Statistics PhD Profile Evaluation

[huxlb]

@bayessays I've heard from professors that a good percentage of masters programs offer funding for students though. Is that not true?

[DJ3Sigma]

@huxlb I am sure that all of them or a good percentage offer some degree of funding (Master's TA, RA, GA) but if you are going to be paying out of state tuition and living expenses I am not sure that the level of funding provided by the schools at the Master's level will be enough. I am at an institution that does not guarantee funding but I am an in state student and already lived in the area. I was able to obtain scholarships for my first year which covered my tuition for the year and I did get a GA position starting the summer before my second year and it covers tuition and gives you a stipend but honestly it is not enough to live on I am just lucky that my parents were still living in the area haha. I imagine that the amount given to a TA/GA/RA varies for every school but let me tell you that the yearly contract that I currently have only offers $11k and tuition wavers for Fall, Spring, and Summer and the yearly contract offered to a PhD student in the same college is $30k+ and tuition wavers for Fall, Spring, and Summer. Now if we assume that they are not purposefully trying to make these PhD students wealthy then one can say that it takes roughly $30k in my area to lead a life that is not horrible which from what I have seen is a true statement. So you can see the great difference between what is offered to a PhD student and what is offered to a Master's student. My experience is limited to the university that I attend so if I were you I would check with the Master's programs that you are interested in and see what level of funding they offer. 

[bayessays]

10 hours ago, huxlb said:

@bayessays I've heard from professors that a good percentage of masters programs offer funding for students though. Is that not true?

Some do, but I think most the top ones don't, or at least it's often not guaranteed in the same way that PhD funding is and not offered to all students. 

[Bayesian1701]

Plus if you apply to Ph.D. program they will sometimes consider you for the masters program if you don't get in. I would apply mainly to PhD programs.

[huxlb]

I'm just worried that my profile is not competitive enough for PhD programs, and that I would be more competitive doing a Masters first and then moving on to a PhD program

[ray92]

@huxlb I think you should be pretty competitive. What you could do is scout out a faculty member you want to work under and email them and ask them if they are accepting graduate students. Some factulity profiles will give instructions on how to contact them and some may ask for a cv or resume. I would contact the faculty member directly (having a friend on the inside is nice). Also seek out graduate admission folks networking is key

[huxlb]

@ray92 Is this a common thing to do before you even get accepted? I've heard some people recommend not doing this in the past

[Bayesian1701]

The very few places (Missouri, Florida State, Iowa State) I know have funding for masters students (based on past results) are places you should be competitive for a PhD.  And if you don't make the PhD cut they should consider you for a masters.  Apply to a lot of places and you should get in somewhere.  You don't want to reach too far but you don't want to limit yourself either.  

Like I have said before we have similar profiles and I honestly believe that you are overcorrecting.  Your first list was too high maybe but you should get into a few of the programs from your list on Monday.   I definitely was very afraid of striking out.   I would have never guessed I would get into both UT and TAMU and be on the waitlist at Duke.  Applying to masters programs would be a waste of your time/money unless you don't want a PhD.  Yes admissions are competitive and random but you should be fine.  

I am not personally a fan of contacting because no one except Baylor actually responded or answered my questions.  Some people have been spotty even post acceptance. 

[ray92]

@huxlb Not only do you have to research on the graduate program you are interested in you will need research the faculty members there as well.  One major factor that goes into a decision is similar research interest as a faculty advisor. If you read faculty on-line profiles it will give instructions (usually) how to contact them if you are interested in being a PhD student. If you are nervous about emailing them or how to start I recommend reading one of their recent publications and say hey I was reading (paper) and I was interested in the area of (stats interests) and i I found it interesting how you did (math magic) but have you thought about coming from this angle.  OR you could end it by stating. I have no idea what I read but I am interested in learning more

Edited March 3, 2018 by ray92
mistake

[ray92]

13 minutes ago, Bayesian1701 said:

The very few places (Missouri, Florida State, Iowa State) I know have funding for masters students (based on past results) are places you should be competitive for a PhD.  And if you don't make the PhD cut they should consider you for a masters.  Apply to a lot of places and you should get in somewhere.  You don't want to reach too far but you don't want to limit yourself either.  

Like I have said before we have similar profiles and I honestly believe that you are overcorrecting.  Your first list was too high maybe but you should get into a few of the programs from your list on Monday.   I definitely was very afraid of striking out.   I would have never guessed I would get into both UT and TAMU and be on the waitlist at Duke.  Applying to masters programs would be a waste of your time/money unless you don't want a PhD.  Yes admissions are competitive and random but you should be fine.  

I am not personally a fan of contacting because no one except Baylor actually responded or answered my questions.  Some people have been spotty even post acceptance. 

I personally think that even if the faculty does not respond it is okay some state that they wont but they know your name and you took the time out to read their paper and email them. This shows you are interested more than someone who just applies without doing any research. They read a lot of applications you have to stand out even a simple email could do the trick. You may show up on their desk and they may say hey i remember Scott Pilgrim he emailed me last month about doing research.

[Bayesian1701]

@ray92  They may not even be on the admission committee though and they get so many emails according to the professors at my institution it may not even help if they are the committee.  They may not even read it and it's easy to fake having read a paper or interest.   I didn't even name professors in my SOP (much less email them) at three of the programs I have been accepted to.  I see your point of view but I don't agree with it.

[ray92]

@Bayesian1701 Ah let my clarify you would want to email the faculty. Admission would be a plus if you could get a hold of someone or meet them at a networking event. It is also easy to detect someone that has not read your paper and is bullshitting.

[bayessays]

Ray's advice may be useful in some disciplines, but statistics is generally not one of them - the risks of doing this outweigh the pros. I did something similar once for other reasons and luckily was met with a kind reply by the professor, but I would recommend not contacting professors unless you have a very good reason. 

[DJ3Sigma]

@huxlb If you want the PhD go for it. If you are really that worried about not getting in anywhere look at schools that are not even ranked on US news but are still in the USA. These ~30+ (this is just an estimate) schools should be easier to get in to. Personally I do not think you have to go that far but if it makes you feel better why not?

Edited March 3, 2018 by DJ3Sigma

[GoPackGo89]

@huxlb you will get an offer from one of the schools on your list. Some advice I found valuable is if only one school on your list offered you funding for a phd, would you still attend it? Eliminate schools until the answer is yes. If that shortens your list to CMU and Duke then sure maybe you need a rigorous masters to help you but even that won't be a guarantee.

Have you looked into biostatistics at all? 

[huxlb]

@GoPackGo89 not really, but it doesn’t really fall within my interests

[DJ3Sigma]

@huxlb You are interested in Bayesian Statistics and machine learning right? I am interested in those two as well and although I did not apply to biostatistics programs I can tell you from experience that both topics play a major role in biostatistics especially in the analysis of gene expression data. In fact one of Duke's programs called Banjo is frequently used in the biostats center that I work for. I did not apply to those programs because my passion is not truly in applying my research to medical data. But, honestly who do you think funds a lot of machine learning type research? NIH does and they are interested in solutions to health problems. What I realized (too late) was that my goals and the goals of the health organizations actually coincide. The organization and I have the same goal and that is for a complex problem to be solved. All they want me to add is the use of health data but who is to say that I can't go ahead and apply my solution elsewhere once I have given them what they want. My guess is that this is often times the way actual professors look at the research they do and the funding they receive. "They are paying me to look at this problem and so I will solve it how they want and then turn around and apply it to what I want." Basically what I am getting at in this long winded commentary is do not limit yourself and remember that whether you are looking at cancer data or netflix movie selection data the problem that you want to solve may still be the same one.

[StatsG0d]

7 hours ago, DJ3Sigma said:

@huxlb You are interested in Bayesian Statistics and machine learning right? I am interested in those two as well and although I did not apply to biostatistics programs I can tell you from experience that both topics play a major role in biostatistics especially in the analysis of gene expression data. In fact one of Duke's programs called Banjo is frequently used in the biostats center that I work for. I did not apply to those programs because my passion is not truly in applying my research to medical data. But, honestly who do you think funds a lot of machine learning type research? NIH does and they are interested in solutions to health problems. What I realized (too late) was that my goals and the goals of the health organizations actually coincide. The organization and I have the same goal and that is for a complex problem to be solved. All they want me to add is the use of health data but who is to say that I can't go ahead and apply my solution elsewhere once I have given them what they want. My guess is that this is often times the way actual professors look at the research they do and the funding they receive. "They are paying me to look at this problem and so I will solve it how they want and then turn around and apply it to what I want." Basically what I am getting at in this long winded commentary is do not limit yourself and remember that whether you are looking at cancer data or netflix movie selection data the problem that you want to solve may still be the same one.

Yeah I think it's worth at least looking into some biostatistics programs. UW and UNC in particular have a lot of faculty working in ML and Bayesian analysis. Shojaie at UW is kind of a young rising star in the field and was trained by Michailidis. UW is very strong in clinical trials and Bayesian statistics plays heavily in that field. Ibrahim at UNC wrote an entire book on Bayesian Survival Analysis and hundreds of papers on Bayesian statistics. There's a good thread on here that was started a few weeks ago about statistics vs. biostatistics. I recommend you read it.

[huxlb]

I have a final list of schools (I think) that I will be applying to in the fall and was wondering if anyone had a opinion on my chances at these schools.

• Virginia Tech
• Penn State
• JHU
• UGA
• Purdue 
• NC State
• Rice
• Missouri at Columbia

Thanks

Updated Profile:

Program: Statistics PhD

Interests: Design of experiments, response surface methodology, stochastic processes

Undergrad Institution: US State School

Undergraduate GPA: 3.81 (cumulative) 3.89 (major)

Undergraduate Major: Statistics

GRE: 154 V, 164 Q, 4.0 AW

Quantitative Courses: Intro Stat (A), Calculus I (A), Discrete Math (A), Calculus II (A), Intermediate Stat (A), Mathematical Reasoning (A), Linear Algebra (A), Combinatorics (A), Mathematical Computing (A), Statistical Computing (A), Differential Equations (B), Advanced Calculus (B), Multivariate Calculus (B), Probability Theory (B), Statistical Inference (A), Mathematical Modeling (A), Nonparametric Statistics (A), Regression (A), Mathematical Statistics I (B), Complex Analysis (A), Time Series (A), Applied Computing w/ R (A), Numerical Methods (A), Multivariate Statistics (A), Industrial Statistics (A)

Work Experience: 1 semester working on a research project in Biostatistics, 1 summer doing research at a national lab (which results in 1 published paper and multiple projects), 1 summer with statistical internship doing Bayesian work, Multiple semesters being a math tutor/TA

Demographics: Domestic White Female

LORs:  1 from my research advisor at national lab, 2 from professors I know very well and have promised outstanding letters

SOPs: Will most likely try to tell the story of my research and how it connects to the research I want to pursue and how that could fit into their department

Edited August 8, 2018 by huxlb

[huxlb]

Bumping this once, in case anyone has some input

[StatsG0d]

That seems like a reasonable list. I'd add maybe 2-3 more schools in the 25-35 range since I think Purdue, NCSU, PSU, and Hopkins will all be competitive (especially Hopkins).