Below is a summary of ongoing areas of research. I have intentionally tried to make this page accessible. If you have questions about technical details of any of this work, please feel free to to send those questions my way.
Spatial Regression Models
I work a lot with spatial regression models, particularly when the inferential focus is on associations between a response and a set of exposures/covariates. My work typically focuses on models that incorporate spatial dependence with Gaussian Markov Random Fields. My current projects involve identifying when/if a spatial regression model can improve inference over a non-spatial model. The most common application areas for the models I use include disease mapping, epidemiology, and health sciences.
Data Science & Criminal Law
I often found myself surprised by how seldom the legal system relies on what the average statistician would consider empirically driven decisions. As a member of the legal community, though, it becomes clear how many barriers there are to collecting data in a way a statistician would want to. I am currently working on projects related to identifying and overcoming such barriers in the American criminal justice system. Relying on work in experimental design and the analysis of missing data, I am also exploring the impact these barriers have on our current understanding of the American criminal justice system.
In my research, I have come across problems related to the reproducibility of published research results. While the inability to reproduce research results is inherently problematic, it is particularly troublesome when the research results are being used to inform public policy decisions or convict a criminal defendant. Working primarily in public health and criminal justice settings, I look at two areas related to reproducibility. The first is computational reproducibility: if I had your code and your data, would I get the same results as you did? The second involves a question of whether the computational techniques used in published research match the description of the mathematical/statistical approach they are supposed to carry out.
I am interested in how statistical methods are (or are not) used to analyze forensic evidence in criminal trials. I work at the intersection of law and science on projects related to the admissibility of various kinds of forensic evidence in criminal trials, the methods that should be used analyze this evidence, the methods that are used to an analyze this evidence, and what tools defense attorneys need in order to adequately explore the limitations of these methods at trial.