Being a Hybrid Statistician
This column is written for statisticians with master’s degrees and highlights areas of employment that will benefit statisticians at the master’s level. Comments and suggestions should be sent to Megan Murphy, Amstat News managing editor, at email@example.com.
Melissa Danielson earned her BA at the University of Virginia and her MS in public health in biostatistics at Emory University’s Rollins School of Public Health. Her work focuses on the epidemiology of ADHD and other mental, emotional, and behavioral conditions among children; the epidemiology of health outcomes among children with disabilities; and the evaluation of the Legacy for Children group-based parenting intervention.
Emily O’Malley Olsen has worked as a biostatistician in the academic, pharmaceutical, and public sectors. Her work focuses on surveillance of adolescent health risk behaviors and analysis of complex survey data. Olsen earned her BS in mathematics from Creighton University and her MS in public health from the Rollins School of Public Health at Emory University.
One type of job available to applied statistics graduates is that of “hybrid statistician,” in which statistical expertise is tied into a more general research-based role. Two master’s-level biostatisticians discuss their hybrid statistician jobs with the Centers for Disease Control and Prevention (CDC) in Atlanta, Georgia. Both Emily Olsen and Melissa Danielson have positions at CDC that allow them to serve as the statistical voice for their projects and also initiate and/or lead their own public health research. Below, they describe their jobs, what they like about being a hybrid statistician, and courses they completed to prepare them for this career.
Emily Olsen, Division of Adolescent and School Health
Around 50% of my time is devoted to what my supervisor and I have termed “office hours.” This includes reviewing sampling and weighting methods for our surveillance systems, writing and documenting statistical and data processes for our scientific publications and surveillance systems, and answering statistical questions from the public and my colleagues. Many of the inquiries from my colleagues evolve into authorship contributions to their manuscripts, and I really enjoy these opportunities.
The other 50% of my time is dedicated to publishing research articles using surveillance data. Since I work with large, established surveillance systems that cover a variety of risk behaviors among adolescents, I am able to focus on health topics that most interest me. I can be the first author and work with a subject matter expert (SME) who knows the health topic extensively. For example, I recently co-wrote an article about texting while driving with an expert on teen driving and transportation injury. Alternatively, I can run the analysis and take second or third authorship in collaboration with other SMEs.
This job is a great fit for me. I like the balance of consulting on a few projects here and there. I also really enjoy the freedom to explore new content areas that the surveillance systems cover such as bullying, nutrition, or pregnancy prevention and working with a variety of SMEs and colleagues. Finally, I like that I can keep coming back to one content area and develop my own knowledge, so that one day I also can be a SME on that topic and on statistics.
I earned my master’s in biostatistics through a school of public health. The required basic public health courses, in addition to my statistics courses, were helpful in preparing me for my career. I took introductory courses in epidemiology, health policy, environmental health, and behavioral science. I use what I’ve learned from these courses often in my statistics career. I also use methods I learned in every single statistical modeling class I took, and wish I had taken more! Fortunately, my employer often offers courses that advance the breadth of my statistical knowledge and skills. Finally, for those looking for a statistics position in public health, one course that really helped me was a course in complex survey data and methods.
Melissa Danielson, Division of Human Development and Disability
My job can generally be described as providing statistical support for a team of scientists working on issues related to child development. Our main sources of data are several longitudinal cohort studies and national health surveys, both of which provide an interesting variety of opportunities to apply different statistical techniques. For our work with health surveys, I usually team up with at least one subject matter expert to define the research question and devise an analytic plan to answer the question. I am responsible for identifying the best statistical approach to answer our question of interest, developing and running analytic programs, and producing output that can be used to disseminate our answer, whether it is in tabular or graphical form.
We often spend a lot of effort deciding the best way to display our results, because those choices significantly affect how our audience receives our message. While this component of statistical analysis isn’t always emphasized, a well thought-out analysis is enhanced by having the appropriate graphs and figures to highlight the key results.
The other area I spend the majority of my work time on is managing and analyzing data from two longitudinal data sets. These cohorts have been recruited and maintained under my team’s supervision, and, as a result, my responsibilities include data management and quality control tasks in addition to analysis of our primary outcomes of interest.
While data management may not be the most exciting part of my job, I really appreciate the time I spend immersed in the data. This allows me to be confident in the quality of the data and also to begin to understand the quirks and characteristics of the data that will improve my ability to analyze them appropriately.
One thing I particularly like about my job is that, since I am working in a specifically defined field, I have the opportunity to develop my knowledge base about the subject area, which can be an important way to improve my analytic capabilities. While I don’t have as much knowledge as the SMEs on my team (at least not yet), I do have a good general sense of the relevant literature and the field in general, and this allows me to be thoughtful and consistent about the hypotheses we test and the use of appropriate analytic techniques. Also, the more I know about the subject matter, the more of a leadership role I can take in the development of our research questions, particularly with consideration for the strengths and limitations of the data we have at hand. Plus, I enjoy learning new things, and I really like being able to expand my knowledge base in specific content areas as well as within the statistical field.
Like Emily, I earned my master’s degree in biostatistics from a school of public health and took a number of public health–related courses in addition to statistics courses. I’ve been able to use a great number of statistical techniques that I initially learned during my course of study, from linear and logistic regression, survival analysis, and mixed modeling of longitudinal data to analysis of complex survey data.
One skill I think was important to hone while in school was the ability to independently learn specific statistical techniques, because when using real world data, you find that sometimes the basics you learned in school are not going to be sufficient to address the questions you would like to answer. It helps to practice figuring things out on your own while you still have professors and other students around to help you confirm you are on the right track.