My teaching interests are in introductory courses and computational science classes, where helping students learn the process of doing science is at least as important as mastery of the particular techniques or facts covered in the curriculum. In these cases, going beyond lecturing, and experimenting with peer- and activity-based learning --- while testing to see what works and what doesn't --- has been valuable.

(This teaching statement is available as a PDF)

I'm an astrophysicist. We live in a day when, through the web, we have access to pictures of the Universe in all its glory, taken by a billion-dollar telescope that orbits the Earth. I personally study the explosions that end the lives of stars - explosions among the largest seen in the history of anything that has ever existed. I've also studied both the birth of galaxies, and the emergence of vast voids in space, emptiness abandoned even by the wisps of gas that inhabit the gaps between stars.

When you have this sort of subject matter to work with, you'd like to think that conveying it to students would be easy. It isn't, of course, because understanding never comes easily, whether one is trying to acquire it or inspire it. Still, the attempt to convey that understanding - of astrophysics, or especially of the scientific process in large - is deeply, genuinely important, and deeply, genuinely satisfying when successful. And teaching that process is important not only in introductory science classes, but throughout a curriculum, perhaps especially in computational science courses, where the focus is too often on the tool rather than the goal.

I have great enthusiasm for what I do, and in helping students to start to understand it. This enthusiasm has helped my students, and has helped me accumulate some kudos and very positive teaching evaluations despite my initially unschooled and somewhat naive approach to science education at the start of my teaching career. As I learn more about the teaching process, and tools and techniques for engaging students and improving classroom and lab techniques, I hope to become a stronger teacher, and leave more students with better understanding of science, the scientific endeavor, and why both matter.

Purpose and Goals: General Science Education

Because the importance of the scientific process extends far beyond laboratories and academic journals, so must the teaching of that process. Being skeptical, testing understanding against data from the real world, and being open to the possibility that one's understanding may fail that test - these are skills with wide application to understanding our world, and which are valuable to all. Part of our job is teaching the scientific approach - to understanding problems and searching for their answers - to students who will go on to careers very different than our own.

This understanding has shaped my ideas of teaching science, especially at the introductory level. It's the scientific process that I've tried to capture - whether teaching a course to new or non-science students, who are seeing it for the first and maybe the only time; or introductory computer programming courses, where abstract problem solving is a similar process; or in supervising undergraduate students through their first serious research projects, where they finally get to apply the process to novel work.

I don't expect a student, five years later, to remember a particular equation, or algorithm; they can look that up if they need to. Five years later, I would like the student to remember the process taken in getting to the stage where an equation or algorithm could be used to solve the problem, to be able to use that process in their own work, or to be able to have a more nuanced understanding of science issues and events when they make the news.

Purpose and Goals: Computational Physics

This process approach is equally relevant in computational science courses, which too often focus exclusively on programming techniques and lists of algorithms. These are important, and must be taught, but the process of doing computational science is more than editing and compiling a code until it doesn't have any obvious bugs.

Experimental science has had a long time to discover how best to train new practitioners - how to convey the importance of design of experiment; integrity of results; and reproducibility to incoming students. Computational science is much newer, and hasn't had as much time to develop a canon of material for all incoming students to study. But computational science can learn much from the history experimental science education. Many of these ideas are just as applicable to learning to do computational science well. The importance of keeping logs; continual testing; `equipment' design; and presenting enough information for other groups to reproduce the results.

Computational courses must also deal with pitfalls and opportunities which are exclusive to computational approaches. In testing their codes, computational scientists have powerful techniques like the `method of manufactured solutions', which rely on changing the physics in the simulation - an approach unavailable to real-world experimenters. However, simulators must also be careful that their techniques may be robust enough that they provide results even when the assumptions that have gone into the codes have failed; so that scientists must develop a habit of being skeptical even of well-tested codes when (as is inevitable) they are applied outside of the domain for which they were originally designed and tested.

Techniques

The importance I place on teaching the process of science as well as the body of knowledge strongly informs how I've approached teaching. Carefully crafted lectures can convey a lot of facts - and it is certainly important for the students to learn those facts - but one can't simply lecture a process. A lecture format is useful for conveying lots of information, and allows the lecturer to demonstrate working through a problem, or to frame difficult concepts for the class. On the other hand, it means that the students role is largely passive, which means that even when the topic is the wonders of the Universe, attention span is short and retention is low.

From January to May 2004, I taught and created the curriculum for a class entitled `The Search For Life In The Universe' at the School of the Art Institute of Chicago.saic.edu. Since this was likely the only science course these students would take through their degree program, the breadth of the course topic was in some ways ideal; the course touched on astronomy, geophysics, chemistry, and biology. However, there were challenges; one almost has to admire the demented genius of whoever scheduled a science course for Art College students as a three-hour lecture session on Friday afternoons; the combination of these factors made it essentially mandatory for me to include some non-lecture-based approaches in the teaching.

This was the first time that I applied several Active Learning and Collaborative Learning techniques - both to keep students involved, and to get more feedback about student learning through the course. Absent real lab facilities, microlabs during the three hours and quick writing assignments helped serve the joint purposes of allowing the students to explore the material, collaborate, and stay involved during the afternoon. Because of the variety of material we covered and the relative inexperience of the students to science, there was often a great disparity between students' ability to absorb the lecture material, with different groups of students `getting it' depending on what the topic was. In this case, collaborative or peer-based activities worked extremely well for improving all of the students understanding of the material.

Given the time and equipment constraints, some of the activities verged dangerously close to goofy, and there was a small group of students who grumbled fairly consistently about them. In a particularly vivid example, where we performed an activity originally suggested for much younger students, the students themselves lined up in polymer chain and enacted protein folding. (My one concession was to not actually suggest that they hold hands.) However, in testing, the material covered by such activities invariably showed better retention and comprehension by the students, teaching me something about the limits of student feedback in like/dislike kind of questions - at least in the case where only a small group of students roll their eyes.

In teaching computational classes, the opportunities for useful activities and collaboration present themselves more easily, so this is less of an issue. However, computing classes have their own problems. I've already mentioned the lack of emphasis on the scientific process; another neglected process is debugging. While effort is put into the teaching the problem solving process of putting together an algorithm, very little is generally spent on the quite different problem solving process of fixing your implementation of it. I have put together some materials on debugging that I try to work into any basic computing course. Indeed, I've found that interactive debuggers are extremely useful at very introductory levels, as they demonstrate quite clearly what elementary constructs do, while also introducing students to a tool with which they will soon become all too familiar.

At a more advanced level, I also currently lead a seminar on astrophysical hydrodynamics, largely attended by graduate students, postdocs, and faculty; and am trying to put another one together for computational astrophysics. This is a much different experience; although there is some teaching involved, in making sure the material is accessible to the newer graduate students, and some preparation required, it is much more of a collaborative exercise.

I enjoy the seminars a great deal, and these sorts of activities, as well as teaching more advanced formal courses, are certainly also very important. The unique challenges and payoffs of teaching students who are just beginning their scientific education, however, makes it especially exciting - even by the standards of someone who blows up stars for a living.