ASA Connect

 View Only
Back to eGroups
Expand all | Collapse all

Is it needed?

  • 1.  Is it needed?

    Posted 01-24-2016 03:58

    Dear Listmates,

    How far can one expect to go in a stats career without a stats degree? Does independent study count for anything?

    Sincerely,

    KeithC.

    ------------------------------
    Keith Chamberlain
    Analytical Chemist
    ------------------------------


  • 2.  RE: Is it needed?

    Posted 01-25-2016 03:35

    There are plenty of ways you can get into stats without a "stats" degree. For example, an economist, computer scientist, epidemiologist or industrial engineer take several stats classes with other coursework. They can usually get a job as a "statistician". 

    I started as a chemist many years ago. In my second class in stats, I learned that everything I was taught about stats from chemists was either 180 degrees wrong or lacking so much important material, that it would be better if I never took A-Chem in the first place. Coming back to chemometrics after 15 stats classes, I find most of the chemometricist have little to no clue about stats. (They tend to prefer 10 pounds of cure to a gram of prevention.) 

    I know what I just wrote sounds harsh. It's just that if the authors I've read took the related stats classes, they'd be failing the courses because of their poor understanding of stat methods. Their arguments are based upon, "Well, another chemist did it this way. Why can't I?" The answer is, sometimes the wrong way and the right way can lead you in the correct direction. The right way has a much better chance of leading you there. 

    With all the nuance needed and used in statistical analysis, self study with a good dose of expert opinion wouldn't be too bad... Assuming you use good material to study. Unfortunately, a lot of the good material is drier than glassware in a 350C oven. (It even hurts my head reading it;-) 

    What are you using for your study? 

    ------------------------------
    Andrew Ekstrom

    Original Message


  • 3.  RE: Is it needed?

    Posted 01-25-2016 08:30

    I am currently a government employee in agricultural statistics.  My experience is that you can certainly do the job well and have a successful career without a stats degree, but only if you are planning to stay within a single organization.  As soon as you start looking at opportunities to advance in other organizations, the lack of credentials will be a small but noticeable problem.   

    To your current employer, your subject matter expertise and proven statistical track record should fully cover any degree you may not have.  To a prospective employer, subject matter expertise and track record are much less clear.    

    ------------------------------
    William Morris
    Statistician
    USDA/National Agricultural Statistics Service

    Original Message


  • 4.  RE: Is it needed?

    Posted 01-26-2016 08:12
    Neither John Tukey nor R. A. Fisher had a stats degree.
    I don't believe that Wm. Cochran or Erich Lehmann did either.
    They did fine.

    H

    Howard Wainer
    Distinguished Research Scientist
    National Board of Medical Examiners
    3750 Market Street
    Philadelphia, PA 19104
    Voice - 215-590-9861
    email - hwainer@nbme.org

    This email message and any attachments may contain privileged and/or confidential business information and are for the sole use of the intended recipient(s). Any unauthorized review, use, disclosure or distribution is prohibited. If you are not the intended recipient, please notify the sender immediately by reply email and destroy all copies of the original message and any attachments.


    Original Message


  • 5.  RE: Is it needed?

    Posted 01-26-2016 09:09

    Interesting comment, but all the giants Dr. Wainer mentions invented our field of statistics.  In other words, someone has to build the house before it can be lived in. 

    ------------------------------
    Morris Olitsky
    Statistician
    USDA

    Original Message


  • 6.  RE: Is it needed?

    Posted 01-25-2016 10:39

    A degree in mathematics or computer science might be acceptable to some employers for an applied statistics position, as long as the person had taken statistics courses or attended workshops.  Not sure if a chemistry degree, instead of statistics degree, would be acceptable, but it might be for some employers.

    ------------------------------
    Brandy Sinco
    Research Associate

    Original Message


  • 7.  RE: Is it needed?

    Posted 01-25-2016 10:43

    How far can you go in a stats career without a stats degree?  It depends upon what kind of stats career you are considering.  If you want a tenured position teaching statistics in a statistics department at a University, you need a stats degree.  If, however, you are looking at an industrial position, results matter - how well can you apply the tools in a practical manner?  I have degrees in both engineering and statistics, and that makes it easier for me to help other engineers in solving their issues.  I have seen some engineers who have learned statistics without a formal degree do quite well.  I have seen people with statistics degrees, but without the practical knowledge, do poorly.

    ------------------------------
    Ellen Barnes
    Sr. Technical Specialist
    Ford Motor Company

    Original Message


  • 8.  RE: Is it needed?

    Posted 01-26-2016 08:30
    How far can one go in a stats career without a stats degree?


    This is a fascinating question.

    I have worked in the pharmaceutical industry for 29 years.   During that time, I have worked with approximately 500 statisticians.

    My unscientific analysis proceeds:


    30% of my colleagues have held a Ph.D. in statistics
    67% of my colleagues have held a Masters in statistics
    3% of my colleagues have not held a degree in statistics.

    Now it gets interesting....

    of the 3% who did not hold a degree in statistics, all but one of them had a background that included courses in statistics, probability, and/or mathematics.

    So, it is possible to have NO degree in statistics and have a career in pharma.

    Here is the issue I have identified among those who do not have a degree in statistics.  Almost all of these folks will encounter a situation where their gap in education limits them to providing sound statistical advice.   If these folks have a "mentor" to approach in these situations, they will be able to use consultation to fill the gaps.  Once the mentor is removed or is unavailable (and I have seen that happen)  the situation exposes the incumbent and he or she is left twisting in the wind, or worse, out of a job.

    For those who would venture into this career choice without a degree in statistics, I would strongly suggest that he or she enrolls in a program in advanced statistics (I recommend a Ph.D.) immediately upon starting in the profession.  Many schools offer programs tailored to those in the workplace, with courses offered in the evening.    

    I am available to discuss further, for any who would like to talk about this.  Look me up on Linkedin.

    Jeff Davidson
    Endo Pharmaceuticals




    Original Message


  • 9.  RE: Is it needed?

    Posted 01-25-2016 11:43

    While I would consider independent study to be an important, even essential addition to a college degree, I also have to consider analytic chemistry to be an analytic degree. For example, my own PhD work is in statistical astrophysics. The degree says Physics / Astrophysics on the label; I also have Master's in Physics and Bachelor's Degrees in Physics and Mathematics. 

    While universities require a degree with the right words on it - for example, I only teach adjunct classes in astronomy - my impression is that the private sector isn't overly concerned about the label on your degree but rather the statistical content of your work, especially since you left college. 

    ------------------------------
    David Corliss
    Analytics Architecture
    Ford Motor Company

    Original Message


  • 10.  RE: Is it needed?

    Posted 01-25-2016 17:21

    All the responses provide good advice but I would suggest narrowing your question somewhat. As you probably know, "statistics" is pretty broad and people that know something about one area of application may know little or nothing about others. Obviously you need basic knowledge of statistics (and probability theory) no matter where you want to apply it. since you have a background in chemistry and are aware of chemometrics this may be a good example. There now is a sizable group of statisticians who have an interest in this but not too long ago it was relatively unknown in the statistics community and the statistics literature. If you want to go into or move into biostatistics you probably need a degree with that in the title of the degree. As several have noted, if you want to go into academia (but depending on the department) you likely need a degree in statistics or at least in mathematics. But the bottom line might be having publications in appropriate journals, establishing connections with known statisticians, i.e. having evidence of your statistical knowledge/competence that is transportable from one employer to another. If you look at history there are a number of very well known "statisticians" (current or past) who changed fields and did not go back to get new degrees. It all needs to be put in context, your current career situation, your age, your self discipline, what you want to do and become (maybe even your family circumstances. i.e. are you married and have children, your financial circumstances and whether they would allow you to follow various options.Statisticians (taken in a broad sense) tend to collaborate with people working in other disciplines and this might open doors for you.

    ------------------------------
    Donald Myers
    Emeritus Professor of Mathematics
    Univ of Arizona

    Original Message


  • 11.  RE: Is it needed?

    Posted 01-25-2016 19:24

    How far can one go in a stats career without a stats degree? Potentially pretty far. Back in the late '90s, half the faculty in LSU's Department of Experimental Statistics had their PhDs in other disciplines. This included the department chairman, the late Dr. Barry Moser, whose PhD was in Forestry. Dr. Geaghan's PhD was in Fisheries and Wildlife Management, and Dr. Blouin's PhD was in Agronomy, I believe. All were effective, inspiring instructors as well as competent and productive statisticians.

    It seem, though, that there is a lot of specialized statistics that fit analytical chemistry like a glove. Do you have a lot of statistical expertise in that area?

    ------------------------------
    Eric Siegel, MS
    Research Associate
    Department of Biostatistics
    Univ. Arkansas Medical Sciences

    Original Message


  • 12.  RE: Is it needed?

    Posted 01-25-2016 23:24

    Something else that you will find is that different areas of a corporation will need different skill sets.

    If you are working in a research/production facility, Design of Experiments, Adv Design of Experiments and QC will be vital.(I like Douglas Montgomery's treatment of these topics.) In a clinical trials facility, DOE and QC are almost unheard of. Methods in Mixed Models ,Survival Analysis and Logistic Regression will be a lot more important.

    What area are you looking to get into?  

    ------------------------------
    Andrew Ekstrom

    Original Message


  • 13.  RE: Is it needed?

    Posted 01-26-2016 09:11

    Is a PhD in statistics needed? Most decidedly No. Much more important is an advanced degree and/or knowledge  in the field to which you wish to apply statistics. Either take a postdoctoral in mathematical statistics, or read it up yourself - I did both, and even now, in retirement, try to follow as much new theoretical statistics as I can.

    ------------------------------
    Robert Elston
    Emeritus Distinguished Professor, Case Western Reserve University

    Original Message


  • 14.  RE: Is it needed?

    Posted 01-27-2016 08:17

    I'm an economist by training and work in statistics.  This is very common.  I studied lots of math and statistics throughout my postsecondary education and passed a field in Econometrics.  I also had job experience in statistics before I got my Ph.D. My current job title is Mathematical Statistician.

    ------------------------------
    Chuck Coleman

    Original Message


  • 15.  RE: Is it needed?

    Posted 01-26-2016 10:05

    A degree in statistics can't hurt.  That said, I worked in a process improvement role for a number of years before I earned my MS in statistics and did just fine.  If I needed additional help, there was a statistics group available.  As for the suggestions to get a PhD, do you really want to spend 4 years doing research in statistics basically doing theoretical mathematics (You'll spend your days proving theorems) or are you interested in applying and practicing statistics?  If it is the latter, an MS in applied statistics can be earned on line from places like Penn State U, RIT, Texas A&M, and Colorado State U.

    ------------------------------
    Brian Cocolicchio


    Original Message


  • 16.  RE: Is it needed?

    Posted 01-26-2016 19:18

    Historically speaking, some of the big names in Statistics started out as chemists.

    Australian-born W.J. Youden (1900-1971) obtained his PhD in Chemistry from Columbia University in 1924. After reading Fisher's classic book Statistical Methods for Research Workers and attending Hotelling's lectures at Columbia, Youden made the move to Stat, resulting in Youden Squares (in a 1937 paper, named such by Fisher and Yates 20 years later).

    Perhaps not surprisingly, Youden's first book, in 1951, was Statistical Methods for Chemists.

    Youden was given the sixth annual Wilks Memorial Award by the ASA in 1969. The W.J. Youden Award in Interlaboratory Testing was established in 1985.

    And of course there is John Tukey (1915-2000), who received BA and MS degrees in Chem from Brown University. Tukey was given the second annual Wilks Memorial Award in 1965, and received the National Medal of Science from President Nixon in 1973.

    ------------------------------
    David Bernklau
    (David Bee on Internet)

    Original Message


  • 17.  RE: Is it needed?

    Posted 01-27-2016 13:13
    • It  seems to me that many jobs announcements I see require a masters or PhDs in statistics.  The  I was trained as a biologist, worked 30 years in Fisheries. I had basic introductory courses in applied statistics.  The agency eventually needed biologists with quantitative skills, so I focused on obtaining a set of statistical tools. My job transitioned to that of a data analyst/statistician.  I spent a lot of time with my nose in the statistical literature.  Having come in through the hawse pipe, I always felt that I was hamstrung somewhat by the lack of formal statistical training.  I could likely find employment within the fishery field, but suspect I would have had  difficulty finding work as a statistician in public health or pharmaceutical industry without credentials.  I am now retired, but I had a nice career.  The transition to statistical work opened a lot doors for me and, more importantly the work was immensely satisfying.
    ------------------------------
    Steven Correia
    Environmental analyst (retired)
    Massachusetts Division of Marine Fisheries

    Original Message


  • 18.  Fisher, Tukey, Cochran and Lehmann

    Posted 01-27-2016 08:14

    It is true that the first three did not have degrees in Statistics, and possibly technically true of Lehmann, whose PhD thesis was written under the supervision of Jerzy Neyman. I think at that time Statistics was not a separate department at Cal, so perhaps his PhD was counted as a math rather than a statistics degree.

    ------------------------------
    Giles Warrack
    North Carolina A&T State University
    ------------------------------


  • 19.  RE: Is it needed?

    Posted 01-28-2016 04:35

    This is a good question.  In my sense, this is one of the reasons why the PStat Accreditation was created.  There have been many discussions about why you would need a PStat with a PhD in Stat/Biostat.  To me, the answer is you don't - the reason why the PStat program is useful is exactly for people who do not have a degree in statistics, but have some experience working with statistics and want their experience to be recognized.  If you do not have a statistics degree, I encourage you to look into accreditation.  For more details, see Accreditation

    Amstat remove preview
    Accreditation
    Welcome to the ASA accreditation web page. The ASA now offers two levels of accreditation: Accredited Professional Statistician (PStat ®) and Graduate Statistician (GStat). GStat is an entry level of accreditation, preparatory for full PStat ® accreditation. To apply for either level accreditation, log in to ASA members only, and click on the PStat/GStat button to begin the process.
    View this on Amstat >

    .

    Michael

    ------------------------------
    Michael Posner
    Associate Professor of Statistics
    Director, Center for Statistics Education
    Villanova University

    Original Message


  • 20.  RE: Is it needed?

    Posted 01-29-2016 01:05

    Dear Listmates,

    Wow, a lot of really good feedback so far. Any apprehensions I had about posting my initial question are long gone. 

    Where to begin. It is interesting to learn about the giant's such as Tukey's & Fischer's careers and impact, but for me I think it would be a bit grandiose to think I can share the company. I do want to make a contribution though, so I have no problem standing on their shoulders.

    Many folks seem to claim that while it's doable without the advanced degree (in industry) it would sure make life easier to have one provided life doesn't make any prohibitions. There's also a good percentage of folks that suggest I get more specific with what I want to do and what in studying. I will do the here. 

    My exposure to stats comes from self study in time series analysis and some grad work I did as an undergrad in statistical programming and glm. On the job in exposed to calibration models and have furthered that work on my own a bit to include multivariate calibration. I had a good dose of design of experiments and the academic side of research design/writing/presenting from the McNair scholars program.

    I've always been interested in academia, but I'm concerned that a sleep disorder will impair my ability to partake on time consistently. It certainly did as an undergrad where life is more forgiving. Industry gives me the flex time I need. 

    Lately I've noticed a need in the supplement and neutriceutical industry for good clinical trials and have been interested in filling that gap. I'm aware of a company that had a clinical trial with just 12 participants total in 4 arms and their selling their findings to the FDA (for Gras submission) and consumers and investors like they found something revolutionary. I think it's unethical use of statistics and wonder how they got sold on the design (and why is being promoted so much). So the supplement industry has some hard knocks coming regarding clinical trials and I'd like to be there for it. 

    Finally, regarding training, I definitely want to apply stats over proving theorems. The MS in applied stats sounds quite interesting. I'll have to look into that.

    Thanks again for all of your wonderful feedback. 

    Sincerely, 

    KeithC. 

    ------------------------------
    Keith Chamberlain
    Analytical Chemist

    Original Message


  • 21.  RE: Is it needed?

    Posted 02-03-2016 13:08

    If you are interested in applied statistics with an industrial focus, I'd look into an MS in industrial Engineering. You can take 50%-75% of your courses in applied stats. 

    If you look for an MS program in applied stats, you might have to look for an on line program like those at Penn State or Rochester Institute of Technology(I think). A lot of the applied stats programs you take in person assume you are a full time student and can take day classes. (At least that is the case by me.) Meanwhile, many more IE programs offer part time, online and/or evening classes. 

    Don't forget about George Box. He started as a chemist. Worked as a toxicologists during WW2. Got mad at the poor methods for analysis he knew about and sought help from Fisher. After WW2, he studied stats. Then applied stats to manufacturing and chemical processes. 

    ------------------------------
    Andrew Ekstrom

    Original Message


  • 22.  RE: Is it needed?

    Posted 01-29-2016 16:02

    As part of history of statistics, most of the founding fathers of both applied and theoretical statistics, did not possess any degrees in statistics (there weren't any).  For example:

    1. RA Fisher- Mathematics and biology

    2  Karl (Carl) Pearson- Mathematics and physics

     3. PC Mahalanobis- Physics

    4. Oscar Kempthorn- Genetics and philosophy of science

    5. RC Bose- Mathematics

    6. SN Roy- Mathematics and theory of relativity

    7. John Tukey- Chemistry and mathematics

    8, WG Cochran- Mathematics

    9. George Snedecor- Mathematics, physics, and Greek

    All these people had solid background in calculus, matrix theory, etc.

    They all taught themselves probability and mathematical statistics.  Their backgrounds in physics, chemistry, genetics, and other biological fields drove them into research and development of applied and bio- statistics (not "SAS Statistics!".

    Ajit K. Thakur, Ph.D.

    Retired Statistician

    ------------------------------
    Ajit Thakur
    Associate Director

    Original Message