Pro Publica: Elon Musk’s Team Decimates Education Department Arm That Tracks National School Performance

 

This article by Jodi S. Cohen and Jennifer Smith Richards in Pro Publica reveals that the raids on federal agencies have come to the U.S. Department of Education's Institute of Education Sciences, unlawfully "canceling" millions of dollars of in research contracts.

The AI-slinging team of course has no idea what they're doing — not only do they not understand the policies they are undermining, but they don't even seem to understand the data they are feigning to analyze. Snatching a gigantic database with elements coded for machine tabulation and then running it willy-nilly through language-based LLMs will inevitably produce nonsensical results, but the team doing it are hooting and hollering because they think they've accomplished something. It would have been efficient to simply ask the data scientists in charge of those databases what they wanted to see, but the new administration has committed itself to ignoring experts.

What the team of invaders almost certainly do not know is that current version of IES was the creation of a GOP Congress who in those long-ago days actually wanted some actual scientists to use scientific models to evaluate all the things that education vendors try to sell to state legislators and school board members.

The background is that nearly all brand-name education products (even the few good ones) are sold like snake oil — full of nonsensical claims about student achievement and attached to costly, insipid, and dull professional development programs that the vendors sell as add-ons. Congress knew this, and wanted someone to keep score.

Educators have always known this, of course, and school principals were once expert practitioners of sly eye rolling and quiet grins while vendors offered their wares. Confusing the process was that real educators and education researchers were always at work studying educational techniques (which can't be patented or branded). Those legitimate education researchers published their findings in academic journals, which historically have only been available to university libraries and are certainly not affordable to schoolteachers and principals.

So what to do? IES (and its What Works Clearinghouse, an ancient but respected education research database known as ERIC, the National Center for Education Statistics, and the U.S. Department of Education's Office of Civil Rights once-amazing Civil Rights Database) was supposed to help fix all this. The idea was that it presented real science to real educators to help them navigate the treacherous and preposterous education marketplace.

Unsurprisingly, education vendors didn't like this because report after report issued by IES indicated that all the high-cost tech bro solutions to problems in schooling didn't really help. Pick almost ANY brand name education product you've heard your children or their teachers mutter about in the last 30 years, and there is an extremely high likelihood that there is an IES-funded report indicating that the product wasn't particularly good at what it claimed to be good at it.

IES then realized that the average schoolteacher and principal needed help navigating the enormous amount of conflicting research and data, so the created a program called the Cooperative System Fellows. This network connected researchers, data scientists, and educators by bringing them to the IES research facilities in Washington, letting them meet with the resident scientists and research teams, and practicing ways to communicate all of this to people working local school districts.

I was one of the first Cooperative Systems Fellows, and for a couple of decades in that role I helped school board members, superintendents, principals, and teachers understand what was going on. Back in the chaotic days of No Child Left Behind and Reading First and the advent of high stakes testing, this network of local, state, federal, and private researchers kept educators in the know about what seemed to working and what seemed to be not working. I was able to send regular research reviews, synopses, and trend data to principals and teachers, and served on numerous district and state teams to use all this information to help schools help kids.

I helped several other people become Cooperative Systems Fellows as well, and combined our work with that of several statewide organizations. For a while there, it seemed like educators were going to be able to manage the deluge of marketing from education vendors, and identify those few effective curriculum products while weeding out the nonsense.

Year by year, the vendors got better at marketing and eventually began pushing whole-district commitments. This was contrary to the false promises of school choice (that every school would be its own laboratory of innovation, etc.), but the marketing won out. Still, IES did its best, and kept on explaining how much money was being wasted on the tech bros instead of being spent on teachers and students.

Moreover, policy analysts got really good at using IES and OCR and NCES databases to show how corrosive legislative meddling had become. Study after study showed that "choice" plans and charter schools and vouchers and tax scholarships and so on were methodically re-segregating schools by race, by ethnicity, and by income. The public, private, and charter schools of 2025 are FAR more segregated than the public, private, and charter schools of 2005, and all of that was deliberate. IES and OCR and NCES helped explain this phenomenon, and this became hated by white nationalists and privatizers.

So this absurd attack on scientists and educators might be at last the decisive victory for vendors and a generational loss to student, teachers, and principals.