At first, I suppose the reason that I entered teaching so enthusiastically was because I felt I had learned a lot about teaching from my courses as an undergraduate at the University of Texas. I thought I knew the mistakes that I had seen made, and I thought I could do a real service for other undergraduates who were in a situation I could sympathize with. To some extent, I succeeded. I got good reviews after my first year, and I knew the students liked the way I taught. I was also fairly sure that they were learning, but I have since learned that this is very difficult to measure.
Since my first few courses teaching Astronomy, I have sharpened my focus in teaching on analysis. The central theme of my classes is typically a presentation of a series of data sets, such as characteristics of stars, observations of the Moon or Sun, etc. and a study of the many different problems involved in analyzing the data set and coming to some conclusions in which you can have some confidence. Along with data analysis comes error analysis and an understanding of what exactly confidence means when it comes to making numerical conclusions.
I've taken this approach in my astronomical teaching as well as my mathematics and physics tutoring/teaching whenever possible. A good summary of the philosophy behind my teaching can be found in a handout I give to my students every quarter, regardless of the content of the class I am teaching: ``Cargo Cult Science'' by Richard Feynman. This reading emphasizes the importance of science as a process of understanding rather than a process of collecting facts. My homeworks, exams and lectures all reflect this philosophy.
If I could change one thing about the way I've taught in the past, it would be the sheer amount of material I've covered in my courses. Since my participation in a K-12 education committee on science education, I have come to realize that my teaching should emphasize quality over quantity in some sense. Thus, instead of trying to cover an entire book (or a big part of it) in a given semester, I intend to experiment in future classes with a much more limited and in-depth approach.
I am worried that students will find this somewhat boring and repetitive, wanting to move on rather than ``beat an issue to death'', but I am willing to accept the challenge to make it exciting for the students and see whether or not this approach leads to better understanding by the students. Better understanding of exactly WHAT is dependent upon the goals of the course, but my goal overall in science education is to give the students the confidence and ability to carry out the process of scientific inquiry on their own.