Profile Evaluation for Masters/PhD Statistics

[coffeehouse]

Hey everyone, I am finishing my third year at a top Canadian University, double majoring Statistics and Economics

 

I want to keep my options open and I don't have any specific research interests at the moment, I would like to know the tier of schools that would be realistic for me for both Masters and PhD. I will be applying for Fall 2014.

 

University: Top tier Canadian

Major: BSc Double Major Stats/Econ

GPA: 4.2 (on a 4.33 scale)

 

Math courses:
Calc I, II (AP Credit)

Calc II: A+

ODE: A+

Matrix Algebra: A+

Intro to Proofs: A

Probability: A+
Stochastic Processes: A

Applied Linear Algebra: A+

Real Analysis I: B-

 

Stat courses:

Elementary Stat: A+
Regression: A+

Stat Inference: A+

Design and Analysis of Experiments: A+

Statistical Learning: B

 

Plenty of Econ courses: the relevant ones that relate to Statistics are: Graduate Econometrics A (basically just regression with more proofs and more rigorous), Optimization (Static and Dynamic): should be an A/A+

 

Research Experience:

Last summer I worked under a machine learning professor doing self research on phylogenetic inference, MCMC methods etc (mostly self learning, no results)

This summer I will be working on PCA analysis, writing some code with another Statistics professor

Both are under NSERC undergraduate research awards (the Canadian equivalent to a NSF undergrad research)

I am also currently an RA for a labour economist

 

I know the main concerns are the B- in Real analysis and the B in statistical learning. How crucial will it be? Keep in mind, Canadian schools have slightly less grade inflation I believe? To my understanding in the states an A- is considered an average grade, whereas in Canada the class averages are usually B-.. I know the real analysis grade is still very mediocre

Next year I will be taking 2 courses in Statistical Theory (for advanced students, has a real analysis prereq, cross listed with first year graduate sequence in our school) and probably time series.

 

What are the range of schools (tiers) for Masters and PhD programs I should be aiming at?

Edited April 13, 2013 by coffeehouse

[biostat_prof]

I assume you are not a U.S. citizen or permanent resident? If you were a U.S. citizen, you would be competitive anywhere with strong recommendations. Unfortunately, if you are a non-citizen, you are going to be relegated to the same pool as the Chinese applicants who have won international math competitions. Still, if you have strong recommendations, you would stand a chance at probably any school in the country. A lot depends on recommendations, though. Try to have a candid talk with your recommenders about how strongly they can recommend you. If you have three recommenders who will say that you are in the top 1% at your school and that you are as good as previous students that have attended top-ranked departments, you are probably golden. If they are only willing to put you in the top 10% or top 25% or something, you still stand a good chance at a top-ranked program, but it will be tougher. I would also try to find a recommender who is willing to note that your advanced calculus grade is approximately average at your university to try to do some damage control. If you do well in your graduate courses and can get a recommender to say that these courses have a real analysis prerequisite, that would also help a lot. In any event, if you are willing to spend the money, I would just carpet bomb all of the top-ranked stat/biostat PhD programs in the country assuming you have some good recommendations. My guess is that at least one will bite. Good luck.

[cyberwulf]

Agree completely with this assessment of your chances and advice. I think being Canadian does give you a leg up within the pool of non-U.S. students, though, because language is a non-issue.

[coffeehouse]

I am Canadian.

 

Thanks for the response. Our Statistics department is quite small, and not many go on and do a PhD in the States (Although it is one of the top 3 Canadian Universities), so top 1% or comparing me to previous students may be a bit tougher. Thanks a lot nontheless!

 

Oh also, would you recommend writing a Math GRE? The general consensus is to avoid it unless I am 100% sure I can succeed? (I'm just thinking of ways to remedy my real analysis grade now)

My GRE scores were 170/154/4.5

Edited April 14, 2013 by coffeehouse

[student12345]

Not sure if this applies to Canadian exam-takers, but as of ~June 2012, I've been able to use ScoreSelect to only send the scores I want to show to each institution. This means you can (realistically) take it as many times as you'd like and not send any scores but your very best, or you could even decide to not send your score at all. So it wouldn't hurt to try! 

 

Also, the Math GRE would not be a suitable replacement for a poor analysis grade. The GRE is a multiple-choice exam that doesn't assess your ability to work out proofs. Real analysis is really the standard for understanding an applicant's strength; in my current stats department (not highly ranked), I've heard from faculty that we simply do not look at applications with poor/no real analysis grade. Personally, I would try to get an A in a second-semester analysis course (if it is offered), or as was suggested above, do very well in an analysis-based course (I believe one of the main strengths of my application was my A in measure-theoretic probability theory and a recommendation from the professor). I can only guess, but I assume this outweighed the B+ I received concurrently in analysis I. 

Edited April 14, 2013 by kimolas

[amillion]

Agree completely with this assessment of your chances and advice. I think being Canadian does give you a leg up within the pool of non-U.S. students, though, because language is a non-issue.

 

I always wondered if there was some some kind of nepotism towards Canadian students versus say Indians and Chinese applicants. The point made here by Cyberwulf makes sens but what about French Canadians (me)? Does the median adcom knows that some Canadians do not speak english very well? Don't want to hijack the thread but it's an interesting topic that hasn't been discussed much here and that would certainly help OP prepare his application for 2014 (and myself as I am in a somewhat similar position).

 

Are Canadians preferred over other internationals by adcoms? If yes are there other reasons than language?

[cyberwulf]

I always wondered if there was some some kind of nepotism towards Canadian students versus say Indians and Chinese applicants. The point made here by Cyberwulf makes sens but what about French Canadians (me)? Does the median adcom knows that some Canadians do not speak english very well? Don't want to hijack the thread but it's an interesting topic that hasn't been discussed much here and that would certainly help OP prepare his application for 2014 (and myself as I am in a somewhat similar position).

 

Are Canadians preferred over other internationals by adcoms? If yes are there other reasons than language?

 

I think that French Canadians are likely to benefit from the assumption that English isn't a concern for Canadian students. And in most cases, French Canadian students' English abilities are well beyond your typical international student. 

 

Other things that can favor Canadians include: 1) the structure and syllabi of courses taken are usually familiar, and 2) Canadian professors writing letters are more likely to be known by adcoms.

Edited April 14, 2013 by cyberwulf

[amillion]

I think that French Canadians are likely to benefit from the assumption that English isn't a concern for Canadian students. And in most cases, French Canadian students' English abilities are well beyond your typical international student. 

 

Other things that can favor Canadians include: 1) the structure and syllabi of courses taken are usually familiar, and 2) Canadian professors writing letters are more likely to be known by adcoms.

 

I would have guessed that the similarities in higher education between Canada and the US would play a larger role than the language issue itself (maybe I'm underestimating the problems associated with admitting PHD students with weak writing skills in quant-related programs).

 

Like OP pointed, I feel there is much less "grade inflation" in Canadian universities than in the US: Is that taken into account by adcoms?

 

Also, I have heard that "internationals" are expected to hold a master's degree to apply for a US PhD: Is it usual to admit Canadian students with only a bachelor's degree (with strong profiles including RAing like OP's)?

 

Thanks cyberwulf for your answer. Concerning OP's question: I'm not sure which tier you should apply in the US as I am in a similar position... Have you considered any Canadian programme. If yes, which ones?

Edited April 14, 2013 by amillion

[coffeehouse]

I would have guessed that the similarities in higher education between Canada and the US would play a larger role than the language issue itself (maybe I'm underestimating the problems associated with admitting PHD students with weak writing skills in quant-related programs).

 

Like OP pointed, I feel there is much less "grade inflation" in Canadian universities than in the US: Is that taken into account by adcoms?

 

Also, I have heard that "internationals" are expected to hold a master's degree to apply for a US PhD: Is it usual to admit Canadian students with only a bachelor's degree (with strong profiles including RAing like OP's)?

 

Thanks cyberwulf for your answer. Concerning OP's question: I'm not sure which tier you should apply in the US as I am in a similar position... Have you considered any Canadian programme. If yes, which ones?

 

I would think that us Canadians have some sort of advantage over International students, precisely due to the reasons Cyberwulf mentioned above.

Familiarity to the program may be a big one.

 

As for whether I will be applying to Canadian schools. To my understanding, Canadian schools don't really take direct entry from Bachelors; they almost always require a Masters. This is because Canadian Masters are usually thesis/research based, and lead up to a PhD. Whereas American schools are mostly applied masters, or many schools don't even have terminal masters. I will probably apply for a few Canadian masters as backups. But I don't believe I will have much problem being admitted into Canadian masters; I am more concerned about my chances at the top American PhD/Masters programs.

[biostat_prof]

Not sure if this applies to Canadian exam-takers, but as of ~June 2012, I've been able to use ScoreSelect to only send the scores I want to show to each institution. This means you can (realistically) take it as many times as you'd like and not send any scores but your very best, or you could even decide to not send your score at all. So it wouldn't hurt to try! 

 

Also, the Math GRE would not be a suitable replacement for a poor analysis grade. The GRE is a multiple-choice exam that doesn't assess your ability to work out proofs. Real analysis is really the standard for understanding an applicant's strength; in my current stats department (not highly ranked), I've heard from faculty that we simply do not look at applications with poor/no real analysis grade. Personally, I would try to get an A in a second-semester analysis course (if it is offered), or as was suggested above, do very well in an analysis-based course (I believe one of the main strengths of my application was my A in measure-theoretic probability theory and a recommendation from the professor). I can only guess, but I assume this outweighed the B+ I received concurrently in analysis I. 

 

It is definitely not true that a lower real analysis grade results in your application being trashed. I know a student who had a mediocre grade in real analysis who literally got in everywhere they applied (and they applied to virtually all of the top-ranked programs). Granted, the real analysis grade was literally the only blemish on the student's resume. The student had some of the strongest recommendations I have seen in my life (think multiple senior faculty saying that they were the best student in 5-10 years at a very highly ranked school) and multiple papers submitted for publication, both in top-ranked statistical methods journals and in high-impact applied journals. But a lower grade in real analysis won't sink you if the rest of your resume is solid (which appears to be true of the OP).

 

That said, I agree that the math GRE isn't like to compensate for a poor real analysis grade. I think most schools have recognized that it covers a lot of material that you never use in statistics and is a poor predictor of how well you write proofs. Your time would be better spent trying to get a better grade in a more advanced analysis course and/or cultivating relationships with potential recommenders.

 

I think that French Canadians are likely to benefit from the assumption that English isn't a concern for Canadian students. And in most cases, French Canadian students' English abilities are well beyond your typical international student. 

 

Other things that can favor Canadians include: 1) the structure and syllabi of courses taken are usually familiar, and 2) Canadian professors writing letters are more likely to be known by adcoms.

 

I don't know about this. The international students that my department has admitted in recent years still tend to be Chinese the vast majority of the time. I can only think of one Canadian we have admitted in recent years, and that was a special circumstance. Granted, I have no idea how many applicants we get from Canada (I am sure there are far more from China). But generally we know that students at even lower-tier Chinese universities have the equivalent of an MS in math in the U.S./Canada, and they have to be very smart to be able to be accepted to a Chinese university in the first place. And the grading system makes it easy to compare applicants; you know that a student with high grades at Tsinghua or Peking is going to be very good. So I'm not sure there is much of an advantage to being Canadian, if any. At the end of the day we want the best applicants possible, and with limited funding available for international students, we tend to favor the Chinese students from top-ranked schools with extremely strong math backgrounds. This may be less true at other departments with more funding for international students, though; I'm not sure.

 

If you want to talk about advantages to certain groups, if you are a minority (i.e. basically African-American, Hispanic, or Native American) the bar is far lower. We get so few applicants from these groups that we are willing to admit almost anyone who we think will survive the program. Being female also helps in statistics departments, although I think that is less true in biostatistics. (The proportion of women in biostatistics programs is always much higher than it is in statistics for reasons I have never fully understood.) Maybe I shouldn't admit to these things in a semi-public forum, but that is the reality.

[wine in coffee cups]

If you want to talk about advantages to certain groups, if you are a minority (i.e. basically African-American, Hispanic, or Native American) the bar is far lower. We get so few applicants from these groups that we are willing to admit almost anyone who we think will survive the program. Being female also helps in statistics departments, although I think that is less true in biostatistics. (The proportion of women in biostatistics programs is always much higher than it is in statistics for reasons I have never fully understood.) Maybe I shouldn't admit to these things in a semi-public forum, but that is the reality.

Here are the most recent acceptance rates publicly available for each group where I can calculate it (although there is uncertainty about the intersection of citizenship/sex/race, and most of these combine across master's/PhD):

• UMN statistics: acceptance rates of 30/121 (25%) for females, 40/173 (23%) for males, 4/14 (29%) minority (includes Asian-American), 53/257 (21%) international, 17/42 (40%) domestic
• UMN biostatistics: 35/68 (51%) female, 28/76 (37%) male, 6/12 (50%) minority, 31/95 (33%) international, 32/49 (65%) domestic
• Duke PhD statistics: 8/75 (11%) female, 8/121 (7%) male, 0/7 (0%) under-represented minority, 8/140 (6%) international, 8/56 (14%) domestic
• UW PhD statistics: 12/156 (8%) female, 23/210 (11%) male, 8/46 (17%) minority, 10/233 (4%) international, 25/133 (19%) domestic
• UW biostatistics: 26/148 (18%) female, 22/113 (19%) male, 7/40 (18%) minority, 22/151 (15%) international, 26/110 (24%) domestic
• UNC statistics/OR: 0/7 (0%) under-represented minority, 45/386 (17%) international, 25/125 (20%) domestic
• UNC biostatistics: 11/17 (65%) under-represented minority, 52/166 (31%) international, 59/96 (61%) domestic

Being a US citizen clearly gives you an advantage: acceptance rates are about twice as high for most of these programs. However, it's not at all obvious that being female and/or an under-represented minority gives you a leg up.

[cyberwulf]

Like OP pointed, I feel there is much less "grade inflation" in Canadian universities than in the US: Is that taken into account by adcoms?

 

Also, I have heard that "internationals" are expected to hold a master's degree to apply for a US PhD: Is it usual to admit Canadian students with only a bachelor's degree (with strong profiles including RAing like OP's)?

 

To the two underlined questions:

 

1) To some extent, and the degree may depend on whether the school has experience with Canadian universities or has someone familiar with them sitting on the admissions committee. Canadian transcripts will often report class averages, so that B- doesn't look so bad when the class average is a C+.

 

2) Yes, absolutely. Many Canadian students apply and are admitted to U.S. PhD programs upon completing their Bachelors degree. Indeed, the more specialized nature of the Canadian/British system means that these students often have more discipline-specific (e.g., mathematical) coursework than their U.S., and even Chinese, counterparts.

[coffeehouse]

Thanks for the advice, but does anyone have anything more specific for the range of schools I should be applying to?

Apart from "carpet bombing the top programs"

 

Using cyberwulf's rankings in a previous thread:

 

TIER I

Stanford; UC - Berkeley; Harvard; Chicago;

 

TIER II

UW - Seattle; CMU; Duke; UW - Madison; UM - Ann Arbor; NC State;

 

TIER III

Wharton; Cornell; Columbia; Minnesota; UCLA; UNC; Yale;

 

What kind of shot do I have at the top 2 tiers (top 10)? What range of schools should I be looking at for safeties for Ph.D? What about Masters?

[biostat_prof]

Here are the most recent acceptance rates publicly available for each group where I can calculate it (although there is uncertainty about the intersection of citizenship/sex/race, and most of these combine across master's/PhD):

• UMN statistics: acceptance rates of 30/121 (25%) for females, 40/173 (23%) for males, 4/14 (29%) minority (includes Asian-American), 53/257 (21%) international, 17/42 (40%) domestic
• UMN biostatistics: 35/68 (51%) female, 28/76 (37%) male, 6/12 (50%) minority, 31/95 (33%) international, 32/49 (65%) domestic
• Duke PhD statistics: 8/75 (11%) female, 8/121 (7%) male, 0/7 (0%) under-represented minority, 8/140 (6%) international, 8/56 (14%) domestic
• UW PhD statistics: 12/156 (8%) female, 23/210 (11%) male, 8/46 (17%) minority, 10/233 (4%) international, 25/133 (19%) domestic
• UW biostatistics: 26/148 (18%) female, 22/113 (19%) male, 7/40 (18%) minority, 22/151 (15%) international, 26/110 (24%) domestic
• UNC statistics/OR: 0/7 (0%) under-represented minority, 45/386 (17%) international, 25/125 (20%) domestic
• UNC biostatistics: 11/17 (65%) under-represented minority, 52/166 (31%) international, 59/96 (61%) domestic

Being a US citizen clearly gives you an advantage: acceptance rates are about twice as high for most of these programs. However, it's not at all obvious that being female and/or an under-represented minority gives you a leg up.

 

These numbers are misleading for a couple reasons. First, it depends on how you define "minority." Sometimes Asians are classified as minorities, and the harsh reality is that being Asian will not help you in the admissions process. (Generally speaking only African-Americans, Hispanics, and Native Americans can expect an admissions boost. This is especially unfortunately for groups like Hmong or Filipinos who get lumped into the "Asian" category even though socioeconomically they are much closer to African-Americans or Hispanics, but it is what it is.) Also, the sad truth is that many minority applicants' have such weak profiles that we have to reject them because we simply don't think that they will be able to survive in the program. But I have seen my department admit minorities with a GPA in the 3.1-3.3 range. That's not going to happen for a non-minority unless the rest of their application is really spectacular.

 

As for gender, once again, you see fewer female applicants with solid grades in upper-division math courses. Unfortunate, but true. Also, being female generally doesn't help nearly as much for international students. But it can be tough to find enough qualified female U.S. citizens to maintain some semblance of gender diversity in an incoming class. We deal with this degree to an issue in my biostat department, and the problem is even more acute in stat departments, which tend to get fewer qualified female applicants. Having talked to several friends involved in admissions in various departments in the country, the bar is definitely lower for female U.S. citizens and the bar is far lower for minorities.

 

Thanks for the advice, but does anyone have anything more specific for the range of schools I should be applying to?

Apart from "carpet bombing the top programs"

 

Using cyberwulf's rankings in a previous thread:

 

TIER I

Stanford; UC - Berkeley; Harvard; Chicago;

 

TIER II

UW - Seattle; CMU; Duke; UW - Madison; UM - Ann Arbor; NC State;

 

TIER III

Wharton; Cornell; Columbia; Minnesota; UCLA; UNC; Yale;

 

What kind of shot do I have at the top 2 tiers (top 10)? What range of schools should I be looking at for safeties for Ph.D? What about Masters?

 

Well, first I feel the need to say that this type of decision should not be based on rankings and that these "tiers" are just one person's opinion. (With the general caveat that I think rankings are silly and that prospective students rely on them too heavily, at the bare minimum I think Stanford and Berkeley should be on a separate tier and Stanford is arguably a tier above Berkeley. But whatever. The fact that reasonable people can disagree about these types of rankings just shows how you shouldn't place too much faith in them.) Also, rankings don't tell you anything about the type of research that actually goes on in a department. In particular, there is no meaningful way to compare a department like Harvard (which is extremely strong in a few narrow areas) to a school like Berkeley (that has a much broader focus).

 

Having said that, I will repeat my earlier advice: Just carpet bomb all the top-ranked schools, assuming you have the time and money to do so. If your recommendations are solid, your application should get a look everywhere. The top-ranked schools will be a reach, but it's not out of the question. (And I would suggest that you apply for some of the top-ranked biostat departments as well. I generally discourage students from focusing exclusively on stat or biostat departments since the curriculum will be very similar in both types of departments.) As for "safety schools," you can apply for MS programs. My guess is that you will have almost no trouble being accepted to the vast majority of MS programs to which you apply, so I don't think you need to worry about that. (And I would definitely recommend that you apply for some biostat MS programs if only because funding sometimes exists for such programs. Funding for MS stat programs is very rare in my experience.) There really is no such thing as a "safety school" when you are applying to PhD programs. Ironically, it's sometimes more difficult to be accepted to a lower-ranked school than a higher-ranked school. Lower-ranked schools tend to have less funding, so their class sizes are smaller, and they tend to place more weight on "fit" than simply having the best credentials. (If you can only admit three students, you don't want to admit someone who will jerk you around until April 15 before choosing a higher-ranked program. Thus, you can easily get a rejection despite strong credentials if the school thinks that you see them as a "safety school.")