Congratulations to Xisi Zhang on his recent publication in Vaccine.
The article titled "Understanding the Shift in COVID-19 Vaccine Hesitancy and Its Associated Factors for Primary and Booster Doses Among University Students: A Cross-Sectional Study" represents an interdisciplinary collaboration among Bara' Abdallah Al Shurman, Shannon E. Majowicz, and Zahid A. Butt from the School of Public Health Sciences, along with Kelly Grindrod from the School of Pharmacy.
This study explores COVID-19 vaccine hesitancy (VH) among university students, a population subject to unique vaccination policies. While many Canadian universities mandate the primary COVID-19 vaccine series, booster doses remain optional. Within this distinctive policy context, the study aims to estimate the prevalence of VH for both the primary series and booster doses among students at the University of Waterloo. It also seeks to identify factors associated with VH and examine how hesitancy shifts between the initial and subsequent doses.
Given the large dataset and numerous variables, traditional variable selection methods proved impractical. To address this, Xisi proposed the use of the Quasi-likelihood under the Independence Model Criterion (QIC) to evaluate model fit and effectively manage the analytical challenges posed by the dataset’s complexity.
Through this project, I have gained hands-on experience in processing survey data, selecting the outcome of interest, and statistical modeling and analysis. I particularly appreciate the idea that developing methods should be grounded in practical applications. This project is not only a great opportunity to collaborate with and learn from senior statisticians and domain experts, but also a valuable chance to apply my knowledge and skills to real-world scientific problems.
Xisi Zhang
Xisi Zhang is a second-year Ph.D. student at the Department of Statistics and Actuarial Science and is currently one of the graduate consultants at SCSRU. Xisi’s primary research includes developing and applying statistical methods for causal inference with time-varying exposures, time-varying confounders, and missing data.