My area of expertise is Biostatistics; specifically, I am interested in the application of statistics to the biomedical field. My primary interest is collaborating with colleagues on campus and at the Indiana University School of Medicine. As a collaborating statistician, it is often difficult to distinguish my "research" from my "professional consulting" activities; they are all part of my professional development. Below are more of the professional activities I engage in as part of this development that I would not classify as my personal research area.

Indiana University School of Medicine in Terre Haute

For the past several years, I have been working with the Indiana University School of Medicine in Terre Haute. This campus hosts students pursuing an interest in rural medicine. As part of their curriculum, students participate in a community health project or conduct research of their own with the aim of a paper at a conference. These research projects are conducted under the direction of Robin Danek and Ellen Ireland; I serve as the statistical support on the projects. This gives the medical students a view of true collaboration as I am involved from the early part of the research and designing the survey through to analyzing and presenting the data. Below is a selected list of these projects, which nearly all have a focus on rural medicine.

InterVarsity Christian Fellowship

I spent the Fall of the 19-20 academic year consulting with InterVarsity. Reporting to the Director of Ministry Intelligence, I was tasked with identifying characteristics of chapters (and their host university) which are associated with higher likelihood of chapter closure. Using some proprietary data combined with publicly available data, I constructed several statistical models to identify these characteristics. My work resulted in a detailed report summarizing my conclusions and recommendations.

R Package: IntroAnalysis

I have taught Engineering Statistics I (an introductory course) nearly every term since joining the faculty at Rose-Hulman. Over time, I began to teach the course from a modeling-based perspective. My course notes eventually were compiled into an online text for students. While the text is not dependent on a particular software package, I use R exclusively in class. While there are some great packages for introducing R to non-statisticians, none (at the time) fully captured the modeling-based perspective while allowing a fluid transition between classical approaches and randomization-based approaches.

In response, I wrote the IntroAnalysis R package for my students. It is built on taking the following approach in an introductory course:

  1. Specify the model for the data generating process.
  2. Estimate the parameters for the model
    • Specify the conditions you are willing to assume about the error terms
    • Use an appropriate method for constructing a confidence interval based on the conditions (classical theory, residual bootstrap, parametric bootstrap, wild bootstrap)
  3. Compare models for hypothesis testing
    • Specify the model for the data generating process under the null hypothesis
    • Specify the conditions you are willing to assume about the error terms
    • Use an appropriate method for computing a p-value which compares the models based on the conditions
The package uses a formula-based notation (similar to the mosaic package) and has a similar framework as the infer package; it is somewhere between these two. It works well with the text and my teaching as the function names are consistent for all operations in the class.