The reflective practice of design teams

What I read

In this article, the authors apply Donald Schön’s approach to reflection in action to observe and map, in detail, the efforts of two design teams working through a problem. They conclude that framing, early in the process, plays an important role in success of design problem solving, and that a detailed video and text-based protocol, and detailed modeling, provide insights into the design process.

The researchers describe how design is often a collaborative activity, with teams of people working together to solve a problem. They indicate that there is no established protocol for studying this type of work, and that a part of this paper will introduce a new method.

They then introduce the nature of Donald Schön’s work on reflective practice. Schön rejected the idea that designing is a technically rational activity; that design tasks are unique, and therefore there must be a design process, but the design process is based on a “kind of knowing [that] is inherent in intelligent action”—but that this form of knowledge is difficult to convey to others, particularly to students who are learning to be designers.

Next, the authors describe their study. They had a unique opportunity to study design teams working together: a week-long competition, where a multidisciplinary set of students worked on a project (in this case, designing a robot that must perform a particular task.) They selected two teams to observe, and focus on the first two days of the project, which included more designing than building.

Schön’s work is to be used primarily in a rich evaluation of the data that has been generated. The authors break his theory into four activities: naming, framing, moving and reflecting; this plays out as “The designers start by naming the relevant issues in the design situation, framing the problem in a certain way, making moves towards a solution, and reflecting on those moves and the current frame.”

They indicate how they coded the data that was generated, using a visual form language that maps these activities. The transcript data is divided into episodes, which then become containers for the various models of activities. One of the most visual parts of the modeling is focused on framing: literally drawing a box around the problems or subproblems, indicating a place where a unique element is discussed. They call this “something to hold on to and to focus on while designing.” They provide a brief example of the coding, by providing a small transcript excerpt and then showing how that maps to an example model. They also note that the transcript has been translated from Dutch.

Next, they provide a detailed analysis of the two teams that they studied. They include a narrative of what the team did at a certain moment in time, interspersed with quotes. They show the entire model for that part of the experience (stretching as long as two hours.) The first team’s model contains no frames and very little reflection: it shows a lot of naming, a little moving, and a frame only appears towards the end of the allotted time necessary for design. Even within that frame, there is still nearly no reflective activity. By contrast, the other team’s model includes a more diverse model, including a number of frames, reflective cycles, and moving.

The authors then discuss their findings. They indicate that, first, using the four-stage mechanism provides visibility into “large-scale strategies.” The approaches the teams used were extremely different, and the authors note that the team with no real frames did not actually complete the design problem at all, while the other team was successful (this was a competition, and that team came in first.) The team that did not succeed primarily focused on naming, and never moved into a deeply reflective set of moves, like drawing. Their time was spent discussing things. The other team spent the majority of their time “moving”—actually changing the design content itself. Additionally, they spent more time reflecting, and reflected throughout the project, while the less effective team reflected only towards the end of the project.

The authors discuss that “the description of team designing in episodes and categorizing them into the four activities provides a good insight in the course of the team design on a project level.” It is a successful framework for discussing how teams design, and they indicate that it can be useful during design education. They note that “all the interesting moments occurred when the team makes a transition between frames. The view the team creates about the context they are working in seems very important in team designing.”

What I learned and what I think

This is a strong illustration of a theory. Most of what I’ve read and explored from Schön has been focused on transcription analysis and discussion, which I’ve found to be really fascinating and valuable, but this is the first time I’ve seen a real modeling attempt at it, and the little diagram of framing and reflecting is very effective in demonstrating exactly the theory: reflection in action, or in practice. In some ways, it’s actually more effective in illustrating reframing in action, which is certainly part of a reflective practice, but is also an almost innate or entirely tacit activity. I’m not sure many designers ever say, or notice, that they reframed a problem or a strategy. I know I do it, but I rarely (ever?) think of it that way. I see a big jump in my own constraint-shifting during the creation of a first visual model, and as I think about the move from interpretative research synthesis to actually making something, it’s often been when I’ve had heads-down time to sketch a small model-based story. That’s always sort of unlocked the next steps.

I seem to find my way back to constraints and flexibility a lot, in this space. There’s a subtlety to what’s happening when constraints are scaffolded and then flexed that doesn’t come across in the language. “Shifting constraints” and “reframing” seems so grandiose and purposeful, but it’s much lighter-touch than that. This research focused on strategy—high level moves and boundaries—and at this level, the framing is a pretty blunt instrument. The protocol focuses on discussion framing, which is pretty different than ideation framing. The constraints I often reflect on are the ones that stand up and tear down in real-time during form giving, not verbal form discussion.

Actually, as I reflect a little more on what I read here, the reframing is a “what should we do” form of overt focus aiming, which I think is different than work reframing. It’s the activity/method/work approach selection, which spins out of control precisely at the experience level they studied: students. It’s also the fear of activity that clients have who aren’t familiar with the process, and need to see a process through its conclusion before trusting it. All useful and poorly researched, and this paper contributes effectively to that, but still a gap in the actual detail, zoomed in practice. Another way of maybe saying this is that the team that did well did well because they agreed on the frame of what to do, or what to talk about, or what design to select, but not on the frame of the actual content itself.

I’m not entirely sure that the sort of framing I’m thinking of can actually be collaborative at a strategic level. It needs to be seen or modeled. It has to play out in real time as an artifact, not as a conversation. I think it’s actually why real-time collaborative canvases are actually effective, and maybe more effective, in doing the design work as compared to doing the strategy work. There’s something in here about the net-positive of a remote working environment; it doesn’t have to be remote—real-time Figma or peer coding can happen in person—but it also works fine in distance. I’ve only seen this play out a few times with my designers, but it’s the joint work with talk and not video on, where both (and not more than both) participants are actually good at both the tool and design itself.

Back to the article; the modeling visualization is great. It’s the missing part, or the wrong part, of the other design-as-science or creativity-can-be-measured articles. This is “just” a way of showing what happened; it’s pre-analysis, and so it’s not making any claims, and I think that’s why I find it effective. Take what you want from it; the authors make it explicit towards the end, and probably either directly or indirectly led me to a conclusion they wanted, but it’s not making a declarative, and I’m just fine with the scientific approach to a sample of two. I think it’s actually a form of discourse analysis, not science. Visual discourse analysis—is that a thing? It’s exactly this thing.