2020 Fall Stat/Bio-stat PhD Profile Evaluation - Non Traditional Applicant

[warmest]

Hello everyone,

I am planning to apply for PhD programs in Statistics and/or Biostatistics. I am seeking advice about the programs I should target. I would like advice based on my background below. Thank you a lot for your time! 

 

A quick summary about me - I have undergraduate degree in environmental science and master’s degree in applied stat. I graduated in 2019 and have one-year work experience. I attach my questions at the end.

Type of Student: International, Asian

Undergraduate Institution: Asian, top-50 by US news ranking  

Majors: Environmental Science 

GPA: 3.46/4.0 by WES (84/100) 

Courses taken (score by WES):

• CS
  • Intro to Computation (A)
  • Data Structure and Algorithm (B) 
• Math
  • 2 calculus courses (A and B）
  • Linear Algebra (A)
  • Probability Theory and Statistics (A)
• Stat
  • Applied Statistics (B) 

Graduate Institution: Master’s degree, top-20 by US news ranking in stats major 

Majors: Applied Statistics (courses only, no thesis), dual degree in environmental science

GPA: 3.6/4.0

• Stat 400-level
  • Intro to Theoretical Stat (B+)
  • Applied Probability (B) 
• Stat 500-level
  • Stat Learning I (A-)
  • Stat Learning II (A-)
  • Stat Inference (A)
  • Stat Consulting (A-)
  • Probability Distribution Theory (A-)
  • Stat Inference (A) 
• Stat 600-level
  • Linear Models (B+)
  • Survival Time Analysis (A)
  • Stat Computing (A-) 

GRE General Test: Q: 170 (96%); V: 152 (56%), W: 3.5 (41%) 

Research Experience

• Second author of a paper that is under peer review now. I did plenty of map visualization on the topic of air pollution. No modern stats method applied. (*big name prof. in biostatistics, not sure how the rec would be)
• One presentation at the School Symposium about a minor project, using spatial statistics method to answer an environmental question.
• Undergraduate thesis is about applying non-linear optimization method in a environmental problem. (*prof. in environmental science, good but not strong rec) 
• Graduate capstone is much like consulting about clean energy, no related to stat. (*prof. in environmental science, might be strong rec)
• Research assistant for a social scientist for half year. I mainly cleaned data in R and SAS, and did some simple tests. (*prof. in social science, strong rec)

Working Experience

• One-year work experience in a software company. Daily work includes programming in Python, R, Git; researching for spatial-temporal statistical methods, such as time series forecast, spatial outlier detection. (*supervisor)

Letters of Recommendation

I listed the potential recommenders with asterisk* in the Research/Work Experience above. 

I can reach out to prof. who taught stat courses but that would be weak rec I guess. 

Currently considering schools

Planning to focus on 20-50 tiers schools in the U.S., such as UCLA, UC Irvine, UC Davis, OSU, Boston U, UUC. In addition, I am planning to apply for British Columbia and McGill in Canada. 

 

My questions

Here is my main question - I like statistics, and based on by background in environmental science, I think biostatistics would provide a good balance between theory and application for me. Yet with little research experience in biostats, I find it hard to compare the programs. What should look for when I scan through the  programs? Any advice about how to filter programs would be appreciated! 

I have another concern about the targets – I am worried my low GPA and weak background in math would become a constraint. Is there anything I should be aware about? For example, does programs usually filter applicants' GPA first?  

In addition, I would love to hear about who you think would make the best combo of recommenders. The fact that most of my rec would not come from prof. in stat concerns me as well. 

Thank you in advance for your input! 

   

Edited September 13, 2020 by warmest

[Casorati]

Your mathematical background is too light to apply for PhD in statistics. At minimum, you should have done one course in real analysis and many applicants have taken much more proof-based courses. Admissions committees care most about your grades in math/stat courses so your undergraduate GPA won't carry much weight. However, your graduate GPA is mediocre/below average given the grade inflation in grad school so that also doesn't help. With that said, most schools you listed are unrealistic. I am attending one of UBC/McGill and most students are math/statistics major from top universities in Canada or USTC, Zhejiang University in China, so I don't think your odds are good at UBC/McGill either. I would mainly apply to biostatistics PhD programs ranked below 50 since lower ranked biostatistics programs are more applied and more lenient on math requirement. The Biostatistics PhD program at UToronto is very applied and I think you have a chance to get in.

Edited September 14, 2020 by Casorati

[warmest]

On 9/13/2020 at 5:14 PM, Casorati said:

Your mathematical background is too light to apply for PhD in statistics. At minimum, you should have done one course in real analysis and many applicants have taken much more proof-based courses. Admissions committees care most about your grades in math/stat courses so your undergraduate GPA won't carry much weight. However, your graduate GPA is mediocre/below average given the grade inflation in grad school so that also doesn't help. With that said, most schools you listed are unrealistic. I am attending one of UBC/McGill and most students are math/statistics major from top universities in Canada or USTC, Zhejiang University in China, so I don't think your odds are good at UBC/McGill either. I would mainly apply to biostatistics PhD programs ranked below 50 since lower ranked biostatistics programs are more applied and more lenient on math requirement. The Biostatistics PhD program at UToronto is very applied and I think you have a chance to get in.

Thanks a lot for your reply. I am also wondering if I take a few more proof-based courses, such as real analysis, through non-degree program, and apply next year, do you think that would change the odd of being admitted to 30-50 ranking schools? 

[Casorati]

You need to convince the admissions committee that you have strong mathematical ability so if you do well in them, it certainly would help. You can apply to some 30-50 range schools but there is no guarantee that you will get in. Even these schools admit very strong international students.

Edited September 18, 2020 by Casorati

[StatsG0d]

7 hours ago, Casorati said:

You need to convince the admissions committee that you have strong mathematical ability so if you do well in them, it certainly would help. You can apply to some 30-50 range schools but there is no guarantee that you will get in. Even these schools admit very strong international students.

I agree with @Casorati. Many 30-50 programs (e.g., Florida, FSU) are considered safeties by applicants from schools renowned for developing strong PhD students, such as Peking, Tsinghua, and ISI. These students have very deep math backgrounds.

I do not think simply taking analysis will cut it. I think you would need to take Analysis I-II and complex analysis, at a minimum, to be competitive. If you throw in abstract algebra or some other high level proof based class, it would help even more. You're probably looking at a 1.5 year commitment in the best case scenario. The only other thing you could do is to get a very high score on the math subject GRE.