How to Grade a Physical Computing Project Without Becoming an Engineer

A five-dimension rubric for grading interactive product prototypes in design studios, with scoring criteria, a worked example and guidance for handling demo-day failures.

The first time a studio of interactive product prototypes gets graded using the wrong frame, the results feel off. Watching each demo and asking "does it work?" seems reasonable, until the students who built the most ambitious projects get the lowest grades because their prototypes crashed during the crit. Meanwhile, the students who played it safe with a simple LED that turned on when you pressed a button walk away with A-minus scores. That outcome points to a structural problem in the grading, not a problem with the students.

What goes wrong is that an engineering grading frame gets quietly imported into a design studio. "Does it work" is a legitimate question in an engineering course. It is a deeply misleading question in a design course where half the point is ambition, failure, iteration and the willingness to attempt something that does not reliably work on the first try.

This post presents a rubric with five dimensions, none of which require evaluating code quality and none of which reward "it worked smoothly on demo day" as the decisive factor. It is designed for ID instructors who feel under-qualified to judge physical computing work and do not want to fake expertise they do not have.

What follows is the rubric itself, the reasoning behind it and some examples of how to apply it in edge cases.

Why the obvious rubrics do not work

Before getting to the rubric, it is worth naming the frames that seem reasonable and are not.

"Does it work." Tempting, concrete, easy to score. It rewards risk aversion and punishes ambition. A student who spent three weeks wrestling with a load cell to make a weight-aware object will be graded below a student who put an LED strip on a button. This is the wrong incentive.

"How clean is the code." Most ID instructors are not qualified to judge this, and honestly, that is fine. Also, the students who write the cleanest code in a physical computing studio are not usually the best designers. They are the ones who already knew how to code. You would be secretly grading prior background.

"How polished is the final photo." A respectable photo is something design students can fake without the object actually functioning. You can spot the fakery if you know what to look for. Most of us do not.

"How well does it match the original concept." Reasonable, but it rewards unchanged thinking. Some of the best studio projects pivot hard in week four. The original concept was wrong. The student noticed. The rubric should not punish that.

The rubric below is designed to sidestep every one of these traps.

The rubric

Five dimensions. Each scored on a four-point scale from Developing (1) to Distinguished (4). No code evaluation anywhere.

Dimension 1: Concept clarity

What is the object for and why does it exist.

This is the question you answer before looking at the physical thing. Does the two-page rationale identify a real human action, a real context and a real reason the object belongs in the world? Or is the rationale a thin paragraph that could have been written about any product?

Distinguished (4). The rationale identifies a specific action in a specific context, names the user's emotional or functional need precisely and explains why a physical object (as opposed to an app, a service or a paper note) is the right answer. The student can defend the concept against an obvious alternative.

Proficient (3). The rationale is clear and coherent but one of the three elements (action, context, need) is softer than the others. The student can mostly defend the concept but may stumble on the "why physical" question.

Developing (2). The rationale describes a plausible object in a plausible situation but feels generic. Could have been written without the student having spent six weeks on it.

Emerging (1). The rationale is a description of what the object does, not why. The concept collapses under any critical question.

Note for graders. The phrase "why physical" is the one worth leaning on hardest. If a student cannot articulate why the thing has to be an object instead of an app, the concept is not finished. This is a design question, not a technical one.

Dimension 2: Interaction quality

How does it feel to use.

You evaluate this by interacting with the object yourself at the critique, not by reading about the interaction in the rationale. Trust your hands. Trust your observations of other people using it.

Distinguished (4). The interaction feels deliberate. The object's response to the human action is legible without explanation. Someone who walks up to it can figure out what it does within about ten seconds. When the interaction fails (and some will), the failure feels meaningful rather than arbitrary.

Proficient (3). The interaction is coherent once explained. A first-time user probably needs one sentence of guidance. The object's response is consistent and feels thought-through.

Developing (2). The interaction works but feels arbitrary. The mapping from action to response has no clear reason behind it. The object could respond in several other ways and none would be worse.

Emerging (1). The interaction is either broken during the critique and cannot be recovered, or it works but the student cannot explain what the user is supposed to do.

Note for graders. Do not confuse "broken right now" with "Emerging." A student whose prototype fails during the crit but who can articulate exactly why, recover it for a second try or show yesterday's video of it working is not at the Emerging level. The Emerging level is for students who either did not build the interaction or cannot articulate what it was supposed to be.

Dimension 3: Form and function integration

How well does the physical form accommodate the electronics and how well do the electronics serve the form.

This is the dimension that most distinguishes ID work from engineering work. A project that scores well here feels like one object, not like a nice shell with a bad gadget hidden inside.

Distinguished (4). The form and the electronics are legible as one design decision. Cable routing has been considered. The USB port is somewhere that does not ruin the silhouette. The sensor is positioned where the interaction wants it to be, not where the breadboard was convenient. If you open the object up, the interior has been thought about too.

Proficient (3). The form accommodates the electronics without visible compromise from the outside. The interior is functional but may be messy. The student has made deliberate choices about where things go even if some were forced.

Developing (2). The form and the electronics coexist. You can see the seams. The USB port is in an awkward place or the cable has to come out of a gap that was not planned.

Emerging (1). The form is a box that contains the electronics. The electronics are loose inside. The object would not survive being moved around.

Note for graders. This is the dimension where you as an ID instructor are actually the most qualified grader in the room. Trust your instincts. You already know what good form work looks like. This rubric item is just asking you to apply that judgment to objects that happen to have chips in them.

Dimension 4: Craft

How well made is the physical object.

Craft is the dimension most design students already know how to think about. The only adjustment for physical computing work is that craft now includes things like wire routing, solder joints, enclosure tolerances and the decision about what to show versus what to hide.

Distinguished (4). The object looks intentional from every angle. Material choices are deliberate. 3D-printed parts have been finished or the rough surfaces are there on purpose. Visible fasteners are chosen, not accidental. The object survives being picked up and handled without falling apart.

Proficient (3). The object is well-made with minor issues. One surface is rougher than it should be, or one fastener was the wrong choice, or a cable is visible where it should not be. Overall the craft is defensible.

Developing (2). The object is functional but looks like a prototype. Hot glue is visible, tape is visible, edges are unfinished. A reasonable person would assume it is unfinished because it is.

Emerging (1). The object is held together by friction, luck or optimism.

Note for graders. Do not penalize students twice for the same mistake. If a student's load cell fails during the crit (Interaction dimension) and the mounting of the load cell is also visibly sloppy (Craft dimension), those are two different observations of two different things. Grade them independently.

Dimension 5: Documentation and reflection

Can the student explain, in writing, what they made and what they learned.

This includes the two-page rationale, the demo video and (critically) the week 5 failure list if the studio brief uses one.

Distinguished (4). The written documentation makes the object legible without the object present. The demo video is usable as a portfolio piece without additional narration. The failure list identifies real, specific failures and describes real, specific fixes. The student has an articulated theory of what they would do differently next time.

Proficient (3). The documentation is complete and coherent. The video shows the interaction clearly. The failure list is genuine but thin, or the reflection is accurate but shallow.

Developing (2). The documentation exists but feels like it was written the night before the crit. The video is usable but badly lit or shaky. The failure list is vague.

Emerging (1). The documentation is incomplete, missing key elements or contradicts what the actual object does.

Note for graders. This dimension deserves more weight over time, not less. The documentation is where the thinking lives. A student who can write clearly about what they made and why is a student who has actually done the work, even if the object itself is modest. A student who built a spectacular object and cannot write a paragraph about it has usually had help they are not disclosing.

Putting it together

The five dimensions work best weighted equally. Some instructors weight Interaction and Concept more heavily. That is fine and depends on what your program is trying to develop. Resist weighting any single dimension above 30% of the total, because the whole point of a five-dimension rubric is to prevent one factor from dominating.

A Distinguished score in all five dimensions is rare. Expect maybe one or two per cohort. Most strong students land at Proficient across the board with one or two dimensions at Distinguished. That is an A-minus or A. The overall grade calibration:

Your program's grading culture will vary. The anchors matter more than the bands.

A worked example

To show how this plays out in practice, here is a real (anonymized) student project.

The project. A tabletop object shaped like a small smooth stone. When you placed your hand on it, a slow warm glow emerged from underneath through a translucent base and faded out after about ten seconds. Built with a capacitive touch surface and a warm-white LED driven through a MOSFET. No WiFi, no display, nothing extra.

Concept clarity: 4. The rationale named a specific context (evening reading, the moment before bed) and a specific emotional need (something to acknowledge presence without demanding attention). The student could defend why it had to be an object: an app on a phone would pull attention the wrong way. This was a strong concept.

Interaction quality: 4. You put your hand on it and it responded. No explanation needed. The fade-out timing felt right (the student had iterated on it for most of week 4). Walking up to the object for the first time in the crit, it took under five seconds to figure out what it did.

Form and function integration: 3. The stone shape was beautiful. The USB-C port was on the back in a small recess, which was acceptable but slightly clumsy (the cable stuck out at an angle). Interior was clean but the battery was held in place with double-sided tape.

Craft: 4. The printed shell had been sanded, painted and finished. The translucent base was diffused with sanded paper inside. The object felt complete. It photographed beautifully.

Documentation: 3. The rationale was good. The video was shot in harsh light and could have been better. The failure list identified three real issues but felt rushed.

Total: 18. A.

The student who built this had never touched a microcontroller before the studio. The code (which was not evaluated) was copy-pasted from an MPR121 example and a basic PWM fade tutorial, slightly modified. If the grading had been based on code quality or on the ambition of the engineering, this would have been a B. It was an A in a design studio because the design was excellent.

What to do when something breaks during the crit

This happens. It will happen in half your crits the first year you run this studio. Here is the recommended approach:

  1. Give the student one minute to try to recover the demo without panicking. A loose cable, a dead battery or a reset-me-and-try-again is not a grade-destroying event.
  2. If it does not recover, ask to see the video shot the day before. Most students will have filmed the object working.
  3. Grade Interaction Quality based on a combination of the video, the brief hands-on you did have and the student's ability to articulate what the interaction is supposed to feel like. Do not grade Interaction Quality on the six seconds of failure you just witnessed.
  4. Write a note on the feedback form about the failure mode and what it suggests the student should have caught during week 5.

This is not generosity. It is accuracy. Grading a six-week design process on its worst thirty seconds is bad measurement.

If you are also wondering how to write the brief that produces the kind of project this rubric is designed to evaluate, the companion post A Physical Computing Studio Brief That Actually Works includes the brief this rubric was built for. If you are earlier in the curriculum planning process, The Cognitive Load Problem explains why the structure of the brief and the rubric matters so much.

Related Guides