By Danielle McNamara, Principal Investigator and Executive Director of the Learning Engineering Institute at Arizona State University
I’ve spent the last few years helping lead a project called Learning@Scale at Arizona State University, which aims to understand how students learn across hundreds—or even thousands—of courses and contexts. We bring together large-scale data from learning management systems, online platforms, and student-facing tools to uncover patterns, evaluate innovations, and ultimately improve student learning outcomes.
From the outside, you might think the hardest part of this work is technical: dealing with messy data, integrating across systems, or building analytics that protect privacy. And yes, there are technical challenges—but technology is often not the biggest obstacle. Institutional structure is.
Despite the increasing availability of tools like SafeInsights—a secure, privacy-preserving research infrastructure designed to support ethical research at scale—I often find myself slowed by institutional processes, priorities, and concerns. The barriers are rarely about whether we can do privacy-preserving research; rather, they’re about whether the institutional environment and structure enables us to do so.
Institutional Complexities
Every step of the Learning@Scale project involves careful collaboration with IRBs, IT security offices, registrars, legal counsel, and institutional research offices. Each has its own policies, definitions of “identifiable data,” and interpretations of FERPA or ethical risk. These actors aren’t opposed to research per se—but in a fragmented governance environment, the default institutional response is often caution or delay, especially in the absence of clear precedent or shared frameworks.
The technical work—aggregating clickstream data, deploying privacy-enhancing technologies, running large-scale models—feels easy compared to aligning the internal politics and governance of educational data.
Understanding Institutional Hesitations
It’s worth unpacking why institutions continue to hesitate and delay, even when we offer secure methods and good intentions:
Institutional stakeholders (IRBs, data stewards, legal teams) are trained to protect the students and university from legal or reputational harm. When presented with something new—like a privacy-enhancing enclave or a novel data-sharing protocol—the default posture may be caution or delay, particularly in absence of precedent.
Many universities lack a centralized data governance framework for research. Instead, we navigate a patchwork of stakeholders—each of whom can veto a project but none of whom feel ownership to move it forward.
Educational research often sits at the intersection of practice and scholarship. When is using student data for “course improvement” different from using it for publication? That ambiguity makes institutions uneasy, especially when students haven’t explicitly opted in.
“Learning at scale” can raise concerns about surveillance. Even when we take every precaution, the idea of automated analytics applied across populations can raise red flags about fairness, transparency, and control.
While researchers focus on evidence and improvement, institutional systems are often oriented around compliance and risk avoidance. Without leadership buy-in, and sometimes even with it, innovation may stall despite potential benefits.
SafeInsights: A Promising Pathway
I am incredibly excited about initiatives like SafeInsights, which provide trusted research environments (TREs) for educational data. Data enclaves are designed to allow researchers to analyze anonymized, well-governed datasets under controlled conditions—without extracting or exposing sensitive information.
They represent an important shift: don’t just build the tech—build the trust infrastructure too. SafeInsights works not only at the technical level, but also brings together cross-institutional partnerships, legal frameworks, and shared norms for ethical research.
And yet—for such tools to reach their full potential—broader institutional alignment is needed to avoid delay and pushback.
We Need Organizational Change, Not Just Technical Innovation
As someone working in the weeds of Learning@Scale, I want to see educational research treated as a public good—not a liability. But for that to happen, we need to evolve how institutions think about research:
SafeInsights is a big step forward. Building research infrastructures is vital to furthering our understanding of how to enhance learning at scale. But infrastructure won’t solve the problem alone. Expanding research at scale requires parallel investments in organizational capacity, cross-functional communication, and a shared understanding of what ethical innovation looks like in education.
If you’re working on similar challenges—or facing your own IRB battles—I’d love to hear from you. The more we talk openly about these barriers, the more likely we are to build the shared infrastructure and institutional collaboration that learning at scale truly requires.
This project is supported by the Institute of Education Sciences, U.S. Department of Education, through Grant R305N210034 to Digital Promise. The opinions expressed are those of the authors and do not represent views of the Institute or the U.S. Department of Education.