Profile Evaluation: 2020 Stats PhD

[MrDonkey]

Hello everyone, 

I'm currently a senior planning on applying to PhD programs in statistics or any related field (eg. operations research) for Fall 2020.

Undergraduate Institution: UCLA

Majors: Mathematics, Statistics (Double major)

GPA: 4.00/4.00 (cumulative) 

Type of Student: Domestic Asian Male

Relavant Courses:

Math: 

• Calculus I, II, III (A+, A+, A+)
• Linear Algebra (lower & upper level) (A+, A+)
• Differential Equations (lower level) (A+)
• Honors Real Analysis I, II, III (A, A+, A+)
• Honors Complex Analysis (A+)
• Applied Numerical Methods (A+)
• Probability Theory (A+)
• Intro to Financial Math (A+)

Statistics:

• Programming with R (A+)
• Intro to Mathematical Statistics (A)
• Linear Models (A)
• Intro to Data Analysis Regression (A+)
• Intro to Design and Analysis of Experiment (A+)
• Intro to Regression and Data Mining (A+)
• Intro to Computational Statistics with R (A+)
• Intro to Computation and Optimization (A+)
• Intro to Monte Carlo Methods (A)
• Python for Data Analysis (A+)

Programming:

• Intermediate Programming C++ (A+)
• Java Language and Application (A+)
• Python for Applications (A+)

GRE General:  165V/169Q/Writing unknown.

GRE Math Subject: 940 (99%)

Research Experience:

• No formal research experience, but will start a research beginning this fall under an assistant professor in the Stats/Communications department on computer vision
• A directed research project for two quarters under the guidance of a stats lecturer (more of a project than of research) on the topic of computer vision
• A Math Directed Reading Program for a quarter under a math PhD student on the topic of linear PDE.

Work Experience:

• Three internships, one in finance, two in actuarial science (one in risk management with a ML project, the other is accounting related)
• One part-time job as lab assistant tutoring programing in a computer lab.

Letters of Recommendation:

• Have not asked yet, but I am thinking about choosing from the following options:
• Three math professors that know me well from my Honors Analysis class (one is a tenure who taught me two real analysis classes, one adjunct who taught one real analysis, and one adjunct who taught me complex analysis) where I got A+ in almost all of them
• One stats lecturer with whom I did a directed research project on computer vision with and whom I took three stats classes (2 A+, 1A)
• One stats lecturer with whom I took two mathematical statistics classes with (I impressed him by getting ~90 on all the exams when the averages were all ~30)

 

As you can see, the weakness in my profile is the lack of research experience, despite of good GPA, GRE and GRE sub. So I am wondering what level of schools do I have a good chance into? I don't have a school list yet, and would like to seek feedback of the range of schools you recommend (like for example, what is the chance of me getting into phd programs of #10-20 ranking schools or what are the specifics schools that I have a good chance for?) and if there are other programs that fit my background other than statistics (eg. should I apply to operations research or something else?).

Thanks for your input!

[Stat Assistant Professor]

Your profile is outstanding. UCLA has one of the best math departments in the world (#7 in USNWR), and your grades and test scores are outstanding.

The lack of research experience should not pose an issue, but just one question: you say that you are getting two letters of recommendation from "lecturers." Are these not tenured professors or tenure-track Assistant Professors? It would be far more preferable for you to get strong letters from tenured or TT professors (not lecturers or adjuncts) who can credibly speak to your research potential. Also, you do not need to submit more than three LORs, and the adcoms may only look at three of them (given time constraints), so I would recommend submitting three of the strongest ones you can from professors. They do not need to come from stats professors; they could all be from math profs.

Anyway, if you apply to all the top 10 programs AND get strong letters of recommendation, you will probably get into at least a few of them.

Edited September 2, 2019 by Stat PhD Now Postdoc

[MrDonkey]

29 minutes ago, Stat PhD Now Postdoc said:

The lack of research experience should not pose an issue, but just one question: you say that you are getting two letters of recommendation from "lecturers." Are these not tenured professors or tenure-track Assistant Professors? It would be far more preferable for you to get strong letters from tenured or TT professors (not lecturers or adjuncts) who can credibly speak to your research potential. Also, you do not need to submit more than three LORs, and the adcoms may only look at three of them (given time constraints), so I would recommend submitting three of the strongest ones you can from professors. They do not need to come from stats professors; they could all be from math profs.

Thanks! The fact that they are stats "lecturers" are because almost all those who teach stats classes at UCLA are lecturers. The tenure-track (assistant) professors in our stats department hardly teach any classes so I do not know any of them. I was just thinking that since one of them has title "senior continuing lecturer," which basically is a "tenure" lecturer, they might be better than adjunct math professors (who are basically post-docs).

[bayessays]

If they know you well and they have PhDs, I wouldn't worry about it.  From the sound of it, I'd take the tenured analysis prof and the two stats people - most people don't have famous statisticians writing their letters, and it's more important that they know you well. Agreed with above that you'll almost certainly get into multiple top 10 programs. 

[Stat Assistant Professor]

I've read and heard in many places that it is best to get letters from TT or tenured faculty whenever possible, and some are quite adamant about this, e.g.: https://www.math.uh.edu/~tomforde/KissesOfDeath.html

I think to be safe, the OP should go with at least two letters from TT or tenured faculty. One letter from a statistics lecturer who is not a TT/tenured professor may be okay, but I would say that it is best to err on the side of caution and try to get at least two letters from TT/tenured faculty. The assistant professor who is supervising the OP on computer vision research would be an excellent choice for a LOR, in my opinion. 

[bayessays]

Yes, I 100% agree with you overall, but looking at the particular situation here, I believe these people have PhDs in statistics (whereas lecturers at many places do not) and he has had significant long-term interaction with them including the computer vision research.  Since "did well in class, don't know much else" letters are essentially useless, OP has to get by with what they have.  After the ideal case of TT math/stat professors whom you have had significant interaction with, I think other statistics PhDs who know your work well is about as close to an ideal letter as you can get.  A UCLA lecturer could probably easily get a job teaching at a small LAC if they wanted to, and I don't think the shift in job title would make their letter more credible.  But yes, I agree people should try to get letters from the most prestigious/qualified people possible, all else being equal.

[ipsumlorem]

Just to piggyback on this discussion, what do you guys think about letters of recommendations from industry? Is there a criteria for that as well, or is it just not recommended at all?

[Stat Assistant Professor]

25 minutes ago, ipsumlorem said:

Just to piggyback on this discussion, what do you guys think about letters of recommendations from industry? Is there a criteria for that as well, or is it just not recommended at all?

If it is for a PhD application, letters of recommendation from industry would probably not be a good idea. You need letters from people who can credibly evaluate your academic research potential.

In general, I would recommend at least two strong letters of recommendation from tenured or tenure-track professors (not postdocs, grad students, or non-tenured professors) -- professors who have academically evaluated you in upper division classes or supervised you on some research project(s) and who can write glowing LORs on your behalf. The third letter can be somewhat generic and/or from someone who is not tenured/TT faculty without being detrimental to the application as a whole (as long as it is not terrible) -- though of course, if you can get three excellent letters that all come from professors, that would be the most ideal situation.
 

Edited September 2, 2019 by Stat PhD Now Postdoc

[bayessays]

2 hours ago, ipsumlorem said:

Just to piggyback on this discussion, what do you guys think about letters of recommendations from industry? Is there a criteria for that as well, or is it just not recommended at all?

I am glad we are getting diverse opinions about this in here, because adcoms are made up of people with different opinions too.  Again, I agree with @Stat PhD Now Postdoc broadly, but think there are exceptions. If you are a data scientist at Facebook or Google, and your boss is a published PhD statistician, I think they are in a great position to evaluate your potential for success in a graduate program. Granted, this is a super rare exception to the rule, just like it's a rare exception for someone to do in-depth research with a non-tenured professor.  If someone has a PhD in statistics and you have done research-like activities with them, IMO those are the best type of letter you could get.  But getting a letter from a boss without a PhD and whom you haven't done real statistics stuff with will not be useful and will probably make you come off as uninformed about what you're getting into.

But in statistics, which is a field with tons of options in industry and in non-TT research positions, your job title often reflects personal choice more than academic success.  Anyone with a PhD spent years taking classes and doing research and finished their degree successfully, so I don't see how their job choice after graduating would affect their credibility as a letter writer, outside of the fact that they're probably not a famous person.  I've had bosses in industry who are immensely more qualified than my undergraduate TT professors and most professors at smaller colleges.  But perhaps some people do hold this bias, so it is good to be aware of it. 

 

Edited September 2, 2019 by bayessays