However, turning a lecture series into chat sessions "Students continually discuss concepts among themselves and with the instructor during class." is pretty dumbed down. The example given of result of a truck and car colliding seem to be in the Intro area, maybe to students who have to take Intro to Physics and aren't all that interested in it?
Yes and no.
It was an intro physics class, but at Harvard (with many pre-meds in the course; not sure if majors or not). Certainly the students were motivated and had had physics in high school. The point is that students could do calculations, but they didn't understand, in a fundamental way, Newton's Laws (kinda the point of an intro physics class). If you really don't understand them, then you haven't learned what that course is all about. Here's my favorite quote when one of the students was given the assessment with that question, that asked about everyday phenomena, like colliding cars and trucks: "How should I answer these questions? According to what you taught me or according to the way I usually think about these things?" (again, a Harvard student taught by someone who was regarded an excellent lecturer). When the results came back on the assessment, sure enough, these students largely didn't understand the fundamentals.
In http://www.youtube.com/watch?v=WwslBPj8GgI and http://www.compadre.org/per/items/detail.cfm?ID=4990 Mazur (sorry, both are quite long) describes the same thing with circuits -- students could do the math, but couldn't describe what would happen if a light bulb was pulled out of a simple one. In short, they didn't fundamentally understand the concepts. There is a large literature in physics education research that on a fundamental level students don't understand the key concepts. One leading paper on this (the leading one?) is http://modeling.asu.edu/r&e/fci.pdf (some 1,000 cites from the scholarly literature) . It makes for sobering reading.
Having a clicker response from everyone to questions every few minutes in your lecture I guess is feedback that your points are getting across or not, but I still think it's dumb. It was the open conversation chatting amongst each other and lecturer that was engaging in Mazur's class, not primarily the clicker.
Rather than "dumb," this literature finds that such techniques leads to students who (i) can do calculations as well as those in a traditional class and (ii) have a better fundamental understanding. This really isn't too surprising as they're actually doing physics with a lot of frequent feedback from their clicker responses and discussions with each other to carefully crafted questions designed to help ferret out their common misconceptions. See http://www.youtube.com/watch?v=WwslBPj8GgI (quite short) where students are much more engaged than in the typical lecture. Part of that is students committing anonymously to a question (via clickers). With no questions, there is a strong temptation for students to say, "Yeah, I understand that." Many times, in fact, they don't.
Yes, the clicker is just a means to an end (shouldn't all technology in teaching be that?) -- getting students to commit anonymously to an answer. As Mazur says in http://www.laspau.harvard.edu/idia/mecesup/readings/Eric_Mazur/Mazur_52364.pdf , you can get the same basic results with cards that students hold up and where they can't easily see each other's cards. As you say, and I'm sure that Mazur agrees, key is the discussion with other students and with Mazur.
One more direct role for technology here is that students at first do on-line homeworks that are used to guide the selection of questions for the week. For a short video of how this all works, see http://www.youtube.com/watch?v=lBYrKPoVFwg (very short, and mentioned above).
Two final points. First, a LOT more faculty than Mazur teach like this. It is becoming increasingly common in physics due to the robust empirical finding that students learn more. Second, on why I posted the links, I very much take to heart Carl Wieman's point in "Why Not Try a Scientific Approach to Science Education?" http://www.cwsei.ubc.ca/resources/files/Wieman-Change_Sept-Oct_2007.pdf on how we should study teaching: "Practices and conclusions based on objective data rather than—as is frequently the case in education—anecdote or tradition. This includes using the results of prior research, such as work on how people learn." Pretty much everything in this discussion has been assertions.
