The National Center for Education Statistics published a request for information on October 29 about redesigning its portfolio of administrative data collections and cross-sectional and longitudinal surveys. NCES wants a framework that can guide which studies it runs, how it collects and distributes data, and how the portfolio can respond as education conditions change. Comments are due December 3, 2026. The notice opens a design process; it does not adopt a framework or change a specific collection.
What changed
NCES says it wants to align its work more closely with state and local needs, produce high-quality evidence faster, and increase the use of that evidence by educators and policymakers. The agency also identifies a missing capability in its current process: it has not adopted strategic foresight to systematically consider possible futures and the information those futures might require.
The RFI asks what data states, districts, educators, and families need now; how collections should prepare for changing conditions; which collection and dissemination methods fit those needs; and which partnerships could support the work. One question uses the agency's chosen term “Super Intelligence or SI,” followed by “formerly referred to as Artificial Intelligence or AI,” and asks how collection and dissemination might be optimized for those applications. A practical design challenge is whether future systems can use the data without losing its definitions, limits, and history.
Two additional questions expose the central trade-offs. NCES wants to reduce respondent burden as survey response rates decline while preserving the public value of its data. It also says that derived variables, imputation, extensive documentation, and products for nontechnical audiences compete with its timeliness goals. The agency is asking how to balance those obligations before it redraws the portfolio.
Why it matters
A portfolio decision determines which questions the national education evidence system will still be able to answer years from now. Faster releases may make current information more useful. Streamlined collections may reduce the work demanded of schools, families, and institutions. Those gains can still carry costs if the redesign weakens documentation, breaks continuity, or drops a baseline that later becomes important.
Longitudinal and cross-sectional surveys do different jobs. A current snapshot can describe conditions at one point in time. A longitudinal study can follow change across years. If a future policy shift, demographic change, or new learning environment creates an urgent question, the answer may depend on observations collected before anyone knew the question would matter. A discontinued baseline cannot be recreated after the fact.
Machine use raises a related design problem. A system can process data quickly while missing a changed definition, an imputed field, a revision to a survey instrument, or a population that the source does not represent well. Addressing that risk would likely require stable identifiers, version history, provenance, documented transformations, and clear access conditions. The RFI does not prescribe those controls. It creates the opening to decide whether they belong in the future portfolio.
The dated signal is the move from improving individual studies to designing the portfolio as a whole. NCES is asking what should be collected, for whom, by which methods, and with what future uses in mind. That puts omission decisions on the same level as collection decisions. What the system stops observing can shape future knowledge as much as what it measures faster.
Verification bottleneck
The RFI provides questions, not a proposed framework. The next useful evidence will show whether NCES makes its portfolio choices inspectable.
A reviewable redesign would identify which collections are continued, modified, combined, or discontinued and explain why. It would show how the agency weighed burden against public utility and timeliness against documentation. It would also preserve a detectable record of changes to definitions, survey instruments, derived variables, imputation practices, and release formats.
The machine-use question needs the same discipline. “Optimized” could mean faster access, standardized formats, richer metadata, new interfaces, or something else. A later framework should specify the intended use, the users it is designed for, the quality controls attached to it, and the route for correcting errors. Without those details, optimization remains a goal rather than an operating design.
Watch next
The first milestone is the December 3 comment deadline. After that, watch for a proposed framework, a portfolio review method, or study-specific implementation notices. The strongest evidence would be a public decision record that connects each major portfolio change to an identified need, trade-off, and revision rule.
Specific signals include named criteria for preserving longitudinal baselines, explicit standards for documentation and versioning, and a clear account of how products for nontechnical users will coexist with machine-readable outputs. Any proposal to retire, combine, or substantially alter a study should also reveal what capability is being gained and which future questions may become harder to answer.
The RFI does not settle those choices. It marks the point at which NCES has put them on the public record, before the architecture is fixed.
Sources
- National Center for Education Statistics, “Request for Information,” Federal Register document 2026-22047, published October 29, 2026.
