Here’s a little advice to students picking electives.
Consider taking classes in those things that would be hardest to learn on your own after you graduate. Taking the most advanced courses available in your major may not be the best choice. Presumably you’ve learned how to learn more about your area of concentration. (If not, your education has failed you.) So the advanced courses might teach you the material you’re best prepared to learn on your own.
Maybe it would be better to take a foundational course in a related area than an advanced course in your main area. For example, I suggested to some statistics graduate students yesterday that they take a really good linear algebra class rather than taking all the statistics they can. If they become professional statisticians, they’ll continue to learn statistics (I hope!) but they may find it harder to take the time to really understand mathematical foundations.
From a statistical viewpoint, I think it would be beneficial to take courses related to the science you are interested in. I have consulted with a few people this year in statistics, and it seems that some of the things that hold me back the most is the science. I could just be an outlier, but I think understanding the science behind experiments can definitely make your staitistics much more meaningful.
With this being said, the cool thing is that most statisticians actually have degrees in other areas and pursue stats later. I actually was a math education major and have recently finished my masters in stats. I’m just looking for a way now to combine these two areas.
Excellent advice. I will be giving this advice to students.
Since this site seems associated with statistics, let me ask a question. Most of the folks I know who deal with statistics think little of the assumptions underlying the question that is being answered using statistics. For example, most of the social science statistical analysis that I am familiar with assume that variables are normally distributed. This leads to brilliant horrid models, like the Risk Assessment Model. I’ve always found that framing the question properly and then writing a statistical analysis package that uses a minimal number of assumptions and that does not tend to hide assumptions works best. I’d be curious to see what the arguement is against doing this, because I rarely see it done.
This is good advice. Sorry for the tangent, but your example is also relevant for machine learning people (well, this is related to statistics). I guess the next best thing for people who are not students anymore is use textbooks or video courses such as MIT’s OCW. Here’s Brendan O’Connor’s Take on poor man’s linear algebra textbook.
Yuval: I didn’t mean to limit this to statistics. I just chose my example from statistics because that’s what prompted this discussion. I’d ask any student to imagine what they might wish they’d taken a course in when they look back in 10 years.
Dan M: Many statisticians don’t understand what they’re doing well enough to be aware of their assumptions or to assess the consequences of their assumptions being only approximately met. But since their clients don’t understand statistics either, they can get away with it.
I think this is generally good advice, but I would recommend taking it much further. I think students picking electives should pick classes as little related to their majors as possible.
So, rather than linear algebra for a statistics major I might recommend history, philosophy, Latin, Chinese, art, anthropology, or some such. A foundational sociology course I took probably has had a more profound effect on the ways I can view the world than all of the math I took (math was my major). A close second was a class on four of the classical problems in philosophy. I am still grateful for the anthropology and history I took, and kick myself for not having taken sculpture (protip for the young: it ain’t all chiseling marble. You also get to weld steel and cast bronze.)
So, if you are majoring in grievance studies, perhaps you should also take a class or two in vector calculus, eh?
I have really appreciated the knowledge I got from industrial engineering electives such as linear programming and queueing theory. Definitely not your typical stats courses.