I was pointed to this piece which made me think about why I disagreed with it. The basic point is that mathcamp, as it is taught in political science, does not make much sense and it would instead make sense to learn some programming. The basic point if I understand, is that given the short time, you can not learn much math in such a short time, and instead, becoming familiar with coding will set you on a fertile path.
There is a part I fully agree with: coding and data analysis is something for which the initial learning curve is steeper but shorter than math and it probably makes sense to set you on track. I also agree that one can not possibly absorbe everything one is supposed to learn in four-six weeks. I would however like to put a nice word for the mathcamp concept, if only to play devils advocate.
- The function of mathcamp is not (only) to teach you math: in my view, mathcamp works to set expectations and measure where each student is. Poli-sci students come with very diverse technical backgrounds. Some of them have not seen any math since high school (that could be 5 or more years), others may have majors in applied math. Yet, in their career as social scientist they will have to deal with a bunch of math- it you are going to take any stats or game theory class that will be the case. Certainly, you can not expect people to learn it in four weeks. However, it helps to set the bar of what is expected of students and a common ground from which classes can start and also it can work as an incentive for students to start learning in advance. It also fulfill an evualative function to help students to choose classes to make it up.
- You can actually learn some math in four weeks: you won’t be able to attend the Bourbaki seminar after 4 weeks of math. But if you have not seen math for a while, being the reminded the basics of calculus (limits, derivation, integration), probability (what does it mean for two variables to be independent?) and optimization, or be exposed to what a write looks like, can really make a difference, if only because concepts will ring bells. Really, you don’t need a crazy amount of math to go through first year of grad school, but you need the minimum (“that is a derivative and I have the basic understanding of what it means“) and that is something you can acquire over 4-6 weeks.
- Vive la différence!: I would like to end this with a reflection. Eventually, most of this discussion is a bit misplaced because an underlying assumption is “what should the ideal political science know”, while in practice, we may be happy with different people knowing different things and exploiting alternative paths. We don’t need to have a community of wise scholars, but one of scientist collaborating and producing, jointly, something that is valuable. But, and here is the important point, people should be able to communicate for that. The goal of teaching math -or formal theory, or any other method- to grad students is not so much to make them learn something they will apply in their own research, is to create a common culture of people who can talk to each other and where members are sensitive to each others understanding. For a large part of people, their first year sequence in game theory and math is probably the last they will ever take. I don’t think everyone should be using formal theory in their arguments, but formal theorists should not become an isolated community quoting each other and becoming irrelevant because they are the only one able to understand their arguments.
Now, none of the above is against the idea that learning some coding is probably a very good use of your time during bootcamp which I think is an excellent idea.