A- in Real Analysis - impact on phd application and remedy?

[shinnyu]

Hi everyone,

 

I took real analysis this past semester and unfortunately got an A-. I heard that nowadays for your application to be considered competitive, getting an A in real analysis is almost a 'must'. I am aiming for T30 programs, and I wonder how much of a negative impact this A- in real analysis will have on my application. A bit more information on my math background (I'm an undergrad double majoring in Econ and Math at the U of Michigan):

 

• Calc I-III (As)
  
• Linear algebra (A-)
  
• Intro to Probability (A)
  
• ODE (A)
  
• Abstract Algebra (A-)
  
• GRE quant (170)
  

 

Also, I wonder what I could do to offset the negative impact of this A- in real analysis. I am a rising senior and will be applying this fall, so I only have this spring/summer semester and the upcoming fall semester. I have thought about:

1. Retake real analysis in the spring and get an A+ (but after second thought I feel like the marginal utility of this strategy is pretty low, considering that taking courses in the spring is expensive and time-consuming, and might affect other work I have over the summer, namely thesis and RA work)
   
2. Take a more advanced analysis class and get a good grade. However, the second course in analysis is not offered in the upcoming spring/summer/fall. Would taking it online be an option? If so, where can I take it?
   
3. Take grad-level micro theory. However, I couldn't get permission to enroll from the instructor due to the fact that I haven't taken game theory ( the policy to enroll in grad-level econ course here is a bit stricter)
   

 

What other things can I do to demonstrate that I am mathematically prepared for grad-level coursework? Any comment/help/advice will be greatly appreciated! Many thanks in advance :)

[Zubrus]

An A- will not sink your application at all. I think you're a strong student, just take an additional (more advanced) analysis class and get an A. Taking grad micro and getting an A would also help a lot.

[shinnyu]

Thank you so much Zubrus! Right now it seems unlikely that I can take either an additional analysis class or grad micro this fall semester (which is the last semester before I apply), since the more advanced analysis class is not offered in the fall term, and I couldn't get permission to enroll in grad micro... But I will try to email the instructor for the grad micro class and see if he will change his mind.

[Cauchycomplete]

1. I don't believe that retaking class at Michigan can hide grades that you had before. Also, A+ is never guaranteed.

 

2. Do Math 490. That helps for application. Or check if you like Math 663 (this is a super helpful class for econ phd coursework). Those are non-standard but useful analysis courses offered in Fall 21.

Other than those two, Math 525 or 555 may help for applications as well. But I don't like the instructor for 525, at least his style when I was taking, not recommending it.

 

Perfomance in online classes (even A+) do not offest A- in Michigan coursework at all. I think you could just choose of the above four, no need to wait 452 &551.

 

3. Micro group won't give you the override permission as they rejected students with way stronger math background than you do due to the lack of game theory. It's not determined by a single instructor. In fact, you could enroll for Phd Econometrics, which includes measure-theoretical probability theory. It is also a good signal.

 

Lastly, I really think that An A in real analysis is never a "must" for T30 program. If one has decent grades in Economic courses, along with good letters and research experiences, literally no reason to reject just because a single A-.

[tutonic]

Good grades aren't anything special for Top 30 admits, since everyone has them. You also need good letters and research experience. Do you have those?

 

You don't need classes on Topology or Non-linear programming since they are only marginally useful for grad school. You've probably already learnt enough point-set topology in your analysis class. The non-linear programming class looks very applied. You'll only really benefit from the convex optimisation bit at the start, so it's not entirely efficient to take the entire course (unless you want to do macro). You also definitely don't need complex analysis (unless you really love math). If you really want to load up on more math, functional analysis, at the very least, is more relevant than complex analysis for grad school purposes.

Lastly, you should consider taking a mathematical probability class or a measure-theoretic probability type class.

[shinnyu]

Cauchycomplete, thank you so much! Nice to see a Michigan peer here.

 

1. Yup, now I am certain that retaking RA isn't a good strategy at all.

 

2. Actually, I am on the waitlist for MATH 525 (also 526). I was a bit worried that since I've taken 425, 525 won't add too much value. But it does seem like those two are quite different, especially that 525 touches upon measure theory at the end. Also taking tutonic's advice into consideration, taking 525 next semester sounds like a great idea. As for 490 and 663, well, tbh I am a bit worried that it will take me a lot of time and effort to excel in those classes. Since next semester I will be working on my applications, I'm trying to keep the course workload manageable...

 

3. Thanks for bringing up the phd econometrics class! I did not know about that somehow... I checked the syllabus and it certainly looks very useful and relevant.

[shinnyu]

Good grades aren't anything special for Top 30 admits, since everyone has them. You also need good letters and research experience. Do you have those?

 

tutonic, thank you so much for your advice!

 

I believe that I will have two strong LORs (fingers crossed, cuz nothing is certain...), one from my thesis advisor, and one from a professor I have taken a seminar class with (final paper graded A+) and will RA for over the summer.

 

As for the third LOR, I am torn between 1) asking the professor whom I took another seminar class with two semester ago (got a good grade but it's been a while and we haven't been keeping in touch) and 2) asking the professor whom I have RAed for since the start of the past semester (worked with his grad student so I never communicated with him directly, but grad student told me that he will update the professor on my progress and a LOR is guaranteed if I need one). What do you think?

[tutonic]

The second one is better, in my opinion. Granted, both aren't really great, but at least with the second one, the letter can be about your RA work, instead of just being a generic letter attesting to your academic ability (which wouldn't be informative since that information can be obtained from your transcripts).

[shinnyu]

The second one is better, in my opinion. Granted, both aren't really great, but at least with the second one, the letter can be about your RA work, instead of just being a generic letter attesting to your academic ability (which wouldn't be informative since that information can be obtained from your transcripts).

 

Make sense! Thank you :)

[econliker]

Math grades are way overblown on this forum. I know people who got far worse grades than A- in analysis (or easier courses) and still got into top 10 programs. Your grades are perfectly in line to get into a T30, at this point it's much more about your research ability/experience

[Cauchycomplete]

A warning: Math 525 may turn out to be not much different than 425, depending on who you're taking with. Nowadays, I believe that most sections have no touch on Measure-theory related stuff. If you want that, go for Econometrics. The end should be either Markov Process (Intro to 526) or Monte-Carlo Methods.

 

That said, the biggest difference is only for the "rigorous" introduction to concepts for convergence (a.s, in Probability and etc), which should be taught in STATS 426 but at lease mine did not do that. Most schools will do convergence stuff in the second Mathematical Stats class (i.e our STATS 426 version).

 

Thus, once again, I highly recommend Math 526 over 525, based on who are teaching the class. You would be able to suceed in 526 without 525.

 

Some other info: In the past, Math 490 was easier than 525, but it seems to change recently because of instructors.

 

Good luck on your choices!

 

 

Cauchycomplete, thank you so much! Nice to see a Michigan peer here.

 

1. Yup, now I am certain that retaking RA isn't a good strategy at all.

 

2. Actually, I am on the waitlist for MATH 525 (also 526). I was a bit worried that since I've taken 425, 525 won't add too much value. But it does seem like those two are quite different, especially that 525 touches upon measure theory at the end. Also taking tutonic's advice into consideration, taking 525 next semester sounds like a great idea. As for 490 and 663, well, tbh I am a bit worried that it will take me a lot of time and effort to excel in those classes. Since next semester I will be working on my applications, I'm trying to keep the course workload manageable...

 

3. Thanks for bringing up the phd econometrics class! I did not know about that somehow... I checked the syllabus and it certainly looks very useful and relevant.

[shinnyu]

Thank you! A friend of mine who took 525 last semester said the same thing. Will go for 526 then :)

 

A warning: Math 525 may turn out to be not much different than 425, depending on who you're taking with. Nowadays, I believe that most sections have no touch on Measure-theory related stuff. If you want that, go for Econometrics. The end should be either Markov Process (Intro to 526) or Monte-Carlo Methods.

 

That said, the biggest difference is only for the "rigorous" introduction to concepts for convergence (a.s, in Probability and etc), which should be taught in STATS 426 but at lease mine did not do that. Most schools will do convergence stuff in the second Mathematical Stats class (i.e our STATS 426 version).

 

Thus, once again, I highly recommend Math 526 over 525, based on who are teaching the class. You would be able to suceed in 526 without 525.

 

Some other info: In the past, Math 490 was easier than 525, but it seems to change recently because of instructors.

 

Good luck on your choices!