Four years ago, after having been a high-school teacher for some 10 years, I decided to take a job as an instructor on a university general-physics course. Naturally, I was scared. Would I be able to teach physics as well as a "real" physicist? Although my background in physics was equal to a master's degree and I had a doctorate in physics education, I had never done any physics research, and always considered myself to be a high-school teacher.

However, when I received my first-semester evaluations I was surprised to read one student make the following comment: "She was the best teaching assistant I ever had, and I am a senior." And that was not an exception, for many other students wrote similar comments. I could not understand why they considered me to be so good - I was not, after all, a "real" physicist.

It was only the following semester, when I became a course co-ordinator and started to observe teaching assistants (TAs) who taught the same course, that I understood my evaluations. These TAs, who were graduate students doing physics research, had no knowledge of the basics of how to teach. They merely followed tutorial instructions, and did not try to find out if the students had understood what they were doing; they were satisfied if the students could answer questions with a "yes" or a "no", and could plug numbers into formulae.

In the laboratories they were more concerned about whether the equipment functioned properly than they were about probing the students' understanding. If a student asked a question about a problem, they would immediately solve the problem for him or her, taking pleasure in showing how well they could do it. They could talk for 30 minutes straight, without noticing that only one student in a group of 20 was listening. They would use terms that students did not understand, without seeming to realize what they were doing. They would spend time with one group of students, while the rest of the class waited, not knowing what to do.

It was depressing to watch. The undergraduates were not learning. They tried unsuccessfully to plug numbers into randomly selected formulae in problem-solving sessions, and in the labs they collected data whose meaning they did not understand.

On the last day of class the students filled out course evaluations, which asked them what they liked most about the course. A common response was: "That it's over." And although we had weekly meetings in which I tried to explain to the TAs how to teach the upcoming material, they were just not interested.

The art of teaching

It is easy to assume that if you know physics, then you know how to teach it. True enough you can teach it, but will people learn anything from you? Teaching and learning do not necessarily happen simultaneously. It was no wonder my students felt different about me: finally, they had a teacher who was trained both in physics and in how to teach it.

I therefore decided to take some action, and offered to teach a graduate course on "the basics of teaching physics" in the physics department at my university. My aim was to help graduate students who were TAs to master what is called "pedagogical content knowledge". This would help the TAs to transfer their understanding of physics into a conscious teaching process, which would in turn help the students to learn.

However, it seemed crucial to me that this should be a regular course that gave students credits towards their degrees: the TAs had to feel from the start that their teaching was as important as their research. After all, if they took courses on learning how to do research, they at least had to take a course on how to teach. I also wanted the course to be attended only by current TAs so that they could practise their newly acquired knowledge and skills straight away. In addition, it was important that I sat in their classrooms, observing their teaching and giving them feedback.

Course content

Despite the novelty of my suggestion the physics department voted to approve the course, and about 15 graduate students have enrolled on it over the past two years. The course is divided into two main parts: learning general physics-teaching methods, and learning teaching methods specific for certain topics.

The "general methods" include topics such as "Socratic questioning", in which the teacher does not lecture but asks logically connected questions that guide the students through a theme. I also cover the idea of "concept mapping", in which students select the most important concepts in a particular theme and arrange them spatially to demonstrate logical connections. Co-operative learning techniques, problem-solving approaches, ways of designing practical classes, and tips on how to structure a course are some of the other topics included.

Each method is first discussed in general, using research papers in physics education and the participants' own experiences of teaching. Participants are then asked to provide examples of how to apply the methods to particular physics topics - for example, how to construct a concept map for the topic of static electricity. They also devise a "perfect laboratory write-up" for a given topic such as wave optics, or write a "perfect syllabus" for an introductory physics course for non-science majors.

In the second part of the course - where we discuss how to teach specific physics topics - we consider, for example, the different ways of introducing the concept of "weight". We analyse the typical difficulties that students have with this concept, and examine the best way for students to overcome these problems.

These topics are covered before the TAs actually have to teach them. I then ask the TAs to record the difficulties that the undergraduates have with each concept and to find out whether their preparation has helped them to teach. Although it is a long and painful process - the TAs do not improve their teaching instantly - at least they become aware that a knowledge of physics is not the same as a knowledge of how to teach physics.

Spreading the message

Observations indicate that my supervision of the TAs has helped them to learn how to select objectives for a class, master question-asking techniques, and transform their knowledge into ideas that students can understand. One would therefore expect that graduate students who are teaching would want to take the course because it would help them to teach more effectively. One might also assume that new lecturing staff would want to take this course as well.

Unfortunately, this has not happened in practice. Indeed, most of the TAs who teach physics are non-native English speakers from outside the US. They have problems with the language, problems with the culture, and are preoccupied with their research. For them, teaching is only a way of paying for their tuition. The new lecturing staff, meanwhile, are more concerned about tenure - and good teaching is not the way to get it.

One solution to attract more people onto the course would, obviously, be to make it compulsory. But that would not solve the fundamental problems. Only the recognition that the quality of teaching and learning are primary concerns of any physics department can solve those problems.

There is another difficulty as well: who will teach such courses? Even an experienced physics professor may not be competent enough to do it. After all, it is not only experience that has to be transferred, but also the knowledge of physics teaching. We therefore need people who are knowledgable in physics, education, and physics-education research. And how many people like this do we have? Should we start training them now?

I propose that physics departments must hire specialist staff and work with colleges of education to bring about these changes. Only then can we properly train the next generation of university teachers.