11/13/2025: 10:00 AM - 11:30 AM EST
Panel
A single source of data can provide insight into a wide variety of interesting questions. Research teams often approach similar problems from different lenses and synthesize their findings to better understand the context of their work. This setup often flourishes when a senior investigator works with junior trainees, as each brings different expertise to the table.
In this panel, sponsored by the Caucus for Women in Statistics, we feature a diverse group of research teams ranging from undergraduate research assistants to tenured faculty. The speakers will discuss their collaborative work in health disparities, phylogenetic diversity, HIV prevention, and disability statistics. For each pair, we highlight how their projects align to answer a broader question. First, Dr. Sarah Lotspeich and mentees Cassandra Hung and Darcy Green will describe how measurement error can pose challenges in quantifying health disparities using a case study on diabetes and the food environment. Next, Dr. Taylor Krajewski and mentee Lydia Owens will apply distinct modeling strategies to understand predictors of pre-exposure prophylaxis. Then, Dr. Laura Kubatko and former mentee Dr. Kristina Wicke will present a fair proportion index to explore the range of outcomes induced by using gene-level and species-level phylogenies to rank species. Finally, Dr. Shiya Cao and mentee Lucia Qin will close the session by discussing the use of multilevel models and quantitative intersectional analysis to discover healthcare disparities and policy implications for inter-categorical groups.
The four speaker teams span a wide array of technical content and statistical applications, making the session valuable and interesting to a general audience. However, all four showcase how team science can build community, support collaboration, and allow both the mentors and the mentees to excel.
collaboration
mentorship
health disparities
biodiversity
public health
Organizer
Ashley Mullan, Vanderbilt University
Target Audience
Beginner
Tracks
Knowledge
Women in Statistics and Data Science 2025
Presentations
Disparities in healthy eating relate to disparities in well-being, leading to disproportionate rates of diseases like type-2 diabetes in communities that face more challenges in accessing nutritious food. Quantifying these disparities is key to developing targeted interventions, and there are limitations with the currently available methods and data that we are working to resolve. Namely, available data on disease rates are usually aggregate, which smooths over details about the individuals and communities within them. Further, aggregate disease data often comprise small area estimates, which carry additional uncertainty. In this project, we investigate the relationship between patients' food environment and the risk of diabetes using individual-level data from electronic health records at a large academic medical center. Using various health disparities methods, we quantify whether patients with worse access to healthy food face a higher burden of prevalent type-2 diabetes. Still, we face measurement error in food access since it is collected using inaccurate distance calculations. Finally, we discuss the impact of using error-prone food environment measures to detect health disparities in these data.
Speaker(s)
Sarah Lotspeich, Wake Forest University
Cassandra Hung, Wake Forest University
Darcy Green, University of Chicago
This presentation highlights a mentor-led, student-driven collaboration to understand predictors of PrEP knowledge using diverse machine learning approaches. Working from a shared public health dataset, each team member applied a distinct modeling strategy—including logistic regression, elastic net, random forests, and deep learning neural networks—to explore the research question from different analytical angles. The talk will describe how these parallel analyses offered complementary perspectives on the research question and reflect on how team-based exploration of modeling strategies supports skill development and applied learning in public health data science.
Speaker(s)
Taylor Krajewski, UNC - Chapel Hill
Lydia Owens, University of North Carolina at Chapel Hill
Phylogenetic diversity indices are often used as a means of ranking species for conservation priority. A crucial issue in this process is the selection of an appropriate phylogenetic tree for application of these indices. We use the Fair Proportion Index to explore the range of outcomes that result from using gene-level vs. species-level phylogenies. We show both theoretically and by application to empirical data that the choice of phylogeny plays a key role in ranking species. We discuss the broad implications of our work to other measures applied to fixed phylogenies.
Speaker(s)
Laura Kubatko, The Ohio State University
Kristina Wicke, New Jersey Institute of Technology
In this talk, we will discuss a collaborative project: Understanding the Sexual Orientation and Gender Intersection-Related Healthcare Disparities from a Social Context Perspective. The project was conducted in the mentor's Disability Inclusion Analytics Lab. The mentor has expertise in disability statistics using public population surveys and designed this project from a social context perspective using multilevel modeling. The mentee, an undergraduate junior majoring in Statistical and Data Sciences, had completed a second course in statistics (Multiple Regression) and a Data Science 2 course (Programming for Data Science in R) before joining the project. She brought her programming, statistical analysis, and visualization skills gained from classrooms and textbooks to the project. The mentor always encourages her students to be creative in problem-solving in her lab. Through weekly lab meetings, the mentor and mentee discussed what statistical analysis makes sense in this particular disability inclusion context, what scientific literature helps identify research gaps and form research questions. First, the mentee built and interpreted baseline logistic regression models with an interaction term of sexual orientation and gender using the National Health Interview Survey data. Second, the mentor introduced multilevel models and deepened the mentee's understanding of her multiple regression coursework. Then, the mentor focused on the quantitative intersectional analysis for logistic regression and complex survey design for easier and clearer interpretations of intersectional results and more actionable policy implications for inter-categorical groups.
Speaker(s)
Shiya Cao, Smith College
Lucia Qin, Smith College