quantitative modeling/calibrated peer review

A key part of the course is for students to grow in their ability to solve complex problems. One vital skill is to take a complex system and to develop an approximate, model description of it that captures the main features, but is simple enough to be tractable.

In the mid-semester projects students must come up with a reasonable model description of an device they regularly encounter and try to quantitatively calculate some aspect of it, e.g. how long it takes for a pizza to cook. In this case trying to calculate the rate of heat transfer to the pizza what it might depend on, all the way to temperature changes etc. The educational goal is for students to develop skills in figuring out how to approximately model a complex device so that the science involved is correct enough to reasonably accurately describe how it performs. Key is often to know what to leave in or out of the model.

It is also a good chance to encourage student writing skills.

The challenge has been how to get good feedback to the students on their work. The TAs are the main source of feedback, but this semester I am also trying peer feedback. Students submit their draft work to a web-site http://cpr.molsci.ucla.edu/ , the website then shuffles the papers, and students are then asked to give feedback on 2-3 other projects. Hopefully the benefit is both ways, by reading other projects students will develop a stronger understanding of modeling, and they will get specific feedback on their project.

So far it has been relatively smooth, with the largest concern being that the only format you can submit is plain text, i.e. all equations, figures, tables, graphs are lost. I will suggest to the designers of the site, that they consider pdf uploads as well