What’s Next for the National Center for Education Statistics? | American Enterprise Institute

The future of the National Center for Education Statistics (NCES), and the Institute of Education Sciences (IES) that houses it, remains unsettled. There’s broad agreement that the nation must keep collecting education data. Far less settled is whether NCES in its current form is the right agency to do it.

The table below shows that NCES has operated differently from other federal principal statistical agencies. The rightmost column shows appropriations per staff member. In 2024, the US Census Bureau averaged about $180,000, the Bureau of Labor Statistics about $340,000. Every other agency clustered between $250,000 and $400,000. NCES spent over $3.8 million per staffer—10 or more times higher than every other agency. After DOGE cut staff to around 20, that ratio became even more extreme.

Sources: Office of Management and Budget and American Statistical Association.

In the 1990s Emerson Elliott, the first presidentially appointed, Senate-confirmed NCES commissioner, built the staffing model behind that table. Increasing NCES staff during the Clinton years was politically difficult because there was bipartisan skepticism of the federal bureaucracy. Instead of fighting that battle, Elliott focused on securing money for programs. As NCES’s budget expanded, Elliott spent a growing portion of it on contractors. He created the Education Statistics Services Institute to provide contractor support for statistical data collections, analysis, writing, and survey development. After the rapid expansion of the National Assessment of Educational Progress (NAEP) following 2002’s No Child Left Behind Act, NCES’s assessment division decided it needed its own support arm and created the NAEP Education Statistics Services Institute. (Both were folded into the Education Statistics Support Institute Network in 2011, continuing the model of paying contractors to do the work of NCES.) This strategy had a clear effect: NCES’s ratio of contractors to civilian employees is far larger than the average across the federal government—which itself is quite high.

Contracting out the bulk of its services turned NCES into a contract management shop. But NCES never built a robust system for effective contract management. It hired PhDs for their subject expertise, then had them manage contracts—a choice akin to a hospital hiring highly trained doctors and sticking them behind the reception desk. It also resisted attempts by IES to establish needed internal control processes—a resistance bolstered by other parts of the federal statistical system and enabled by the greater autonomy given to the NCES commissioner relative to other IES leaders.

The result: NCES relies on a small circle of preferred contractors, supported by outside groups like the American Statistical Association, that have spent years defending legacy collections and the structure of NCES as sacrosanct—all the while neglecting growing problems with NCES’s data collections, including lack of timeliness, high costs, and questionable usefulness. Contractors had sunk a lot of money and many years into expensive, inefficient data collection methods. When better, cheaper alternatives came along, the answer was always some version of, “why mess with the gold standard?” Moreover, because it was captured by contractors, NCES avoided addressing the harder questions: Can these collections be combined? Can better data come from elsewhere? Should some of NCES’s data collections stop entirely?

The same instinct is at work in current efforts to rebuild NCES, which have included hiring a couple of new people to work on NAEP’s technology and reinstating some former staffers to what look like their old roles. In other words, despite being nearly DOGE’d into oblivion—and given an opportunity to remake itself as a leaner, nimbler operation—NCES has more or less stuck with the clunky status quo.

If the current “rebuild” does nothing more than downsize a broken structure, NCES won’t (and shouldn’t) survive the next decades.

Before asking how many staff NCES should have, we must ask what NCES is for and whether the current model can deliver it. There is also a more fundamental question at play: Does NCES need to exist as a standalone federal agency at all? Some of its work is irreplaceable—NAEP is the only nationally comparable measure of student achievement we have, and the Integrated Postsecondary Education Data System is the backbone of national higher ed. data. But being irreplaceable doesn’t entitle an agency to a freestanding structure, a presidentially appointed commissioner with far too much autonomy, and a green light to continue its bad habits.

Focusing its data collection on what the nation needs and identifying modern ways of collecting those data will help ensure a healthy future for NCES.