PDA

View Full Version : College sophomore - how much more math to take?



younggirl
02-16-2015, 03:28 AM
Hi all,

First time poster. I'm currently a sophomore, Econ & Stats major at a top 20 Liberal Arts College.

I plan to eventually go to a top 15 Econ PhD program. Do you guys have any advice on how much more math to take? So far I've had Calc III (A-), Intro Linear Algebra (A), Undergrad Intro Analysis (A), Intro Stats (A), Econometrics (currently taking), and the standard Econ sequence (all As)

To finish the Stats major, I need to take Mathematical Stats, Undergrad Probability Theory (not analysis-based), Regression, and Data Mining. I plan to also take Stochastic Processes. Are there any other classes I need to take for Top 15? I'm thinking of Ordinary Differential Equations, Real Analysis II, Time Series, or grad level Micro/Macro. Are any of those useful?

I know the more Math the better. However, as much as I like Math, I still want to have a life, learn languages, study abroad, and be well-rounded.

Any opinion will be greatly appreciated!

Food4Thought
02-16-2015, 04:14 AM
Life? Study abroad? Be well-rounded? HA!

As a general rule, the more math the better. But taking a ton of math and doing horribly is worse than taking less math and doing well.

Given that you will have taken a lot of statistics, you probably give yourself some room to have less math than another applicant. However, for the top 15 programs, you should still have all of your bases covered. For example, look at UPenn's economics page. They suggest having exposure to real analysis, point-set topology, and some measure theoretic probability in order for ones application to be competitive. This is a lot of advanced math. Most people only get one of those boxes checked.

Since it is early in your career, I suggest also trying to get a summer internship at a Fed bank or something of that nature. Good liberal arts colleges usually have really good connections to places which take summer RAs. That will look very good on an application to grad school; it will allow you to say, "Hey!! I've done actual research. I know what I'm doing and know what I'm getting myself into! Accept me!!"

younggirl
02-16-2015, 05:00 AM
My school doesn't even offer those classes haha.

I checked the nearby school and the measure theoretic probability course requires grad level real analysis. I'm afraid I will do horribly because such classes seems to be for Math PhDs, and are not even from my school (so I can't just charm profs to be lenient with grades lol)

As for RA, do you think it's better to work at the Fed/similar places or at least spend 1 summer working for profs from my school first? Because I think I should get at least 1 good rec from a prof at my school. I also plan to write honor thesis and work as an RA before applying to grad school (to avoid the GRE senior year and also make some money).

Food4Thought
02-16-2015, 05:08 AM
My school doesn't even offer those classes haha.

I checked the nearby school and the measure theoretic probability course requires grad level real analysis. I'm afraid I will do horribly because such classes seems to be for Math PhDs, and are not even from my school (so I can't just charm profs to be lenient with grades lol)

As for RA, do you think it's better to work at the Fed/similar places or at least spend 1 summer working for profs from my school first? Because I think I should get at least 1 good rec from a prof at my school. I also plan to write honor thesis and work as an RA before applying to grad school (to avoid the GRE senior year and also make some money).

Exactly. It is sort of ridiculous to assume students will have measure-theoretic probability theory under their belt, because that exclusively reserved as a graduate-level topic. Just take up to real analysis and you should be fine so long as your grades are good. Since you come from a liberal arts college, the important thing will be for your letters to be very strong.

There is no formula as to which one is better. I suggested Fed, because that is more prestigious than just working for your professor. That being said, the latter will garner a much better recommendation, since you will likely have taken courses with said professor and discussed grad school extensively. An RA is a very good thing to try to get if you are coming from a less than ideal situation. For those in liberal arts colleges or large state universities, they can often be a golden ticket to a top-10 admit.

younggirl
02-16-2015, 05:50 AM
I got an A in Real Analysis 1. Should I take one of the following courses:

1. Topology
2. Mathematical Analysis: Lebesgue measure, Banach-Tarski paradox, Axioms of Choice, Nelsons Internal Set Theory,...
3. Real Analysis 2: Topology of the euclidean space and functions of several variables. Introduction to Fourier analysis. Metric spaces and normed spaces.

Which one is more useful and which is easier?

Also, how to get a "prestigious" summer RA position? Should I ask profs to connect me to such places? Profs that know me are looking for RAs themselves. I don't want to ruin my chance to work with them (as a back-up plan) by telling them I want to work somewhere else.

teddyb
02-21-2015, 11:59 AM
3. Real Analysis 2 would be the most useful. Then you can do 1. Topology, which is a generalization of what you do in 3 to abstract spaces.

As for 2.....I don't think you'll ever need Banach-Tarski paradox (I might be wrong), but Lebesgue measure is a step into measure-theoretic probability.

But to be frank, any kind of proof-oriented math would do. It's mathematical maturity that you'll really need in the end.

Mathew952
02-22-2015, 02:39 PM
I'm in a grad Math For Economists course right now, and I can say that having Calc III plus linear under your belt already is a really useful set of classes, and especially experience with proofs. To expand on that, I'd suggest Diff Eq, it's apparently very useful for macro, and basically just more proofs courses. To give you an idea, I took:

Honors Calc I-III
Linear Algebra
Diff Eq
Probability Theory
Math Stats
Basic Concepts of Mathematics (Basically intro analysis at my school)
Math for Economists (now)

And my highest rank acceptance right now is OSU (#27 by USN), and I got fellowship to Vanderbilt. I'd say you have plenty of Quant skills as it stands. I think what helped me so much was having really good letters of rec, and doing a lot of research. I think you'd probably get more out of that than carving ever deeper into abstract math. Definitely continue to study them and push yourself, but I think you have a strong enough base there that you could do more to boost other parts of your app.

srslee
02-23-2015, 06:50 PM
I've found my class on measure theory and lebesgue integration to be rather helpful with the game theory class I'm taking now (e.g. the set of static games with an odd number of Nash Equilibria are a set of measure zero) but that can be a bit overkill.

I will be taking topology next year (and have questions on probability vs measure theoretic probability theory, which will appear in another thread) so PM me if you have any questions!