Inexplicable Cuts at the Dept. of Education

Recently, DOGE has bragged about terminating 89 contracts worth a collective $881 million at the Institute for Education Sciences at the US Department of Education.

Penny-wise but pound-foolish. They have unwittingly canceled some of the best education research out there, along with major national surveys and tests that are crucial to tracking America’s educational performance.

[Their claim is also highly misleading: as discussed below, many if not all the contracts they terminated were already well underway, and much of the $881 million must have been already paid out and spent.]

***

The Department of Education’s own Twitter account has attempted to explain:

These are the only examples of supposed waste thus far. But reporter Matt Berg has the full list of canceled projects here (so far, I haven’t been able to find the actual contracts on “zoom and in-person meetings” or “mailing and clerical operations” but maybe they’ll turn up someday).

As someone whose actual PhD is in education policy, and who has been following these issues for about two decades, I’ll say this: Much of what DOGE cancelled here is not waste or abuse under any conceivable definition of the term. Indeed, DOGE has canceled many high-quality research projects that had already been underway for years, which is far more wasteful than letting them finish!

***

Take a step back. What is the Institute for Education Sciences, or IES?

It was created by the Education Sciences Reform Act of 2002, a major project of George W. Bush. IES is specifically mandated by Congress, so it is not something that a presidential administration could lawfully abolish:

The official mission is to sponsor major national research projects and surveys that track “the condition and progress of education in the United States” and to study programs that may “improve academic achievement and access to educational opportunities for all students.”

IES includes various subunits, including the National Center for Education Statistics, which is the main source for a ton of incredibly valuable resources to track national progress, including the Common Core of Data (CCD) and the National Assessment of Educational Progress (NAEP, pronounced like the word “nape”). Whenever you see a chart like this, you’re viewing the work of IES and NCES.

IES has been controversial in the education community for the past 2 decades or so, ironically because of its relentless focus on high-quality randomized trials to test everything from curriculum, to teacher training programs, to dropout-prevention programs. As we’ve seen before, the education community can be surprisingly resistant to randomized trials, and you can see this resistance in a 2022 report from the National Academies:

I’m sure that, like any government program or agency, there are cases of arguable waste and overspending at IES. But for now, it’s the one place in the federal government that produces the highest-quality education research and statistics, even in the face of occasional resistance.

***

With that as background, what did DOGE cancel the other day?

There isn’t time or space to review everything, but here are a few representative examples as I go through the list and try to find more information online.

FIRST EXAMPLE:

Here’s an impact evaluation of reading programs in early elementary school. Some screenshots in case the webpage is deleted:

In other words, this was a large-scale randomized trial across 140 schools nationwide, in order to determine how best to help struggling students in early grades learn to read. Is the contract too expensive? Who knows, maybe so — I’ve reviewed contracts/grants exactly like this when I was in philanthropy (and in one case, we actually made a $20 million grant to the research firm MDRC to do some major studies on preschool). Research firms can sometimes pad the bill by adding a ton of survey work that isn’t 100% necessary.

But that’s beside the point, because it is guaranteed that no one dug into the exact details and activities enough to know if that was happening. And if there’s anything we should be doing in education research, it’s funding large randomized trials on how to more effectively teach kids to read.

In short, here we have a research project that was already underway for the past several years. Even if the budget was somewhat padded, it is MUCH more wasteful to pull the plug halfway than to let the project finish. If the project finished up, it would deliver some useful findings. Pulling the plug now means that all the millions spent to date are wasted!

SECOND EXAMPLE:

There’s a contract to Abt (another research firm) to evaluate the DC Opportunity Scholarship Program, i.e., the voucher program that has operated in Washington DC for over a decade now.

Here is Abt’s page about it, which correctly notes that Congress itself mandated this exact research!

[In the interest of full disclosure, I’ll admit to being a little biased here, in that Anna Egalite is an old friend of mine, and an outstanding education researcher. But I didn’t know she was involved with this contract until literally just now!]

This is valuable education research on a long-running voucher program, and it was mandated by Congress no less. And we're canceling it . . . why?

THIRD EXAMPLE:

In the No Child Left Behind Act of 2002, Congress created and funded so-called “21st Century Community Learning Centers,” which were essentially a ton of afterschool programs in high-poverty schools with the goal of reducing drugs and violence, offering kids arts and music programs, and more. Congress spends some $1.3 billion a year on this initiative.

Mathematica (the research firm, not the software firm) had a contract originally awarded in 2019 to study how this initiative has been implemented and what its impacts are. That contract got canceled by DOGE.

When we’re spending over $1 billion a year on this program, does it make sense to cancel a multi-year contract worth about 1 percent of that amount that might tell us how well the program is working (if at all)? Again, it’s much more wasteful to cancel a contract like this midstream rather than wait for it to deliver results.

FOURTH EXAMPLE:

The DOGE team unwittingly canceled some of the most famous, long-running, and useful studies in all of education research. This is truly disheartening.

• The High School and Beyond Longitudinal Study. This is a continuation of a large-scale national study that originally launched in 1980, and has led to many hugely important educational findings.

• The Early Childhood Longitudinal Study, Kindergarten Class of 2022-23. This is another continuation of a long-running effort to improve kindergarden education in America.

• Trends in International Mathematics and Science Study (TIMSS) 2023. If you’ve ever seen charts comparing American 4th and 8th graders to other countries as to math and science performance, that’s due to TIMSS. It’s a huge undertaking, but important.

• The School Survey on Crime and Safety. This is a survey of about 4,800 schools, and is the “primary source of school-level data on crime and safety” in the United States.

• Programme for International Student Assessment, or PISA. This is a worldwide effort by OECD to track high schoolers’ reading, math, and science knowledge.

If there’s anything the federal government can do well, it is to collect national statistics on how we’re doing as a nation. DOGE is trying to cancel many such efforts for no apparent reason.

***

All of the above are illustrative examples. If I had the time, patience, and access to information (many of the contracts aren’t readily available online), I’m reasonably confident that there’s a good case for keeping almost all of these contracts.

None of the above is wasteful, or “woke,” or anything else that would be objectionable to 95% of Americans (barring the occasional nihilist who doesn’t care about American education and thinks that no amount of data or research could possibly matter).

DOGE needs to sit down for a month with IES staff, outside experts in educational policy, and a bunch of the contractors themselves, so that they can get up to speed on what these contracts are about, along with learning some history about the past 50 years of educational statistics and at least the past 25 years of national legislation on education.

Canceling government waste is a great idea, but people with no background knowledge shouldn't cancel things based on a line they saw in a spreadsheet. It's hard to come up with an analogy that truly captures what happened here, but it's as if someone who had never watched a basketball game showed up at the NBA and bragged about all the millions he had saved by deleting contracts, and it turned out that he had fired LeBron James, Steph Curry, and Wemby. Alternatively, it's akin to a McKinsey consultant going to Tesla in 2018 with a plan for how they could save money by firing some random guy named Andrej Karpathy.

UPDATE: I didn't even notice this, but contracts for the Common Core of Data and the Private School Survey also got cut! Unbelievable.

Here's an interview with my old friend Lynn Woodworth (who ran the National Center for Education Statistics under the first Trump administration):

Q: Why is ending these contracts such a big deal for NCES?

A: Unlike other federal statistical agencies, NCES can use only a tiny slice of the money IES gets from Congress to hire staff to carry out these duties. So it has to contract out almost all of its work. NCES has fewer than 100 employees, and more than 1000 contractors.

Q: What’s the immediate impact on the work now going on?

A: Some of these surveys are now in the field. For others, researchers are analyzing the data that’s been collected. All of that work is being stopped, immediately, which means all the money that’s been spent getting to that point is just wasted.

Q: What will happen to the data?

A: It’s not clear. NCES doesn’t have its own data center, because NCES has never been given the funds to set one up and hire people to run it. So the data are held by the contractors. And when their contract is terminated, is the money for data storage also being terminated?

Q: The Department of Education has said its decision won’t affect the

National Assessment of Educational Progress

(NAEP), dubbed the nation’s “report card,” a massive activity managed by NCES. But it relies on data from other NCES surveys that have now had their contracts terminated. What’s your take?

A: NAEP is based on the test results of a small but representative sample of U.S. students. To figure out which students or which school should be included in your sample, you need the data from the CCD [Common Core of Data, an NCES-managed database on students in U.S. public schools]. Another NCES survey, the PSS [Private School Survey], provides NAEP with the same data for private schools. Without the data from the CCD and the PSS [whose contracts are now terminated], you can’t select and create a proper sample. And that is true not just for NAEP. It will affect every researcher in the country who uses CCD as the frame for sampling and weighing of their survey population.

Q: Department of Education officials have said they hope to save the government money by rebidding the contracts. How feasible is that?

A: There’s a very high entry cost to doing this work, because of the expertise needed to do the data collection and also to understand how the U.S. educational system is structured. Contractors have to do everything. They have to create the sample frame [the pool of potential respondents]. They have to provide the analysts, the computer system, and the equipment and people to actually do the data collection for dozens of longitudinal surveys. It’s very hard for new companies to participate. Sometimes there are only two or three legitimate bidders on a contract.

Q: Some researchers see canceling contracts as a precursor to the Trump administration eliminating NCES and IES because their work doesn’t align with the president’s executive orders—in particular, those ending diversity programs. Are you worried about that?

A: IES has been authorized by Congress [under the 2002 Education Sciences Reform Act], so I don’t think they can legally shut it down. And the law orders NCES to collect information on a long list of topics. But they can starve it of resources and, in effect, accomplish the same thing.

It’s also important to remember that all 13 federal statistical agencies, including NCES, don’t collect data just for the executive branch. They are collecting data requested by the legislative branch, and by the judiciary for making its decisions, as well as data for the American public. Plus, they’re not collecting data just for the current chief executive. The data are also to help future presidents make good decisions. That’s why it’s important for the federal statistical system to be independent.