Robyn C. Wallace is a student in the MS in Government Analytics program with a concentration in statistical analysis.
I am a scientific data analyst with Northrop Grumman, a major US defense and technology firm. Through a contract with the company, I work at the Centers for Disease Control and Prevention (CDC) in Atlanta where I work on a variety of programs within the Division of HIV/AIDS Prevention (DHAP) and the National Center for Chronic Disease Prevention and Health Promotion (NCCDPHP). As an analyst, having the skills to thoughtfully design experiments, analyze large amounts of data, and interpret and leverage data are key factors to successfully solve problems and optimize value within public health programs. The MS in Government Analytics program is expertly designed to expand my proficiency in the latest analytics technologies, applications, and practices that are actively reshaping the public sector.
This skills learned in this program are directly applicable to my work as an analyst within CDC’s Population Health Division, Behavioral Risk Factor Branch. Working in this role, while pursuing this degree, has afforded me the platform to apply fundamental techniques learned to a practical setting within the Public Health sector. A large component of my position involves developing data standards and methodologies for Population Health Surveillance data in tandem with designing and coding statistical models to determine chronic health status at the state and county level. Analyzing health status at the county level is particularly important since state and local governments require sub-state geographically based information in health policy planning and program implementation. However, most public health data, including the Behavioral Risk Factor Surveillance System (BRFSS), are collected nationally and not designed to produce direct estimates for chronic health conditions at the county level as the sample sizes are too small. Hence, the estimates at the county level are not reliable or stable. To mitigate this, the Behavior Risk Factor Branch has designed a methodology, primarily using SAS, to produce accurate chronic health county estimates when aggregated at the state level. As the lead analyst on this project, I’ve worked to develop BRFSS’ small area estimation method and have recently published a paper in Preventing Chronic Disease on this research.
With the knowledge gained in the MS program, I plan to expand current modeling techniques within the Behavioral Risk Factor Branch that will offer additional information about national health data outcomes and introduce new software, such as R and Stata, into BRFSS’ functionality. Owing to the skills learned in courses such as Advanced Quantitative Methods, my goal is to streamline the current small estimation method by utilizing multiple imputation to address survey data incompleteness. I also plan to explore bootstrapping as a method to address underestimated standard errors, a major concern in the BRFSS small area estimation method resulting from the model not taking into account complex design variables. Exploring bootstrapping and other methods will set the stage for future papers on small area estimation for the Behavioral Risk Factor Surveillance System.