Field-Driven Research Synthesis: Understanding the Character of Innovative Interdisciplinary Exploratory Research

Equitable Co-Designed Learning and Practice

Creating contextually rich learning environments: Ross (AI and Robotics), Rebecca and Colleen (Augmented Reality), Lorna (Accessibility and Learning), Brett (Accessibility and Simulations), Ying (Collaborative Learning and AI), and Roxanne (AI)

PIs wanted to develop more equitable technologies and experiences for learning. Interdisciplinary, exploratory research allowed them the time to investigate contextually rich learning environments using equitable co-design approaches that a) were context driven, customizable, and took learner variability into account; b) incorporated inclusive approaches and frameworks for teaching and learning; and c) flexibly addressed challenges in co-design.

Opportunities to investigate contextually rich learning environments aligned with PIs’ goals to develop emerging technologies for learning that are context driven and equitable. For Ross’s AI and Robotics project, Using AI to Focus Teacher-Student Troubleshooting in Classroom Robotics, this meant discovering that a “one size fits all” approach was not going to work for the diverse users he and his team were aiming to serve:

From the beginning, we intended several different environments, including online environments, remote classrooms, inner city public schools, suburban ones. We intended to find commonalities among these environments that we could design our system around. It turns out not to really work that way. What we ended up observing was that the ways in which people teach, especially troubleshooting, are very much shaped by the environments in which they are teaching.

Ross’s team learned that developing equitable emerging technologies for learning required addressing user-directed customization and learner variability across diverse learning environments:

We would have designed a system that would only have worked in a few places—and it probably would have been the places that didn’t need it—had we not looked broadly and with intent to do something helpful. There’s a bit of a subtle context shift from looking for the common element that serves everybody to trying to create a system that everybody can use in their own way. So, one thing we did learn is that the diversity of users has a huge impact on the usage of systems.

PIs we interviewed pivoted toward the idea that a technology is only as useful as users find it to be; this represented a shift in their thinking of equity-focused design to a more user-centered perspective. In Ross’s case, taking an exploratory approach allowed his team to shape technology systems to meet users’ needs across diverse learning environments.

PIs used these context- and user-driven design adjustments to customize spaces for teaching and learning. Their approaches included adopting inclusive design frameworks, such as Universal Design for Learning (UDL), to support diverse users, as Rebecca describes below:

When I was a high school teacher, we would go to amusement parks, go on roller coasters, and collect data. But we only had a few commercial devices, so we started using smartphones. My husband developed a tool that takes all the sensors from smartphones and turns it into useful data. You might think, “why is that even important?” The thing is, if you can plot your data in real time—and you can embody that motion as a physics or a math teacher—you can now help students understand things like rates and mathematical modeling. You start to lay the foundations for things like calculus in a way which is completely hands-on, whole body.

In Lorna’s project, UDL principles were used to support teaching and learning by developing a VR technology to build ASL fluencies with graduated levels of support for deaf individuals and their families:

Imagine that you have parents who have a baby that was just identified as deaf. We can imagine that as new parents, they might not have the time or the easy access to arrive at in-person classes on a scheduled timeline. So, they might use resources that are available like YouTube or books. But what if they could learn American sign language from the comfort of their own home using an off-the-shelf VR device at their own pace? And receive feedback while they’re learning? Simulating in-person instruction with a deaf signer—that’s our use case that drives this work to get more people signing.

PIs also incorporated UDL principles into their co-design work to support multiple modes of user engagement, as Brett described in the Accessibility and Simulations project Inclusively Designing Sensory Extensions for STEM Inquiry Learning:

The simulations themselves are built in a web environment and are also built to be offline, downloadable, so that they can be used in a variety of contexts—especially rural contexts and in places with low internet connectivity. I work in inclusive design and accessibility trying to figure out how to scale both the design and implementation of accessibility. Many people do not have accessibility in their products, and they definitely don’t do it from the start because it’s hard. This idea of trying to design multiple modes at once, correlating how things sound, the words that are spoken, the timing of things… You can design all those, but if they overlap, or they fire at the wrong times, it’s going to obscure your pedagogy at the end.

Brett was particularly interested in providing multiple ways for users to engage with content. He aimed to ensure that accessible simulations could be seamlessly integrated into users’ learning.

In addition to taking environmental context into account in the design of emerging technologies for learning, PIs discussed the role of teacher experiences as critical to equitable co-design experiences. In the Collaborative Learning and AI project Building a Teacher-AI Collaborative System, Ying discussed her team’s commitment to developing equitable technologies for learning:

One of the primary goals we’re trying to address is to find the right balance between involving teachers but not increasing the workload. If you think about a spectrum, on the one hand, the teachers manually create instructional materials. And on the other hand, AI generates everything for them. This Teacher-AI collaboration approach is really finding a sweet spot within the spectrum. What we learned from our engagement with teachers is that there’s no one sense or solution for the right balance. We saw a wide variety among teachers.

Similarly to Ross, Ying reiterates the importance of accounting for user variability—in this case, the context of teacher experience—to support user-centered customization:

So, it’s going back to how to develop a system that is not too complicated but can accommodate all these different needs. We want it to offer autonomy to seasoned educators but also provide stronger guidance and support for those who are newer to the profession. For us, the whole purpose of this proposal is to offer customization. But on top of that, we need to allow teachers to customize how much customization they need in this system.

However, the creation of co-designed spaces for teaching and learning sometimes resulted in challenges that required pivots in approach. Flexibility to change course to meet emergent project needs was built into NSF EXPs. Brett described how this flexibility afforded him the freedom to address the challenges of losing key partners:

We had our expertise in multimodality, inclusivity, and virtual design, and we lost the person on the physical side. Given the timeline between the pandemic, that’s pretty rough. Our ability to work with partners was actually a big issue across the board. A lot of research partnerships with academia, high schools, and districts started to fall through. And they’ve been slow to get back. I think that hit us, and probably many other researchers, pretty hard—especially those interested in co-design. The exploratory nature of the RETTL funding was very powerful because we had a little bit more room within the questions. The fact that it was exploratory in nature meant that the research questions that we’ve set out to solve were not so bounded. That meant we could make a few pivots within what expertise we still had.

Similarly, situational shifts forced Roxanne to make pivots in her original project design—which was to develop a ChatGPT-like tool to help teachers scaffold questions for students— when ChatGPT was publicly rolled out ahead of her project timeline:

Our Principal PI was aware of GPT before it became ChatGPT, but it wasn’t commercially available. So, when ChatGPT came out, he knew it was coming but nobody knew what it was going to mean. The question became, “How do I parse this design challenge?”

The flexibility afforded by NSF programs such as RETTL enabled PIs to address sizable emergent challenges—such as the pandemic or launch of a competing technology—while keeping their commitments to co-design.

Enabling equity in co-design and practice: Sheryl (Accessibility and Learning), Zachary (Virtual and Augmented Reality), Ying (Collaborative Learning and AI), Lorna (Accessibility and Learning), and Chad and Bill (Data Analytics and Simulations)

Enabling equity in the co-design of emerging technologies for learning required a) taking users’ intersecting identities and needs into account when making design decisions, b) engaging in co-design practices with practitioners, and c) drawing on PIs own lived experiences to inform their values and mission-driven work.

PIs felt that exploratory approaches were essential for understanding users’ intersecting identities and needs through the exploration of disability and other plural and mutually occurring identities, as Sheryl explains:

I really wanted to make sure that people with disabilities had access to microcomputers, as we called them back then. A lot of projects, particularly if they focus on disability, narrow it to specific disabilities. But they don’t consider intersectionality. I’d love to make the point that once you choose a group, you need to think about intersectionality. You need to think about the other characteristics they might have. It could be another disability. Or language, or culture, or gender and racial minority status that could impact their success and getting access to technology.

When co-designing with users, PIs took intersectionality into account and used exploratory approaches to actively seek input from underrepresented voices.

This included co-designing with partners to inclusively address the needs of diverse users. Zachary, in discussing the Virtual and Augmented Reality project Using Augmented Reality to Enhance Attention in STEM Learning, shared these inclusive practices:

I think a key part is the differences in perspectives and backgrounds that the co-designers themselves have brought to this space. Some of them are interested in artistic visual design and digital design. Others are in computer science and really interested in the algorithm itself. Others are interested in working in neurodiversity and so they bring this perspective of how they can best develop these kinds of tools with an eye towards universal design to create tools that are inclusive and accessible.

Zachary highlighted the diverse perspectives and expertise of the co-designers involved in the project. This diversity enriched the project, as each co-designer brought a unique experience and viewpoint, to develop tools that are inclusive and accessible to a broad range of users.

For many PIs, exploratory approaches also supported the creation of infrastructures for equitable co-design with practitioners that lowered barriers to participation and uplifted practitioner expertise and autonomy. Ying shared how her team engaged in equitable co-design with practitioners:

The goal of this project is to bring back the teacher’s expertise because they are central to students’ day-to-day interactions, and they bring so much knowledge. We want to support them to go beyond simply being consumers to actively shaping the technology’s development and usage. We see this as an opportunity to expand the traditional co-design and participatory design research.

We appreciate the co-design method to develop AI materials, but one limitation is that you can only work with a small group of educational practitioners, and once the product is developed, they do not have the autonomy to make modifications of those products. We hope that by creating a collaborative system, we’re extending the time where the teachers can stay involved… This could be an opportunity to make the materials even more tailored to the instruction.

Another opportunity related to this point would be lowering the technical barriers for teachers to participate in this development. If you think about traditional development, AI-based learning materials are controlled by people with technical backgrounds, and teachers don’t normally have the opportunity to do it themselves. With Generative AI and with these more accessible systems that we’re going to develop, we’re really bringing the teachers back to this so that AI is no longer dominated by a small sector of professions.

Expanding traditional design approaches to create a collaborative co-design system that included teachers as co-designers was a central element in Ying’s research. Essential to this equity approach is centering teachers in the development of the tools that will eventually end up in their classrooms.

Finally, adopting exploratory approaches allowed PIs to draw on their own lived experiences to inform their research and community engagement. For example, Sheryl shared that her background as a working-class first-generation college student turned mathematics teacher helped her connect with research participants and co-designers from diverse backgrounds:

I feel like I could pull this off because of my diverse background and my comfort. My dad was a used car salesman. So, I was always around these mechanics and welders and watched my brother be encouraged to do that kind of stuff and me just stand back. Finding my place and going on to college was a big step in my family. I was a very good student, and I knew that. But it was only people in school that thought I should, including my math instructor. He said, “Well, you going to college? You’re so good at math. You ought to do this.” So, I became a math teacher.

Sheryl’s affinity for research at the intersection of disability, technology, and education was enabled by her interpersonal skill set born from her lived experiences. Sheryl credited her salt-of-the-earth, blue-collar upbringing as laying a foundation for attending to the needs of marginalized communities with empathetic understanding.

In addition to the impact of PIs’ lived experiences on their research, engaging in exploratory approaches allowed PIs to design their research to be values and mission driven, as Lorna shared:

One of the missions of the Motion Light Lab is that every deaf child in the world has a right to accessible language, which means signed language—whatever is indigenous to their community and their culture since there’s many different sign languages across the world. So really, Motion Light Lab’s goal is to end language deprivation in which deaf children do not have access to language from birth.

Motion Light Lab’s mission was both personalized and equity centered—addressing barriers to ASL access and the marginalization of ASL as a language.

For other PIs we interviewed, their missions were aspirational and aimed to arrive at agentic and impactful outcomes for users of their technologies, as Chad described in his and Bill’s Data Analytics and Simulations projects, Inquiry Space and Data in Space and Time:

We are implicitly and explicitly founded on the principles of encouraging exploration within STEM students’ ownership and inquiry. Those have been at the heart of everything that we’ve done. Science, engineering, mathematics are the mediums that we work within but not necessarily an end in themselves. Really, the end all is empowering students to recognize that you can ask and answer questions that are unique and important to you, and that you can use problem-solving skills in lots of domains. And it turns out that science, math and engineering are really good training grounds for developing those skills that are usable in all different aspects.

For Chad and Bill, shaping how people engage with ideas and data to address knowledge and skills gaps in scientific sense-making was central to their organization’s mission.

Translating the mission and vision for their specific projects to address broader societal goals was also deeply related to PIs’ hopes for the future impact of their work, as we detail in the next section.

Back to topNext section