Biostats PhD/Masters 2021: Profile Eval

[cctvwp]

Student Type: Domestic Asian Male

Undergrad: Top 5 U.S Public School

Major: Statistics

GPA: 3.63 (Major GPA is 3.8ish)

Math & stat classes: Linear Algebra (A), CALC III (A-), Differential Equations (A), Data Science (A), Optimization (A), Probability (A), Advanced Linear Models (this fall), Machine Learning (this fall), Stochastic Modeling (this fall). Intro to Programming (C+)

I goofed in my Intro to Programming course (took it Freshman year), but I've done well in all my higher level stats courses that have used R so hopefully that makes up for it!

GRE: 164Q, 160 V, 4.0W

Research: None. Was accepted into SIBS this summer but got cancelled due to COVID.

Letters of Recommendation:  All three from my Stat professors. They should be decent. 

Hi everyone, I have been reading posts on this forum for a while and am looking for feedback. I'm looking to apply for BIOS graduate programs this fall but I'm not sure if I should be applying to PhD or MS programs. 

I would ideally like to apply for PhD programs (don't really want to pay for Grad school) but I don't think my application will be strong enough due to low GPA and lack of research experience. 

Applications: Currently looking to apply to Emory (MS), UNC (MS), UCLA (MS), UC-Denver (PhD),  Houston Medical Center (Phd), Vanderbuilt (MS)

Judging my profile, do you think I'm overrating or underrating myself? Should I only be applying to MS programs even at lower ranked schools (Denver, Houston)? Or do I have a shot at PhD programs? Other program recommendations would be greatly appreciated!

 

Edited August 25, 2020 by cctvwp
Typo

[StatsG0d]

I think if you take real analysis and bump your GRE-Q score up to 167+ you could probably have a good shot at any PhD program outside the top 3. You should swap out one of those 3 courses this fall (preferably ML or stochastic modeling) and take real analysis instead.

[cctvwp]

Thanks for the replay @StatsG0d, I'm already 2 weeks into the semester so I don't think I'll be able to switch into real analysis. Will this hurt my chances even if I take real analysis in the spring?

[bayessays]

It's less of a boost to your application since the grade won't be on there, but I'd definitely take it and not worry about it.  Real analysis isn't a hard requirement for biostatistics programs, especially for lower-ranked programs.  Even some people in top programs just take it their first year anyways.  I think your list of schools is definitely reasonable.

[cctvwp]

@bayessays I don't really want to pay for masters programs, do you think I would be a competitive applicant for PhD programs at schools like Emory/UCLA?

[bayessays]

I think they're worth applying to, but not guaranteed and you should also have lower ranked schools similar to the PhD programs in your original post

[cctvwp]

Should I just apply to all PhD programs then? If I applied to all PhDs, would I also be considered for Masters programs?

[bayessays]

17 minutes ago, cctvwp said:

Should I just apply to all PhD programs then? If I applied to all PhDs, would I also be considered for Masters programs?

Usually, yes, there is an option to be considered for the MS program but you'll have to look on each individual application to figure out the details.

[StatsG0d]

2 hours ago, cctvwp said:

Thanks for the replay @StatsG0d, I'm already 2 weeks into the semester so I don't think I'll be able to switch into real analysis. Will this hurt my chances even if I take real analysis in the spring?

Yes it will definitely hurt your chances compared to if you were able to take it this semester. Without having the grade, you probably have a shot at programs ranked Michigan and below.

I will mention UCLA is pretty competitive relative to its ranking. 

[cctvwp]

@cyberwulf any advice on what range of schools I should be applying to if I'm looking to apply for PhD Bios programs?

[cyberwulf]

On 8/27/2020 at 8:41 PM, cctvwp said:

@cyberwulf any advice on what range of schools I should be applying to if I'm looking to apply for PhD Bios programs?

I agree with the above advice. You should take a shot at a couple of top 10 PhD programs (which will almost surely admit you for a Masters if you don't get into the PhD) and probably focus your apps on programs in the 10-20 ranking range.

[cctvwp]

Update: admitted to UNC!

Edited March 5, 2021 by cctvwp

[csheehan10]

On 8/25/2020 at 1:30 PM, StatsG0d said:

I think if you take real analysis and bump your GRE-Q score up to 167+ you could probably have a good shot at any PhD program outside the top 3. You should swap out one of those 3 courses this fall (preferably ML or stochastic modeling) and take real analysis instead.

Which ones do you mean by the top 3? Is there a definitive ranking for stats phd programmes?

[bayessays]

7 minutes ago, csheehan10 said:

Which ones do you mean by the top 3? Is there a definitive ranking for stats phd programmes?

Within biostat specifically, Washington, Hopkins, Harvard a considered a step above other programs. 

[csheehan10]

3 minutes ago, bayessays said:

Within biostat specifically, Washington, Hopkins, Harvard a considered a step above other programs. 

oh fair, I don't know much about biostats, i was just thinking of pure stats.

[bayessays]

1 hour ago, csheehan10 said:

oh fair, I don't know much about biostats, i was just thinking of pure stats.

Besides Stanford being #1 by a mile, I think the rest of the top 10 on US News rankings are pretty similar and there isn't a clear of a hierarchy as in biostatistics.  In statistics, there are renowned professors at departments ranked in the 40s, and you could reasonably attend a school like that over a top 10 program.  But in biostatistics, the drop-off is much more severe.

[statsnow]

@bayessays I think a lot of people would disagree with you that Stanford is #1 by a mile.  A lot of people for a variety of reasons think Berkeley is far superior.  What do you base your opinion on?  Also there are a number of differences between the schools in the top ten.  

[bayessays]

10 hours ago, statsnow said:

@bayessays I think a lot of people would disagree with you that Stanford is #1 by a mile.  A lot of people for a variety of reasons think Berkeley is far superior.  What do you base your opinion on?  Also there are a number of differences between the schools in the top ten.  

Berkeley perhaps is much closer, you are right. But does Berkeley have five people who come close to the influence of Efron, Hastie, Tibshirani, Diaconis, Candes, and Donoho?  I can understand thinking they are similar, but why do you think Berkeley is *far* superior?

[Stat Phd]

On 8/25/2020 at 1:25 AM, cctvwp said:

Student Type: Domestic Asian Male

Undergrad: Top 5 U.S Public School

Major: Statistics

GPA: 3.63 (Major GPA is 3.8ish)

Math & stat classes: Linear Algebra (A), CALC III (A-), Differential Equations (A), Data Science (A), Optimization (A), Probability (A), Advanced Linear Models (this fall), Machine Learning (this fall), Stochastic Modeling (this fall). Intro to Programming (C+)

I goofed in my Intro to Programming course (took it Freshman year), but I've done well in all my higher level stats courses that have used R so hopefully that makes up for it!

GRE: 164Q, 160 V, 4.0W

Research: None. Was accepted into SIBS this summer but got cancelled due to COVID.

Letters of Recommendation:  All three from my Stat professors. They should be decent. 

Hi everyone, I have been reading posts on this forum for a while and am looking for feedback. I'm looking to apply for BIOS graduate programs this fall but I'm not sure if I should be applying to PhD or MS programs. 

I would ideally like to apply for PhD programs (don't really want to pay for Grad school) but I don't think my application will be strong enough due to low GPA and lack of research experience. 

Applications: Currently looking to apply to Emory (MS), UNC (MS), UCLA (MS), UC-Denver (PhD),  Houston Medical Center (Phd), Vanderbuilt (MS)

Judging my profile, do you think I'm overrating or underrating myself? Should I only be applying to MS programs even at lower ranked schools (Denver, Houston)? Or do I have a shot at PhD programs? Other program recommendations would be greatly appreciated!

 

I think you are underestimating yourself, you should also try some top programs like UW, UM, and UNC.

[StatsG0d]

On 3/5/2021 at 11:32 PM, statsnow said:

@bayessays I think a lot of people would disagree with you that Stanford is #1 by a mile.  A lot of people for a variety of reasons think Berkeley is far superior.  What do you base your opinion on?  Also there are a number of differences between the schools in the top ten.  

If people disagree then it's just their ego talking. The people @bayessays mentioned have revolutionized the field. You will scarcely find a single person, let alone multiple, that have had the influence those people have had at most departments.

 

On 3/5/2021 at 1:47 PM, bayessays said:

Besides Stanford being #1 by a mile, I think the rest of the top 10 on US News rankings are pretty similar and there isn't a clear of a hierarchy as in biostatistics.  In statistics, there are renowned professors at departments ranked in the 40s, and you could reasonably attend a school like that over a top 10 program.  But in biostatistics, the drop-off is much more severe.

I somewhat agree with this. I personally think Harvard's program is slightly overrated because they are lacking in theoretical training compared to Hopkins and (especially) Washington. It is true that outside the top-5 (UW, JHU, Harvard, UNC, Michigan), there's a huge drop-off, but I don't think anyone will argue that of those 5, Michigan is, by a substantial amount, the best place to be to do genetics research (and probably the worst place of the top-5 to do much of anything else). Similarly, I don't think many people would disagree that Harvard is the best of the 5 at causal inference. The other 3 are probably the best options if you're not sure what you want to do (which I think is the case of most people whether they are aware of it or not). UNC's main pull is that they have a faculty with largely diverse research interests, but I doubt you'll find anyone who truly believes their department is superior to UW / JHU in any way other than student paper awards.

[statsnow]

@statsGod  I think if you want to learn theoretical statistics you are probably correct about Stanford.  However a department is not totally defined by a few stars.  There are other factors that go into an education .  Michael Jordan is the leader in artificial intelligence and a few other areas. That is greatly valued by many people.  Practical application of statistics and methodology is important to very many people.  I am not sure there are many people at Stanford that do Causal Inference.   Departments hire by what is their perceived needs and that can vary from year to year.   Also the lack of diversity at Stanford can and has been seen as a negative by many people.    

 

 

[StatsG0d]

11 hours ago, statsnow said:

@statsGod  I think if you want to learn theoretical statistics you are probably correct about Stanford.  However a department is not totally defined by a few stars.  There are other factors that go into an education .  Michael Jordan is the leader in artificial intelligence and a few other areas. That is greatly valued by many people.  Practical application of statistics and methodology is important to very many people.  I am not sure there are many people at Stanford that do Causal Inference.   Departments hire by what is their perceived needs and that can vary from year to year.   Also the lack of diversity at Stanford can and has been seen as a negative by many people.    

 

 

The question wasn’t what defines a department nor did I claim that purely theoretical statistics is the only thing that matters. The question was how far ahead is Stanford compared to the others. 

Do not confuse what I’m saying to think that Berkeley is a bad department, it’s just a far cry away from Stanford, but everywhere except for Stanford is. That’s the point of @bayessaysand me. 
 

Also, I would caution saying that Jordan is *the* leader in AI. AI is a broad field, and one of his students, Ng, who has also revolutionized AI, is affiliated with Stanford. 
 

I think it’s totally fair to criticize Stanford for its lack of diversity, but unfortunately that’s not how department prestige is determined. Moreover, the lack of diversity is a problem across the field—certainly not Stanford specific. 
 

Finally, to your point about a diversity of research areas (e.g. causal inference), this is a fundamental problem with rankings. See my comment above regarding Michigan biostatistics, which I think everyone would agree is the best place to do genetics research in the world. If a diversity of research areas factored into rankings, Purdue, TAMU, and PSU would be at the top every year simply by having huge departments. 

[Stat Assistant Professor]

Yeah, Stanford truly is in a class of its own. No other program has as many faculty who revolutionized the field -- e.g. Tibshirani, Hastie, Johnstone, Candes, Efron, Friedman, etc. Lasso, bootstrapping methods, compressed sensing, additive models (including gradient boosting), etc. were basically invented at Stanford.

I would note also that Stanford is also very strong in applied statistics. Efron is the founding editor of Annals of Applied Statistics, and Stanford also has people like Susan Holmes, Trevor Hastie, and Robert Tibshirani. Some of these folks have also worked on statistical theory too, but they also work a lot on computational statistics and on applications in biostatistics, bioinformatics, and health policy.

[statsnow]

Actually gradient boosting  came from Berkeley from Leo Breiman if history is correct.  Freidman also got his phd at Berkeley .  Bartlett who is at Berkeley now developed functional gradient boosting which is a more general concept.  It was developed at the same time Friedman published his papers.  I dont really think you can say that one group revolutionized the field if you look at all the historical facts

[Stat Assistant Professor]

Gradient boosting is essentially an additive model tailored to decision trees, and the concept of additive models was first developed by Friedman and Stuetzle at Stanford. It is possible that somebody else suggested the idea of boosting for tree-based models, but the gradient boosting machine (GBM) paper that gets cited the most often was written by Friedman when he was at Stanford. The generalized additive model (GAM), an important development in nonparametric regression, was also developed by two Stanford statistics faculty, Hastie and Tibshirani (albeit this was before they joined the Stanford faculty).

Nobody is claiming that there are *no* people who have revolutionized the field at other programs. Obviously, at UC Berkeley, you have Michael Jordan and Martin Wainwright, and Lucien Le Cam also spent his career at UCB. But there is certainly a higher concentration of such "field-changing" folks at Stanford than at any other school. Bootstrapping, compressed sensing, lasso/L1 regularization methods, additive models -- all very "revolutionary" developments in Stats -- came from people who are at Stanford.

Edited March 13, 2021 by Stat Assistant Professor