Every EdTech product comes with a claim: this improves reading, this raises attainment, this changes lives. Some of those claims are well founded. Many are not. For schools and families spending scarce money and precious time, telling the difference is genuinely hard, and it’s one of the reasons I do the research I do.
I sit in an unusual position here. I evaluate technology, sometimes for the companies that make it, and I also care deeply about research integrity. Holding both of those honestly, rather than pretending the tension doesn’t exist, is the whole game. This article is about how that’s done well.
The uncomfortable truth about EdTech evidence
Here’s the problem in a nutshell. The organisations most motivated to fund evaluations of a product are usually the companies that make it. Public research councils tend to avoid funding evaluations of single products, to steer clear of handing any one company a taxpayer-funded advantage. So philanthropy and companies fill the gap, and that creates grey areas.
The sharpest of those grey areas is what happens to negative results. If a study funded by a company finds the product doesn’t work, does that finding ever see daylight? Or does it quietly become an internal report? A company may be tempted to protect its brand by limiting unfavourable findings, or even by prolonging a study in the hope the numbers improve. Researchers, especially early in their careers, feel their own pull towards positive, publishable results. None of this requires anyone to be a villain. It’s just what happens without clear rules.
What good practice actually looks like
This is exactly the problem I worked on with an international group of colleagues in a commentary for Child and Adolescent Mental Health (Kucirkova et al., 2025). We argued for clear, shared guidance so that collaborations between researchers and technology companies can be both useful and trustworthy. A few principles sit at the heart of it.
Contracts should be honest about difficult outcomes. The agreement should say, up front, what happens to negative or inconclusive results, who owns the data, and when findings will be published. Sorting this out before the study, rather than after an awkward result, protects everyone.
The purpose of the research should be transparent. There’s a real difference between research funded by a company to improve its own product, where keeping some findings internal can be legitimate, and independent evaluation intended to tell the wider world whether something works, where the ethical standard is to publish everything: positive, negative and neutral. Muddling the two is where trust breaks down.
Everyone with a stake should be involved. That means the children, parents, teachers and professionals who actually use a tool, not just the company and the academic.
And the field needs shared infrastructure. We argued for formalised data-sharing and, ideally, a shared, anonymised data repository, so that findings can be compared across studies rather than each one sitting in isolation. That would help funders, researchers and policymakers make genuinely evidence-based decisions.
Practising what I preach
Principles are easy to write and harder to live by, so I try to hold myself to these in my own work.
When I studied a reading device in my OrCam research, the study was funded by the manufacturer, and the paper says so plainly. When I evaluated the Lili Lamp, I ran it as an honest case study and genuinely welcomed a family who chose not to take part, because people feeling free to say no is a sign the research isn’t coercive. And in the commentary itself, my co-authors and I declared our consulting relationships with EdTech organisations openly. Disclosure isn’t an admission of bias. It’s the thing that lets readers weigh the work for themselves.
Why this matters to you
If you’re a school leader, a policymaker or a parent, the practical takeaway is a set of questions worth asking of any EdTech claim. Who funded the research? Were negative results able to be published? Was the evaluation independent, or was it designed to support the product? Is the underlying data available, or just the headline? Good products stand up to those questions. The rest tend not to.
This commitment to honest evaluation is the foundation of my research, consultancy and expert witness work, and it’s why I collaborate internationally on getting it right. Technology can do real good for dyslexic and neurodivergent learners. The evidence for it just has to be trustworthy, or it’s worth nothing at all.
Questions people often ask
Can research funded by an EdTech company be trusted? It can, but only with safeguards: transparent disclosure of funding, clear rules about publishing negative results, and honesty about whether the study is product development or independent evaluation. Funding disclosed openly is far more trustworthy than funding hidden.
Why don’t public bodies just fund all the evaluations? Public research councils generally avoid funding evaluations of single commercial products, to avoid giving one company an unfair advantage. That leaves a funding gap that companies and philanthropy fill, which is why clear guidance for those collaborations matters.
What should I ask before believing an EdTech claim? Who funded the study, whether negative results could have been published, whether the evaluation was independent, and whether the underlying data is available. Robust products withstand these questions.
Isn’t it a conflict of interest for you to evaluate products you’re paid to study? The tension is real, which is why I disclose funding openly, welcome participants declining to take part, and follow the shared principles my colleagues and I have argued for. Transparency, not pretending there’s no tension, is what protects integrity.
References and further reading
- Kucirkova, N., Cherner, T., Dube, A.K., Pasquarella, A., Pitchford, N. and Ross, H. (2025) Commentary: Suggestions for guidance by academics who collaborate with digital companies, Child and Adolescent Mental Health, 30(3), 292-295. https://doi.org/10.1111/camh.70021
- International Centre for EdTech Impact (2025) Case Studies 2025: Celebrating Academia-Industry Collaborations in EdTech, Stavanger, Norway. (Includes Dr Ross’s evaluations.)
- Ross, H. (2024) Impact of OrCam Learn on reading comprehension, Support for Learning, 39(3), 95-103. https://doi.org/10.1111/1467-9604.12474
See the full list of publications.