Nobody runs randomized controlled trials for educational products because proving they work doesn't make money
Schools adopt teaching materials based on sales pitches, not evidence: no regulatory requirement for effectiveness, no profit in proving one product beats another, therefore no company will fund validation.
The video won't play here. Listen to the audio instead:
The argument · timestamps estimated from transcript position
Educational adoption runs on salesmanship because efficacy proof generates no profit
Jason Yeatman pointed out that schools use teaching materials and interventions primarily through sales representatives pitching to teachers and district leaders, with no regulatory requirement to prove effectiveness and no systematic comparison of teaching methods to identify the best approach. He stated directly that a large-scale randomized controlled trial costs enormous sums, and if the results don't generate profit, no company will fund validation of whether their product actually works. This isn't a failure of teachers or schools—it's a system-level gap: there's no mechanism to bring scientific evidence into the purchasing process.
— Jason YeatmanAcademics chase first authorship; parents want their child helped—these goals diverge
Karen Parker traced the root of cross-institutional collaboration barriers to misaligned incentives: academics prioritize first authorship and all trial outcomes; parents care only about their own child. She noted that physics adopted a 'working group' model three decades ago—detailed contribution statements replacing author-order disputes. Neuroscience still lacks a solution for equitable credit distribution in large-team collaborations.
— Karen ParkerCareer stage determines whether researchers can actually collaborate across disciplines
Karen Parker explained how career stage gates cooperation: junior researchers must chase first authorship; at promotion, they compete for last authorship; only after tenure as full professor does the pressure vanish, finally enabling their best cross-disciplinary work. The academic promotion system actually suppresses collaboration in its early stages—trainees can't really collaborate unless the incentive structure changes first.
— Karen ParkerBuild trust with schools first by solving their real problems, research partnerships follow
Jason Yeatman described his playbook for school partnerships: don't enter asking teachers to make time for your research questions. Instead, deliver a tool they need immediately that solves their real problem, embedded in direct service. Once trust is built, schools open up to research questions. Many of his graduate students and postdocs found their research directions reshaped by real problems schools brought to them—for example, how to interpret assessments for English-language learners.
— Jason YeatmanAutism research stumbles over sensory barriers and extreme participant diversity
Karen Parker outlined challenges in autism research that don't appear in typical-development studies. Many children's sensory sensitivities make brain imaging scans impossible; extensive family preparation is required beforehand. More difficult: age spans from three to seventeen years but requires a single outcome measure. The autism population is extremely heterogeneous—the same cohort might include both children with intellectual disability and a child with an IQ of 150, making trial design extraordinarily complex.
— Karen ParkerA shared registry compressed autism recruitment from years to six weeks
Karen Parker described how Simons Foundation's Spark project—a national autism family registry—compressed recruitment timelines. Recruiting hundreds used to cost approximately $5,000,000 and take years. Now a single survey-based questionnaire draws thousands of responses in six weeks, and the system directly routes families to trials matching their clinical criteria, dramatically accelerating research.
— Karen ParkerA reading assessment platform serving 1.2 million students has become a ready-made research cohort
Jason Yeatman explained that his reading assessment platform now serves approximately 1.2 million students across nearly 3,000 schools in daily use, accumulating continuous longitudinal data. He aspires to make it foundational infrastructure: the next generation of students and postdocs won't need to build cohorts from scratch but can pose the most important research questions. If they want to study high schoolers, years of data from elementary school forward may already be in hand.
— Jason YeatmanAI can fabricate references so perfectly they'll pass funding review
Karen Parker shared a real lesson: a colleague used AI at 2 a.m. to draft a funding application; the AI-generated references were complete with journal, volume, and page numbers—all fabricated. Her conclusion: AI is a powerful research tool but can't replace thinking. Jason Yeatman added that he and his daughter built an interactive math website combining open-source curriculum with AI programming, but as a researcher he's more concerned that AI in education is advancing too fast with no mechanism in place to measure its effects.
In their own words · checked verbatim
Why would you run a large RCT to show something is better if that's not going to drive profit?
Jason Yeatman7:14
As we're working our way up the academic ranks, you want to show that you're last author.
Karen Parker12:10
we're all writing grants at 2:00 in the morning, and these were all hallucinations, but they had had all of the authors that had the issue.
Karen Parker25:36
it's like the dog has to wag the tail, not the tail wag the dog.
Karen Parker26:18
I think the magic happens at the interdisciplinary boundaries.
Karen Parker28:55
Meeting the needs of every learner.
Jason Yeatman29:17
Figures
| Reading assessment platform users | approximately 1.2 million students | 23:10 |
| Research-practice partnership schools | approximately 3,000 schools | 18:40 |
| Pre-Spark autism RCT recruitment cost | approximately $5,000,000 | 21:57 |
| Spark survey response time | thousands of responses in 6 weeks | 22:53 |
Glossary
- RCT
- Experimental design comparing intervention and control group outcomes—the gold-standard research method.
- Spark
- Simons Foundation's national registry for recruiting autistic families into research studies.
How to listen
Entrepreneurs, education product managers, and researchers concerned with evidence-based mechanisms in edtech, academic-to-practice translation, and autism research methodology.
The opening host introduction and closing quick-fire Q&A are low on information and can be skipped.