The Setting of Undergraduate Courses in the Major of Statistics

谢益辉 2007-04-27

Just now I chatted with Teacher Yan JIANG for about two hours on the topic of course setting for undergraduates (when we finished the discussion of a project from the China Customs). There are really so many problems that I can list examples one after another. I’ve always been holding a belief that elder students in higher grades should communicate with those in lower grades, telling them some useful experience – no matter in the aspect of academic research or daily life. This is only too important. We need to simplify, or more precisely speaking, clarify their disordered directions.

Statistics is a discipline with strong logics. From traditional regressions (modeling the mean of Y) to generalized linear models (modeling a function of a certain parameter of Y’s distribution), and to quantile regressions (modeling the quantiles of Y); from sampling without replacement (traditional sampling) to with replacement (the idea of bootstrap); from the central limit theorem (CLT) in probability theory to the confidence interval of the sample mean in sampling survey; … There are just too many connected “lines”, and accordingly, many orbits in which this discipline can develop. Teachers should put great emphasis on the training of the thinking ability of students and tell them how to make innovations (after they’ve mastered the basic knowledge).

For postgraduates, currently I have noticed a strange phenomenon, i.e. we are pursuing the “frontiers” of statistics, but lack some very very basic knowledge of statistics. For example, we’re discussing LASSO every day, but we know little about computational statistics, or even we’re using the principle of least squares every day, but we still have no idea about the computation of standard errors of the coefficients in a nonlinear least square (NLS) regression.

Graduate courses

Problems always exist. The only thing we can do is neither worry nor complain, but make efforts to improve the current situation. Again, there’s a long long way to go. Some day I shall discuss these problems with our teachers and students again.

By the way, this Sunday I should go to the China Customs to help them with the usage of my R program, so I’m worried about small’s going home. Just hope I can get back at noon. Besides, it seems that I’ll be extremely busy in the “May Day” holidays; mountains of things to do…