My new website is at: https://bcheggeseth.github.io/. Please see that site for updated information.
Applied Collaborations
Ongoing work with scholars in biology, social psychology, computer science, and political science.
Longitudinal Regression Trees
Ongoing work with Anna Neufeld to review and compare of the existing methods to create a regression tree based on longitudinal outcome data and then improving the methodology to be able to detect additive and interactive relationships between baseline covariates and growth patterns. splinetree R package is available on CRAN to implement this work.
Clustering Longitudinal Data
Ongoing methodology research in cluster analysis of longitudinal data. Past projects include studying the impact of covariance misspecification in multivariate mixture models, proposing a preprocessing technique to cluster longitudinal data based on the shape of trajectories over time, and describing the consequences of the centering longitudinal data on covariance and the modeling implications. Additionally, we have compared model-based clustering algorithms with standard longitudinal methods in their ability estimate non-monotonic relationships between baseline exposures and non-linear growth patterns.
Childhood Developmental Trajectories and Environmental Exposures
Collaborated with researchers at the Center for Environmental Research and Children’s Health to detect associations between environmental exposures and their developmental patterns. I have worked with Kim Harley, Associate Director for Health Effects, and Vitaly Volberg, PhD, who was an environmental health sciences student at UC Berkeley, to apply clustering longitudinal data methodology to child body mass index growth curves as well as hormone trajectories.
Parity-related Weight Gain
Worked with Barbara Abrams, professor of Epidemiology, Maternal and Child Health, and Public Health Nutrition at UC Berkeley, to estimate the association between parity and weight gain over a woman’s lifetime.
World Health Organization
Worked with Dr. Claudia Stein, WHO Department of Food Safety, Zoonoses and Foodborne Dis- eases, on determining the global mortality burden of foodborne disease using vital registration data provided by WHO and creating a model to predict foodborne mortality according to ICD9 code for countries with missing data. Presented results to Foodborne Disease Burden Epidemiology Reference Group (FERG) in Geneva, Switzerland in November 2007.
St. Olaf Financial Aid Office
Collaborated with the Financial Aid Office in the analysis of current student data and used multiple imputation for missing data to estimate the student household income distribution.
Summer Institute for Training in Biostatistics
Explored the field of biostatistics at North Carolina State University, Raleigh, NC through lecture, computing, and a final project using data from the Duke Clinical Research Institute under the supervision of Marie Davidian and Dennis Boos.