Skip to main content
Explore URMC

SMD Logo

menu
URMC / Education / Graduate Education / URBest Blog / April 2017 / Get Prepared To Become A Successful Statistician in a Collaborative Research Environment

Get Prepared To Become A Successful Statistician in a Collaborative Research Environment

Career Story by Aiyi Liu, PhD, Senior Investigator, Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD), Biostatistics & Bioinformatics Branch

Rochester Data Science Society

Statistics is perhaps one of the few professions that has seen steady job growth in the past 30 years or so, and the need for statisticians continues to grow today. Due to the applied nature of statistics, graduate students often find themselves landing a job in a highly collaborative environment (e.g., medicine, public health) that requires not only good training in statistics, but also a fair understanding of subject matter and, perhaps more importantly, skills needed to be able to collaborate as a team member with non-statisticians. Most likely these important skills are not taught in classrooms and could potentially hinder the career growth of a statistician. 

I had a Bachelor of Science degree in mathematics and a Master of Science degree in mathematical statistics from the Department of Mathematics at the University of Science and Technology of China, one of the most prestigious universities in China.  At that time, I saw no apparent distinction between mathematics and statistics, viewing the latter as just one branch of the former. 

Only after I became a graduate student in the Department of Statistics at the University of Rochester in 1993, did my view and understanding of statistics change.  I had the good fortune to work on my dissertation under the supervision of the late Professor W. Jack Hall, who was then the principal statistician on the Multicenter Automatic Defibrillator Implantation Trial (MADIT) to study whether prophylactic therapy with an implanted cardioverter-defibrillator, as compared with conventional medical therapy, would improve survival in high-risk patients. Many statistical problems arose from that trial, some were very mathematically challenging.  With such a clear connection between statistics and real-world applications, I felt very motivated and energized, completing my doctoral dissertation and addressing and answering a number of problems. The unsolved problems later became my research focus for many years. 

Over the years many PhD students have asked me for advice about choosing an advisor and a thesis topic, and I often share with them my experiences from the University of Rochester. A few things are worth considering, including the extent of self-interest in the topic, self-evaluation of the ability to solve the problems (What skills that might be needed to solve the problems? Are the problems too easy or too difficult to solve?), the research portfolio and personality of the professor (Has the professor been an active researcher in the area?  Is the professor a role model for you to follow both as a researcher and a future colleague?) and the continuity of research stemming from the thesis (Will this work extrapolate for at least a few years after graduation). 

As a statistician working in a collaborative environment in public health and medicine, I consider a “successful biostatistician” as one who is enthusiastic and able to collaborate with biomedical investigators of various backgrounds to advance the sciences in medicine and public health, and who takes pride and is productive in developing more efficient, new innovative statistical methods to address issues uncovered during the collaboration. 

The job of a biostatistician in biomedical research is not just data analysis. The job involves all aspects of the study: study design, implementation, interim monitoring of data quality and study compliance, data analysis and interpretation, and manuscript preparation. All these study components require a combination of many skills, besides a profound understanding of various statistical topics. These requirements include (but are not limited to) communication skills, a fair knowledge of the subject matters (e.g. breast cancer, genetics, human reproductivity), a good personality, people skills, and the ability to bridge statistical and biomedical topics. While many of these skills will not be learned in the classroom, some programs do provide courses/workshops on statistical consulting, building collaborations or communicating effectively. I strongly encourage graduate students to explore as many opportunities (as time will allow) that will help them to become better scientists and better collaborators. Please join me at the Rochester Data Science Society's Inaugural Alumni Seminar for more discussion April 28 from 11 am – noon in the Center for Experiential Learning in room 2-7520. Refreshments will be provided.

 

Global Administrator | 4/26/2017

You may also like

No related posts found.