Implementing Bayesian Estimation Under Complex Survey Sampling (Online)

This course will take place over two mornings (10/22/24 and 10/24/24), 4 hours per morning, and will be offered via Zoom. Attendance is required as the course will not be recorded.

Complex survey sampling techniques (e.g., clustering, stratification, oversampling) allow for cost efficient estimation for large, dispersed populations. As such, they are frequently used in demographic, health, and public opinion survey research settings. In recent years, Bayesian statistical methods, due to their flexibility and intuitive interpretation relative to frequentist methods, have become increasingly popular for analyzing complex survey sample data; however, complex survey sampling introduces certain features (e.g., unequal selection probabilities, dependencies between observations) that violate traditional statistical assumptions and can bias survey estimates.
This course provides a practical introduction to the csSampling R package, which addresses these issues by implementing Bayesian estimation under complex survey sampling. The course will begin with an introduction to Bayesian statistical methods and complex survey sampling as well as the differences between Bayesian and frequentist methods to account for complex survey design. The bulk of the course will focus on a guided tutorial of the csSampling package with sample data and R code. Finally, the instructors will present use cases of how they have used the package in their own research.
Learning Objectives for the course are:
● Define key concepts in complex sample survey data collection and analysis.
● Define key concepts in Bayesian statistical methods.
● Compare Bayesian and frequentist approaches to account for complex survey sampling designs.
● Understand the functionality of the csSampling R package.
● Describe prior applications of the csSampling R package to real-world data analysis problems.
● Implement sample R code to build Bayesian models analyzing complex sample survey data.
Prior experience using R to manipulate and analyze data will be assumed. Some familiarity with either Bayesian statistics and/or complex sample survey designs will also be helpful but is not required.

This course will count as 8.0 CSS short course credits.

Instructors: Hunter McGuire, Matt Williams, and Stephanie Wu

Hunter McGuire is a PhD candidate in Public Health Sciences at Washington University in St. Louis. His dissertation uses Bayesian multilevel modeling applied to complex sample survey data to estimate the population distribution and determinants of mental health concerns at the intersection of multiple axes of social identity and position (e.g., race/ethnicity, gender, sexual orientation, weight status). Hunter received a BA in Public Policy and a Master of Public Health from the University of North Carolina at Chapel Hill.

Matt Williams is a senior research statistician with RTI where he works with federal clients on topics such as privacy, survey design and analysis, and complex Bayesian models. Prior to RTI, Matt worked for over a decade as a statistician at several federal statistical agencies. Matt has also consulted on international development projects related to agriculture and vaccination. Matt received his PhD in statistics from Virginia Tech and a dual BS in biology and mathematics from Clarkson University. Matt developed the initial csSampling R package with the goal of bridging complex survey designs with Bayesian modeling approaches.

Stephanie Wu is a PhD candidate in Biostatistics at the Harvard T.H. Chan School of Public Health. Her thesis work, co-advised by Professor Briana Stephenson and Professor Michael Hughes, focuses on Bayesian model-based clustering methods for survey data with applications to nutritional epidemiology and health disparities research. Stephanie received her bachelor’s in Statistics and Public Health-Global Health from the University of Washington in 2019 and has also worked on projects relating to planetary health and HIV drug resistance surveillance in low- and middle-income countries.

Registration fees:

• UNC Chapel Hill Students (CSS Students will earn CSS short course credit): $0, with a $35 deposit to hold your spot (deposit is refundable upon your attendance for at least 66% of the course)
• UNC Chapel Hill Faculty/Staff/Postdoc/Resident/Visiting Scholars: $80
• University (Non UNC Chapel Hill) Student/Employee (must have active university email): $105
• Government/Non-Profit/Corporate: $130

Additional course information:

<ul

• Cancellation/ Refund Policy: A full refund will be given to those who cancel their registration no later than 10 days prior to the course. If you cancel within the 10 days prior to the class, no refund will be given. Please allow 30 days to receive your refund.
• Zoom link for this course will be sent prior to the course. Registration must be made prior to the registration cutoff date to receive the Zoom link.

For questions regarding this class, please contact Jill Stevens at jill_stevens@unc.edu