Reading up on the Futures Forum EdTech Tools Competition has prompted me to think about the best ways that different organizations can contribute to the field of learning engineering, and in particular, a more vibrant and open education research ecosystem. I’ve come to appreciate that this is a complex, multi-faceted problem. This post is my attempt to organize my thinking around how different types of EdTech organizations can feasibly make meaningful contributions to the field.
An “Enabling Environment” for Education Research
While reading up on this topic, I kept coming back to the concept of the enabling environment, an idea I’ve encountered in my global health work. The term refers to conditions (such as policies, laws, institutional capacities, incentives, and even norms) that do or don’t contribute to a goal or social good. EdTech sits in a complex ecosystem with many different actors, including educational institutions, governments, companies, foundations, and nonprofits. While all of the different actors can contribute something to the field of learning engineering, what each organization can best contribute depends on its role. This may seem obvious, but I think it’s a useful framework to figure out how different organizations can pitch in. We can see this with a concrete example of how EdTech organizations with different business models can make fundamentally different contributions to education research and learning engineering.
Roles and Responsibilities of EdTech Companies
Some EdTech providers (think Duolingo and Udacity) provide education services directly to end-users. These companies, as well as universities, have B2C business models. When it comes to student data, these organizations are gatekeepers. They have a lot of say over how student data is used. By contrast, other organizations are service providers for the gatekeepers (think Canvas working with a university). They have B2B business models. As such, these organizations have obligations to process and store data for the gatekeepers. Service providers’ permitted uses of the data are contractually defined and likely include the functions necessary to provide and improve the service itself. In a situation like this, the decision to conduct publishable education research or share data likely lies with the gatekeeper, while the service provider’s powers are more limited.
These gatekeeper and service provider categories aren’t arbitrary: they are products of the system. For example, the EU defines different responsibilities for data “controllers” (gatekeepers) and “processors” (service providers). When it comes to contractual norms, most educational institutions (reasonably) expect to retain control over their own student data. The recent acquisition of Instructure provides a good example of this: when some university staff raised concerns about the impact of the acquisition on student privacy, the company pointed out that its existing agreements with universities set legal limits on how it can use and share data.
Why go into all of this detail? Because it demonstrates that not every organization with access to student data has equivalent decision-making power with respect to those data. This feeds into the actions different organizations can take to bolster education research.
What Gatekeepers Can Do
Gatekeepers have great power, and with it great responsibility. Here are three things that gatekeepers might do to contribute to a better education research ecosystem.
- Set up policies and systems that promote research. An organization’s stated priorities have a big impact on what its employees do. Gatekeepers can prove that they value education research through their actions. This could include creating a transparent and fast (yet robust) ethical approval process, adopting tools that simplify experimentation and data de-identification, and providing incentives to publish research findings.
- Be transparent with users about sharing de-identified data. Not all gatekeepers explicitly tell their users that they may share de-identified data for research purposes. For gatekeepers interested in taking this step, the Khan Academy privacy policy provides a nice example.
- Publish de-identified data. This is the big one! Published de-identified platform log data is the exception, not the rule. In a quick search for de-identified datasets from big EdTech companies, I didn’t find much. For many companies, I suspect that the perceived risks of sharing their data outweigh the rewards. This ties back directly to the enabling environment: organizations don’t bother to share in the absence of policies, norms, and incentives that promote sharing.
What Service Providers Can Do
Service providers aren’t in the same position as gatekeepers, but they can still make a big difference! Here are three things service providers might do to contribute to a better education research ecosystem.
- Develop better tools for experimentation. Faculty and staff who work at gatekeeper organizations are usually excited about doing research. With better tools to conduct education research (particularly tools that facilitate experiments), more research might be happening. Terracotta looks like an interesting example of an organization taking this approach.
- Create better tools for safe data de-identification and sharing. Data de-identification is not a trivial problem. Providers could create tools for gatekeepers to de-identify data, along with tools to validate that de-identified data is safe to share. This might be particularly useful for small universities and direct-to-consumer startups that don’t have dedicated staff who are up to the task.
- Collaborate closely with gatekeepers on research studies. Service providers have a deep understanding of their technology, gatekeepers understand their students and content, and academics can bring in research expertise. These powers combined could lead to great new insights. BIRI is a recently-launched example of such a collaborative research initiative.
What Everyone Can Do
There are some things that most actors in the EdTech space can do, regardless of their role. Here are three things we can all do to contribute to a better education research ecosystem.
- Adopt an open access policy. Such policies are commitments to making research papers (and often the underlying data), publicly accessible. Many universities, governments, and foundations already have such policies. Why not EdTech companies as well?
- Work towards consensus metrics for evaluation. No metric is perfect, but some consensus metrics could be hugely beneficial. For example, standard survey questions that everyone agrees to use for collecting learner feedback would be so useful for making apples-to-apples comparisons of different products and services.
- Work on the enabling environment itself. This could come in many different forms, such as funding innovative new ideas, strengthening and joining existing research collaboratives, supporting and contributing to data repositories, or lobbying for institutional and government policies that promote education research and data sharing.
I’m confident that there are many more actions that companies can take to contribute. Issues of research, data sharing, contracts, and privacy concerns involve a host of legal and ethical considerations. Please keep in mind that I’m not a lawyer and any organization looking to take these steps should seek professional guidance.
To read other musings on EdTech, check out Empiricism.