Articles

The Insights Association warned that proposed changes to the 2026 Census Test, a huge piece of preparations for the 2030 Census, will make it a huge waste of resources that undermines an accurate headcount of the U.S. population.

Howard Fienberg, Senior VP Advocacy for the Insights Association (the leading nonprofit organization representing the market research, insights and analytics industry), commented that, “This new proposed approach to the 2026 Census Test – an essential piece of planning and executing an accurate 2030 Census – could save time and energy by just lighting a huge pile of taxpayer money on fire.”

Fienberg continued, “In the middle of the last decade, shortsighted decisions and regular funding interruptions forced the elimination of almost all field testing, which helped lead to inaccuracies in the 2020 Census count, particularly in rural and remote areas, and among small populations. The Administration’s new plan for the 2026 Census Test – reducing it from 6 test sites to only 2, eliminating testing in rural, remote and tribal areas, not testing the counting of group quarters, using a lengthy American Community Survey (ACS) questionnaire instead of a normal decennial questionnaire, limiting internet self-response, and trying to use U.S. postal workers as census takers – is putting the U.S. on a dangerous course that could jeopardize the 2030 Census.”

He concluded that, “An inaccurate decennial census will hobble most every other research study in the country for the next decade, public and private, since most rely on census data for statistical reliability and representativeness.”

Driven by these concerns, the Insights Association joined a March 4, 2026 letter urging Commerce Secretary Howard Lutnick (who oversees the Census Bureau) “to reconsider the proposed redesign of the 2026 Census Test and to restore promising elements of the test.”

The letter, led by the Population Association of America (PAA), came in response to a February 3 Federal Register notice announcing the proposed changes, and addressed many of our concerns in detail:

• On the issue of reducing the number of test sites “from six to two locations without a clear rationale,” the letter noted that those original six sites “were chosen after years of research to ensure the Census Bureau could adequately test innovative enumeration strategies and data collection technologies with the goal of making the 2030 Census more accurate and cost effective. The sites that were eliminated (tribal lands within Arizona; Colorado Springs, Colorado; Western Texas; and Western North Carolina) encompassed rural communities, tribal lands, and military installations as well as large regions with limited cell phone service and few physical mailing addresses. We are concerned that removing these sites will ultimately exacerbate the undercount of historically hard-to-count populations, including American Indians, young children, rural residents, Latino, and Black communities. As a reminder, in the 2020 Census, Latinos experienced a 4.99 % net undercount (more than three times the rate of the previous decade); Native Americans on reservations experienced a 5.64 % undercount; and Black communities experienced a 3.30 % undercount. Further, the 2020 Census resulted in the highest undercount of young children (ages 0-5) since the Census Bureau started collecting these data in 1950. When populations are undercounted, data quality is compromised. Reducing the scope and size of the 2026 Census Test jeopardizes the outcome of the 2030 Census, setting the stage for potentially high undercounts of regions and populations in the next decennial. We urge the Department of Commerce to reconsider their decision and fully restore the originally proposed six sites nationwide.”
• On the use of the lengthy ACS questionnaire instead of a simple decennial questionnaire, the letter continued, “The intent of the Census Test is to field the decennial questionnaire form. The ACS is a much longer form that asks different questions than those asked in the decennial questionnaire, rendering it inappropriate for use in the 2026 Census Test. Further, using the ACS form comprises the use of data for informing decennial census methodology. Using any questionnaire other than the decennial census form to test census operations is not scientifically valid and will not produce useful information to improve 2030 Census operations and outreach. Therefore, we urge the Department of Commerce to field the decennial census questionnaire, and not the ACS, in the 2026 Census Test.”
• On the limitations on internet self-response (ISR), including offering it only in English, the letter noted, “The original design of the 2026 Census Test had proposed providing ISR in English, Spanish, and Chinese and allowing self-response via several modes (internet, phone, and paper questionnaire). Research has shown that making self-response available in multiple languages ultimately increases self-response, reduces in-field costs, and increases data accuracy. By limiting ISR to English only, more in-field staff will be forced to visit homes that otherwise would have responded online, increasing both costs and respondent burden. Given these factors, the Department of Commerce should reconsider its decision to limit ISR to English only as well as consider expanding the modes of self-response.”
• On using postal workers as census takers, the letter commented that it was “not an original idea and has been tested in previous simulations. A 2011 report issued by the Government Accountability Office (GAO) reported that using mail carriers to conduct census field operations at USPS pay rates would not be cost-effective. In 2010, USPS mail carriers cost, on average, about $41 (city) or $34 (rural) per hour compared to about $15 per hour for census part-time enumerators. The GAO concluded it would be more efficient to hire both retired and active postal workers as Census temporary field staff to help with enumeration. … Revisiting issues that have already been thoroughly tested diverts funding from developing and assessing other promising strategies that could be used to make the decennial headcount more efficient and accurate. If the Department, however, continues with its plans to test the use of postal service employees in the 2026 Census Test, they should be transparent about the methodology and how this evaluation is going to be different than those conducted in previous years.”

- PAA letter to Commerce Secretary joined by the Insights Association (3/4/26)