Xuming He writes:
I went to college because I thought that I would know all about mathematics after four years in college. By the time I completed my PhD, I felt that I had hardly scratched the surface of my specialization, statistics. As the wise would say, the more you know, the less you think you know.

Probably many junior faculty face the same dilemma today. After getting their PhDs, or perhaps after a couple of years of postdoc experience, they plunge into the exciting new world. As they celebrate their achievements, and luck in landing a position as Assistant Professor, they suddenly are swamped with teaching, research, and service. They are told to get more papers published. They are told to get more external funding through grants. At the same, they feel that there is so much that they have to learn, and re-learn. Because of time pressures, many choose to cut back on the time spent teaching by taking on the same courses over and over again. Teaching the same courses may save them preparation time, but I would argue that it is not a good way to advance their careers.

I will start with the obvious: the best way to learn is to teach. Teaching new courses is a sure way to broaden and deepen our understanding of our own discipline. Since junior researchers have the greatest need to learn more, they should be the first in line to teach new courses. My personal experience, if I try to quantify it, tells me that 80% of what I know today was learned after PhD, but more than half of that was accumulated in my earlier years as Assistant Professor. Of course, I cannot stand behind those numbers, but the message is loud and clear. As an Assistant Professor some years ago, I was fortunate enough to be able to teach a wide variety of courses in statistics. I taught the basic graduate level courses such as mathematical statistics and linear models. I taught topics that I was not so familiar with, such as design of experiments and multivariate analysis. Whenever I had a chance to teach a topic course, I included the topics that I wanted to learn myself. That was how I learned Bayesian computation, nonparametric estimation, variable selection and so on. I knew little about these currently active topics. If I had not taught those classes, I would have been at least 50% out of date.

The field of statistics has changed a lot in the last 20 years. I cannot make useful predictions about what will emerge 20 years from now, but I can say with 99% confidence that the challenges that we will be facing will be different from what we know today. Even if you are doing excellent research now in your own field, you will need to be broader, and will need to continue learning both the classical and the emerging ideas and tools, in order to stay ahead. Teaching a wide variety of courses can help achieve this goal.

I suggest that junior faculty teach one new course each year. Department administration needs to provide the necessary support to make this possible. This support should include offering them opportunities to teach a wide range of courses, and helping them develop a broad-based and well-balanced research career. They do not want to become jack-of-all-trades, but narrow-minded researchers risk becoming frogs at the bottom of a well.

The benefit of teaching new courses speak for itself. If you teach the same course many times, you will lose enthusiasm. You can be a more inspiring teacher if you are enthusiastic about the material yourself. When you teach a new course, you will better stimulate your brain and think better not just about teaching but about your own research as well. Statistics has developed in such a way that most of its sub-fields are intricately related, so that knowing one subject deeply often aids understanding of another subject. You may develop a new idea on your own research problem partly because of what you see in a seemingly unrelated subject. If you take the challenge to teach one new course each year, you will become a happier statistician over time.

Where can junior faculty find time to teach one new course every year? There are at least three ways that department administration can help. First, offer a reasonable teaching load to junior faculty. Some departments are still asking its faculty to teach two courses each semester. The time needed to teach a new course roughly equals 1.5 times that for a repeat course. It makes sense to equate two new courses to a three-course load. Second, offer good TA support. A good TA can help faculty manage their time better. Third, offer non-conventional topic courses, where two or three faculty members share teaching. This is one way to bring faculty and students together to stay abreast of the latest developments in our field.

Junior faculty may not have a lot of say in their departments, so I hope they get unconditional support from their mentors and their department chairs. By doing so, everyone benefits in the long run and the department can only grow stronger. ■