biostatistics programs: logistical questions

[prospectivestudent2013]

hello,

 

I am currently attempting to decide between the M.S. program at Mich, Ann Arbor and the Ph.D. program at Pitt. I would like to pursue an academic career (or at least have the option of doing so). Mich unquestionably seems to be the better option in this regard; however, funding decisions remain in the air. 

I'm posting this because my situation has led me to two questions that may be of interest to future applicants:

 

(1) If one were to enter a funded Ph.D. program (in my case Pitt), how detrimental would it be to attempt to transfer to a stronger program after earning the M.S.? An obvious problem is related to letters of recommendation; also, academic communities tend to be relatively small, and one wouldn't want to alienate one's peers...

 

(2) If switching programs is not a viable option, how difficult would it be to enter an (ideally upper-mid tier) academic position having come from Pitt (or a similarly positioned program)? Might it be better to simply wait til next cycle given the goal of working as an academic?

To provide more specific information, my limitations arise primarily because of a lack of mathematics courses. My GRE was stellar, GPA solid, and research activity high. The benefit from taking an extra year would likely be considerable. 

Thanks, and I would greatly appreciate any feedback.

[cyberwulf]

This is a common question without an easy answer. While some fraction of students do 'upgrade' to a stronger PhD program after completing an MS, the reality is that admissions committees consider the entirety of a student's academic record when making PhD admission decisions, and so good performance in an MS program doesn't negate mediocre undergraduate performance. The following student types are most likely to benefit from completing a Masters degree from a highly-ranked department before applying to PhD programs:

 

Type I: Students with very strong records (e.g., GPA 3.8-3.9+) from relatively unknown undergraduate institutions

Type II: Students with excellent academic credentials, but severely lacking in mathematical preparation

 

Those least likely to benefit are:

 

Type III: Students with solid but unspectacular records with adequate mathematical preparation from good schools.

 

As a general rule, adcoms at top-tier departments are seeking talent and intelligence more than a particular set of statistical skills. With Type I students, the talent/intelligence level may be unclear because of uncertainty about the level of competition and rigor they faced as undergraduate students. Strong performance in an elite Masters program can go a long way towards showing that the excellent performance wasn't just a mirage. With Type II students, the Masters may be useful to confirm that their talents extend beyond a single field to encompass mathematics/statistics as well. With Type III students, good performance in a Masters program is basically expected, and adds relatively little information about that student's abilities. 

 

Of course, every individual's situation is different, but I think that students too often fall into the trap of thinking that stellar performance in a Masters program will cause everyone to ignore the rest of their academic record, while that is rarely if ever the case. 

[chrishacker]

This is a common question without an easy answer. While some fraction of students do 'upgrade' to a stronger PhD program after completing an MS, the reality is that admissions committees consider the entirety of a student's academic record when making PhD admission decisions, and so good performance in an MS program doesn't negate mediocre undergraduate performance. The following student types are most likely to benefit from completing a Masters degree from a highly-ranked department before applying to PhD programs:

 

Type I: Students with very strong records (e.g., GPA 3.8-3.9+) from relatively unknown undergraduate institutions

Type II: Students with excellent academic credentials, but severely lacking in mathematical preparation

 

Those least likely to benefit are:

 

Type III: Students with solid but unspectacular records with adequate mathematical preparation from good schools.

 

As a general rule, adcoms at top-tier departments are seeking talent and intelligence more than a particular set of statistical skills. With Type I students, the talent/intelligence level may be unclear because of uncertainty about the level of competition and rigor they faced as undergraduate students. Strong performance in an elite Masters program can go a long way towards showing that the excellent performance wasn't just a mirage. With Type II students, the Masters may be useful to confirm that their talents extend beyond a single field to encompass mathematics/statistics as well. With Type III students, good performance in a Masters program is basically expected, and adds relatively little information about that student's abilities. 

 

Of course, every individual's situation is different, but I think that students too often fall into the trap of thinking that stellar performance in a Masters program will cause everyone to ignore the rest of their academic record, while that is rarely if ever the case. 

Well, I am an econ major and I plan to transfer to stat MA. I wonder whether they will give my undergrad courses a less weight. I have a GPA less then 3.5 ( You know in some places outside US...) If it is still very important then  .......

Edited March 27, 2013 by chrishacker

[biostat_prof]

If you are dead set on academia, my guess is that your best bet is to do the MA at Michigan. What cyberwulf says is true to a point, but my (highly ranked department) recently admitted a student who had a downright ugly undergraduate record but absolutely killed it in an MS program. I probably can't provide any more details for privacy reasons, but it was a really terrible undergraduate record but they did spectacularly well as an MS student. It was enough to convince the committee to overlook the undergraduate record. So if you enroll in an MA program and do well, it certainly can boost your chances of being admitted to a good PhD program and later finding an academic job. I don't have any hard data about Pitt's placement in academia, but my guess is that it's going to be tough to find a job in a strong department coming out of Pitt. A quick glance at their faculty's web pages indicates that very little (if any) methodological work is being done in that department. I suppose transferring out of Pitt is a possibility, but you would need to do very well (and asking for recommendations could be awkward).

 

That said, think long and hard about whether you really want to be an academic bad enough to pursue this path. Even attending a fully funded PhD program at a top-ranked school is a questionable decision from a strictly financial perspective. Most people will earn more over the course of a career with an MS, to be honest (once you consider the additional 3-4 years of lost earnings to get a PhD). So from a financial perspective, going into debt to get a PhD is almost always a bad decision. Personally I don't recommend it unless you want an academic job so badly that you're willing to live like a student for a very long time (probably even after you finish your PhD) or you are independently wealthy. Also remember that admission to a top-ranked PhD program after earning an MA from Michigan is not guaranteed, nor are you guaranteed to land an academic job after attending a top-ranked PhD program. You will have to decide if you want to make this gamble. Good luck with your decision.

[cyberwulf]

Even attending a fully funded PhD program at a top-ranked school is a questionable decision from a strictly financial perspective. Most people will earn more over the course of a career with an MS, to be honest (once you consider the additional 3-4 years of lost earnings to get a PhD). 

 

I'm going to disagree with you here. In my geographical area, starting salaries for M.S. students in private industry average around $70k, while PhDs start near $100k. Even if you figure four years of lost earnings = $280,000 (which is an overestimate for a funded PhD student earning a stipend of $20-$25,000 per year), a PhD makes up that difference within 10 years, and probably less given that the salary gap is likely to increase over time.

[biostat_prof]

I'm going to disagree with you here. In my geographical area, starting salaries for M.S. students in private industry average around $70k, while PhDs start near $100k. Even if you figure four years of lost earnings = $280,000 (which is an overestimate for a funded PhD student earning a stipend of $20-$25,000 per year), a PhD makes up that difference within 10 years, and probably less given that the salary gap is likely to increase over time.

 

I want to know where every PhD graduate gets $100k per year in industry. My department is highly ranked and has some strong ties to industry, but a $100k starting salary isn't even close to the market rate in industry around here. Someone with a strong record who gets a job in Big Pharma might get $100k, but many of our PhD graduates in industry start out at lower salaries than our MS graduates. Remember that the market for PhD statisticians is much tighter than the market for MS statisticians, so if you are picky at all about where you want to live, you're probably going to take a big salary hit. I know several recent PhD graduates who have ended up working for $20 per hour without benefits as contractors at local companies. Granted, all of them applied for jobs only in one city, and all of them eventually found better jobs. But saying that anyone with a stat/biostat PhD will earn 6 figures at graduation is grossly unrealistic; no offense. And that also doesn't take into account the fact that many PhD graduates go into academia, where they may only get a $70k starting salary, or even less at teaching-oriented schools (particularly in a staistics department). I agree that the upside for PhD's is higher, but most (virtually all?) PhD graduates won't earn the salaries at the high end of the scale. I saw a study once that said that people with an MS earn about 25% more than those with comparable experience who only had a BS, but the difference between a BS is a PhD was only about 28%. (I can try to find the study if anyone is interested.) Thus, the vast majority of PhD graduates will not earn enough additional salary to justify several years of lost earnings. Hence, my advice to students is always to pursue a PhD only if you love research, because for most people it's a bad decision financially.

[cyberwulf]

I want to know where every PhD graduate gets $100k per year in industry. My department is highly ranked and has some strong ties to industry, but a $100k starting salary isn't even close to the market rate in industry around here. Someone with a strong record who gets a job in Big Pharma might get $100k, but many of our PhD graduates in industry start out at lower salaries than our MS graduates. 

 

Just to be clear, I never said that 6 figures was a stone cold lock for any biostat PhD graduate, only that it seems like roughly the average for our graduates entering private industry (mostly big pharma or other medical research companies).  I'm surprised to hear that in your area MS students are earning more than PhDs, because that doesn't appear to be the case in our neck of the woods. Do you think it's a function of your being in a region which is flush with PhD statisticians? Why would companies pay more for Masters statisticians if PhD statisticians were available?

 

many PhD graduates go into academia, where they may only get a $70k starting salary, or even less at teaching-oriented schools (particularly in a staistics department).

 

Salaries for new assistant profs in academic biostat departments are much higher than $70k; indeed, the latest figures from the  AmStat News biostatistics salary survey (see below) shows that the median starting salary for a newly hired faculty member is... $102,000. Presumably these numbers are being driven upwards by salary pressure from private industry. 

 

Edited March 29, 2013 by cyberwulf

[cyberwulf]

And, FWIW, here's the AmStat News survey results for statisticians in government, industry and business (salaries are in $1,000s):

Edited March 29, 2013 by cyberwulf

[wine in coffee cups]

Really interesting discussion. One thing I wonder is how generalizable those high academic starting salaries in biostatistics actually are to biostatistics PhDs, though? My university's biostatistics recent hires and job candidates seem to mostly have been trained at the top theoretical statistics programs (whether direct from grad school or after a postdoc). My suspicion is that if I were to hunt down the 15 respondents who started in biostatistics departments last fall and 19 from the year before is that maybe half of those actually have biostatistics degrees and that they would come from just a few universities.

[cyberwulf]

Really interesting discussion. One thing I wonder is how generalizable those high academic starting salaries in biostatistics actually are to biostatistics PhDs, though? My university's biostatistics recent hires and job candidates seem to mostly have been trained at the top theoretical statistics programs (whether direct from grad school or after a postdoc). My suspicion is that if I were to hunt down the 15 respondents who started in biostatistics departments last fall and 19 from the year before is that maybe half of those actually have biostatistics degrees and that they would come from just a few universities.

 

Excellent point. 

[biostat_prof]

First, all students should not that those AmStat numbers are wildly optimistic. You would think that statisticians would be smart enough to worry about nonresponse bias when they conduct a survey like this, but apparently not. (Actually, I think the purpose of this survey is to help people negotiate starting salary offers rather than to provide an accurate picture of the market, so AmStat deliberately ignores the nonresponse bias problem.) The issue is that people who earning six figures in Big Pharma or a tenure-track job at Harvard are far more willing to fill out this type of survey than someone who is doing contract work for $20 an hour. I think some Google research indicates that the response rate for the AmStat salary survey is around 33%, so take it with a very healthy grain of salt. The median starting salary for a biostat PhD in academia is $102K? Really? I'm not sure whether to laugh at the absurdity of that number or cry that some students might believe it. I don't want to potentially embarrass any faculty members by publicly posting their salaries on this board, but suffice it to say that the salaries for all faculty members (including biostatistics faculty members) at UW, Michigan, and UNC are publicly available. You can find these databases and enter in the names of a few of their junior faculty and you'll see most of them are around $100k (and less in some cases). If the very best departments are paying around $100k, that indicates that the median is probably lower than that. I had one student on the job market this year who got a couple offers in the $80k-$90k range in various (not particularly highly ranked) biostat departments. It's only one data point, but I figure it's better than nothing. And that's in biostat. Salaries in stat departments tend to be lower (mainly because it is usually a 9-month salary rather than a 12-month salary). My student had a couple offers from stat departments around $70k at research-oriented universities and around $50k at a couple teaching-oriented universities.

 

As for industry jobs, as I said before, the job market is much tighter for PhD's than it is for MS-level people, which drives down salaries. Employers don't want to hire PhD's for MS-level positions because they assume that they will get bored and quit after a year or two. If you have a PhD and you are willing to move anywhere in the country and sell your services to the highest bidder, you can probably earn more than an MS graduate. But if you have a spouse and a child and a mortgage in a given city and you don't want to move anywhere else, earning $20 an hour as a contractor is a realistic possibility. I continue to stand by my advice that it's usually not a good idea to get a PhD unless you really love research. I don't know; if you know that you want to advance to the top of your field and you'd be bored as a SAS monkey, it's worth thinking about getting a PhD, I suppose. But it's still dangerous. Most students are single when they start grad school and think that it will never be a problem, but if you get married in grad school and have to find a job in the same city as your significant other, there is a good chance that the PhD will actually hurt both your job prospects and your earnings.