Assessment #

The problem #

A traditional grading scheme, in which you are asked to meet carefully defined requirements and then awarded points in proportion to the number of those requirements you meet successfully, doesn’t work well in a class like this for three reasons.

The first is that if I grade you with any static set of requirements, most of you will stop as soon as you meet those requirements, i.e. you will try to do the minimum to get an A. This is because you have been trained to maximize your grades since elementary school.

The second is that, when it comes to fabrication, we have a wide range of experience in this class. Some of you have built lots of stuff; some have you have built very little. This is very different than, say, differential equations, where we have a huge cadre of calculus teachers trying to prepare you for the class. Those of you with experience building stuff will have gained it either at home, or at a summer program, or some extracurricular robotics club. There is no AP Building Stuff exam. If I try to hold you all to the same standard, some of you will be underchallenged, while some of you will be crushed.

The third reason, which is probably the most important, is that in engineering, failure is the immediate precursor of learning. The moment you find yourself thinking, “Hmmm, that’s funny, that definitely wasn’t what I was expecting,” is the moment your learning opportunity arrives. Once we understand that, it is immediately clear that you should not be penalized for failing. Instead, I want you to make predictions, model the behavior of systems, and try to understand how stuff works at a deep level. Explore the boundaries of your knowledge. That’s where the learning happens. You can’t do that if you’re getting punished for every misstep.

The plan #

The best solution I have to this situation is to treat you like responsible adults, who are genuinely interested in learning. For each project that we do, I will ask you to develop a short list of learning goals: basically, what do you want to get out of this project? If you’re not exactly sure, I can give you some suggestions. Then, at each prototype milestone and final project submission, we will have a “quiz” on Canvas, in which you attest to the degree to which you met your goals and reflect on your successes and failures. At the end of the semester, your grade will just be the percentage of your goals that you met.

Here are a few effects that this should have.

First, you do not have to worry about getting an A at all. To get an A, just set your goals very low and meet those goals– easy! Cool, that’s out of the way; now you can focus on learning.

Second, I hope you’ll find that this learning model can help you transition to how you will learn after you graduate. Continuing to learn and grow as an engineer is absolutely crucial– if you can’t develop as a professional, you’ve got a real problem on your hands. This course will give you a chance to see what happens when you’re the one deciding what you’re trying to learn and how to go about doing it. My hope is that as you set learning goals and try meeting them, you get some insight into what works well or not so well for you. You are, of course, welcome to try different strategies during the semester.

One warning: you may think to yourself, “Sweet! This class is going to be so easy! I’ll just set my goals to learn virtually nothing, and then succeed at that! Graduation, here I come!” If you take a disciplined approach, that might work, but the far more likely outcome is that the class will be engaging to you, and you will spend a bunch of time learning stuff, despite your best efforts. If you just want a reliably easy course, take something predictable with a traditional grading structure, where there is no danger of your natural enthusiasm for learning fouling everything up.

Learning goal examples #

Your learning goals should be personalized to your experience, situation, and interests. They should be achievable given a level of effort you think you can muster. They should be customized to the project.

Here are some example goals.

  • Use the laser cutter (which I have never used before).
  • Use the laser cutter to cut gears (which I have never designed before).
  • Use the waterjet to cut at least one metal gear (which I have never done before).
  • Get more comfortable with cordless drills and at least one other hand tool.
  • Test my system to failure, then rebuild it stronger.
  • Complete my project 24 hours early.
  • Model, predict and subsequently measure at least one mechanical property of my project.
  • Complete the project in less than 6 hours of focused effort.
  • Make at least one part out of steel.
  • Turn a part on a lathe at Bray.
  • Estimate the power requirements of the project ahead of time, and then measure them afterwards.
  • Spend at least 1/3 of my effort on the aesthetics of the project.
  • Make my most refined 3D print ever.