Research Interests
Likelihood ratio and receiver operating characteristic (ROC) models in forensics
My research in this area is to strengthen error rate assessment and evidence interpretation by utilizing novel and rigorous statistical methodology. By investigating the covariate effects on the forensic accuracy and relating the ROC curve to the likelihood ratio, the goals are to enhance understanding of accuracy and its inherent variability in forensic science disciplines, and provide guidance to rigorous interpretation of these measures.
Group sequential designs and adaptive sampling procedures in comparative diagnostic trials
My research in this area is to develop novel design methods to estimate and compare common diagnostic accuracy measures, including ROC curve, precision curve, and F-measure curve.
Analysis of correlated ROC curves, location ROC curves, and free-response ROC curve
My research in this area concerns the development of new statistical methods for estimation and hypothesis testing in correlated ROC curves.
Statistical methods in substance abuse treatment
My research in this area involves collaborations with investigators in the areas of criminology and substance abuse treatment. I have worked on statistical methods to synthesize systematic reviews with different types of outcomes. In addition, I have developed models for estimating the trajectories of certain crime activities in probationers.
Clinical trial design in personalized medicine
My research in this area is on the development of new clinical trial design methods to validate markers for personalized medicine.