Modeling in experimental research
Today's Physics Colloquium was given by Prof. Heather Lewandowski from University of Colorado, Boulder, on "Engaging Students in Authentic Scientific Practices in Physics Lab Courses". While the talk specifically addressed how to improve undergraduate physics labs, it was also helpful for more advanced researchers, like myself, for training graduate students and in my own research.
A couple of important takeaways:
- Students should be taught to model the measurement apparatus as well as the physical system. Blaming the equipment to justify discrepancies between the measured and expected outcomes is not allowed. Systematic errors and apparatus limitations need to be built into the model.
- Experiments are an iterative process, where measurement equipment and physical models are continuously revisited and refined. Those of us who have conducted graduate-level research know this fact very well. Therefore, in the context of undergraduate teaching, we should eschew labs with a simple inear design.
- We need to identify and understand the underlying assumptions made in the modeling process. This is hard if you don't have a good enough grasp of the fundamental concepts involved. This leads to...
- "Incomplete prior knowledge can hinder the sophistication of the model". This point is crucial. In my experience, undergraduate labs are usually timed to coincide with the lectures, occurring sometimes before and sometimes shortly after the concept is formally introduced. The idea is that when students learn about a topic in a lecture, they can associate it with their experience in the lab, or vice-versa. The problem is that, often, students don't fully understand the concepts when they step into the lab and end up mindlessly following the recipe presented to them. There is no space for critical thinking, building a model, testing, and revising it, iteratively. I completely agree with Prof. Lewandowski's declaration that "labs don't support theory; theory supports labs!"
- This final point is more of practical solution: use electronic note-taking software like Microsoft's OneNote instead of old-school lab notebooks. This makes complete sense. How many times have I wished for a search function or the option of adding a hyperlink!
My notes from Prof. Heather Lewandowski's physics colloquium today