiment.pdf , see p. xii. A census which used only tax records would miss many more native born people than this administrative census which used 31 sources, and also would miss many immigrants. We also note that the administrative census used a probabilistic approach to verifying addresses and even so failed to verify many addresses.
There are other concerns with the tax record requirement. Many people do not use their residence for their tax address. IRS tax addresses are not legally required to be residences, and many people use post office boxes, a military address, their business address, their accountant, or just a mailing address they consider to be reliable. People who move during the proposed residency period or who filed a return extension request might have tax records with a different address than their residence on April 1 (or over the three month period). We also note that the proposed residence rule which considers that people might have multiple residences on a given day would potentially require that children whose parents have joint physical custody might actually spend the most days at and should be counted at their school or child care rather than with either parent (and this of course could not be confirmed by tax records.)
We are deeply concerned that the new Commerce Department order on data privacy will greatly reduce the precision of Census data on young children, suppress important data on young children, affect the allocation of federal and state funds for programs benefiting children such as the Title 1 funding allocated using SAIPE data, and undermine efforts to measure the undercount of young children in 2030. We ask the ICSAC to oppose it.
We note that since children are a subset of the total population, and young children even a smaller subset, this order may affect the quality of the data even for fairly large geographies. This data is essential for many purposes, including re-apportionment and redistricting, governmental planning, policy analysis, and business.
We particularly note that when the data is suppressed it may adversely affect the allocation of federal funds. The Project on Government Oversight found 70 federal programs use census data to allocate funds that benefit children, allocating $793 billion dollars in FY 2023. For some of those programs, the data used to allocate the funds is broken out by small populations and geographies, and if that data is coarsened or suppressed, the intended funds may not go to the communities that most need it and were supposed to get it. We are particularly concerned about the Title 1 School funding, which uses the Census Bureau’s Small Area Income and Poverty Estimates (SAIPE) data, and where the funding formula is established by statute. We further note that if the data on child poverty in local education areas and communities is coarsened or suppressed, it may be harder for schools to use the Community Eligibility Provision for school meals, for schools to provide summer meals to all children, and for child care programs to provide meals to all children in their care. (The latter two programs use data from three census tracts to determine eligibility.) The results may be many more children going hungry. We continue to assess what other program funding may be affected.
We further note that even when the federal government uses data that is not coarsened to allocate funding to states, states may need data that has been coarsened or suppressed to allocate those federal funds—or other state funds—to communities.
Thus, this order eliminating important privacy protection options is likely to mean that for some programs, the communities that most need federal or state funds may not get their statutory allocations in full.
We are concerned that this order may reduce the precision and/or availability data from the Census Bureau’s demographic analysis technique for measuring the undercount of young children by county in the 2030 Census.
In addition, the ACS release has been delayed for 2026 in order to comply with the Commerce Department order, and we anticipate that other data, such as the SAIPE data, may also be delayed. The older the data is, the less useful it is. In some cases the data will come out after it is needed for allocating federal funds, and the funding agency will have to use other data such as the prior year (we understand that was the typical approach with pandemic data delays.) In particular, using past year data will be a problem for growing communities.
We urge the ISCAC to recommend that the Census Bureau review how any data they are considering suppressing or coarsening was used in the past, particularly in a funding context, and make suggestions about what data they should use in its place. We recommend several places they can look to see how the data they stop producing has been used in the past: this report by the Census Project, https://thecensusproject.org/acs/; the resources on this page, https://thecensusproject.org/acs/; the Casey Foundation KIDS COUNT project; and this survey of how policy analysts at the state and local level use the data. https://countallkids.org/wp-content/uploads/2022/03/Use-of-the-American-Community-Survey-by-state-and-local-Child-Advocacy-Organizations-3-8-2022-1.pdf
Because young children have a higher net undercount than any other age group and the net undercount of young children has grown steadily from 1980 to 2020, we ask the ICSAC to recommend that the Bureau test new strategies for counting young children now, in time to implement them in 2030. We regret that the Census Bureau did not use the 2026 test to investigate better ways to collect data on young children. It is essential to test multiple new strategies since some may not work; therefore we are disturbed that while the original plans for 2026 Census test included testing such new strategies, the revised 2026 census test as implemented did not include those strategies. We note that the GAO recommended testing all the new strategies originally planned, and the Commerce Department accepted that recommendation; we encourage the Bureau to test those new strategies promptly so that any necessary adjustments can be made.
We ask the ICSAC to recommend that the Bureau conduct a Post Enumeration Survey for 2030. We are excited by the potential of the alternative approach the Bureau is testing to measure census accuracy in 2030, but it is entirely untested and there are a number of features that need to be worked out. Until that approach is tested, the Bureau needs to conduct a tested, accurate approach. Furthermore, a Post Enumeration Survey would be essential for measuring the accuracy of this alternative approach.
We ask the ICSAC to recommend that the Bureau evaluate how many people counted by the non-ID response approach in 2020 would otherwise have been missed, and which populations were most likely to be so counted. That will inform decisions about how to expand and support that program in 2030. (We note that it is far cheaper and likely more accurate than sending enumerators to homes.) In 2020 17,590,000 non-ID responses were submitted and 9.4% of households were counted using this innovative approach. It is critical to understand which kinds of households were enumerated using this approach. For example, was this a good way to improve the count of young children and should the Bureau expand its outreach regarding this approach? We also understand that in many cases the address used was not on the master address file and had to be verified in the field (https://www.nationalacademies.org/read/27150/chapter/9#152 , p 152 ). We urge the Bureau to make sure that its verification process recognizes that many of these addresses may be for ADUs or other nontraditional housing (such as garages or businesses that also serve as residences.)
We ask the ICSAC to urge the Bureau to work with experts on low literacy to develop approaches to help people with low literacy skills to self-respond at a higher rate. We note that this is both more accurate and cheaper than reaching them via NRFU. The best available data indicates 20% of all adults lack basic literacy skills (among English language users) and another 30% can manage generally but struggle with more complex materials.Young children are disproportionately missed in households that have no adult with a high school degree. We are delighted that the Bureau tried to use simple English in its outreach materials (not the actual questionnaire) for the 2026 Census test. However, this is only one aspect of reaching people with low literacy skills. We urge the Bureau to work with literacy experts to determine the best way to reach these people (for example, TV, radio, online music stations, online game platforms?) and how they prefer to respond (phone?) and then advertise to them how they can 12 easily respond.
The ICSAC should ask the Census Bureau to begin planning its 2030 Communications campaign and also share Professor Rocco’s research with states and localities showing that communications outreach increases self-response rates. Since self-response is both the most accurate response and the cheapest, both the Census Bureau and the states should invest in a communications campaign urging self-response. Professor Rocco’s research is described here and has been submitted to a journal for publication. We note that state outreach was particularly effective in communities with many people of color, many people with linguistic barriers, and communities with high numbers of renters—all of which are communities which also have large undercounts of young children.
Thank you for the opportunity to submit these comments. If you have any questions about these comments, please contact Chris Kang at ckang@chn.org.
