This workshop covers the fundamentals of statistical analysis using complex survey data, blending theoretical foundations with practical examples. It emphasizes the importance of a solid inferential framework when making inferences about a conceptual or finite population, focusing on model-based and design-based approaches. Choosing the appropriate framework may not be straightforward, and this workshop will clarify the contexts in which each approach is applicable. Additionally, the workshop will demonstrate how to adapt classical data analysis methods to survey data using the design-based approach. Key topics include the estimation of regression model parameters, means and proportions, variance estimation, estimation for domains of interest, confidence intervals, and comparisons between groups.

The use of health care claims datasets often encounters criticism due to the pervasive issues of omitted variables and inaccuracies or mis-measurements in available confounders. Ultimately, the treatment effects estimated utilizing such data sources may be subject to residual confounding. Digital electronic administrative records routinely collect a large volume of health-related information; and many of which are usually not considered in conventional pharmacoepidemiological studies. A high-dimensional propensity score (hdPS) algorithm was proposed that utilizes such information as surrogates or proxies for mismeasured and unobserved confounders in an effort to reduce residual confounding bias. Since then, many machine learning and semi-parametric extensions of this algorithm have been proposed to better exploit the wealth of high-dimensional proxy information. In this tutorial, we will (i) demonstrate logic, steps and implementation guidelines of hdPS utilizing an open data source as an example (using reproducible R codes), (ii) familiarize readers with the key difference between propensity score vs. hdPS, as well as the requisite sensitivity analyses, (iii) explain the rationale for using the machine learning and double robust extensions of hdPS, and (iv) discuss advantages, controversies, and hdPS reporting guidelines while writing a manuscript.

DOI: https://doi.org/10.1080/00031305.2024.2368794

This workshop aims to familiarize participants with available tools in R for analyzing time series of functional data objects. Topics will include visualization, forecasting, autocorrelation analysis, and change point analysis.

Learning outcomes include:

• Preprocessing of functional data objects

• Rainbow plots for visualization of functional time series

• Forecasting models and model selection/goodness-of-fit

• Change point analysis

Stochastic process models in one and two dimensions are increasingly important in applied modelling of real-world data sets in finance, epidemic and infectious disease modelling and other health-data analysis.  In this workshop we will study the Bayesian analysis of data for which such stochastic process models are widely deemed to be appropriate. Building on methods developed for non-stochastic (ODE) models, we will illustrate the use of Markov chain Monte Carlo (MCMC) approaches for the analysis of stochastic process models driven by Brownian motion but also more a complicated driving process, the Levy process.  Methods studied will include those based on linear noise approximation, particle methods and (where possible) exact Bayesian computation. We will also study Bayesian methods and MCMC approaches to spatial (point process) data.

This Workshop will be in person but, due to unforeseen circumstances, the presenter (Prof. David Stephens) of this Workshop will be virtual.  He will have his PhD student on site to coordinate the delivery of the lectures.

Research is at the core of scientific inquiry. Science of Statistics has a broad range of applications in research such as health, medicine, finance, psychology, etc. Mentoring students in research in the course of their education gives them opportunities to critically examine and affirm their commitment to their future career. Mentoring undergraduate researchers has additional benefits. It can help senior mentors educate new researchers and create more inclusive disciplines while advancing their own science in partnership with smart, enthusiastic, passionate, and hard-working students. It also provides junior mentors with opportunities to develop professional skills in advising and teaching. Both these forms of teaching and mentoring measure success by enhancements in understanding of complex topics and critical thinking skills. Each mentor has a different approach and uses different methods in working with their students, but there are common approaches and practices that mentors can use to contribute to the success of every student researcher. In this workshop, we facilitate activities drawn from best practices in Statistics education by utilizing the work by Aaron M. Elllison and Manisha V. Patel: Success in Mentoring Your Student Researchers, Moving STEMM Forward, as well as sharing our own practical ideas for involving and mentoring students in research. 

Biographies:

Omidali Aghababaei Jazi is an Assistant Professor, Teaching Stream, in the Department of Mathematical and Computational Sciences at the University of Toronto Mississauga (UTM). He has taught a variety of courses including Probability and Statistics, Stochastic Processes, and Experimental Design, supervised undergraduate students for their research projects, and facilitated SSC sessions and the case study competition over the past years. His research interests are Statistics education, analysis of biased survival data, and analysis of longitudinal data with informative follow-up.

Diana Skrzydlo is an Associate Professor, Teaching Stream in the Department of Statistics and Actuarial Science and the current Math Faculty Teaching Fellow. She has been teaching at the University of Waterloo since 2007 and has spoken widely on innovative teaching and assessment techniques, including in Indonesia with the READI project. She has a BMath (2006) and MMath (2007) from UW, and achieved her ASA designation from the Society of Actuaries in 2018.

Lijia Wang is an Assistant Professor, Teaching Stream (LTA), in the Department of Statistical Sciences at the University of Toronto. He has taught Statistics courses including Probability and Statistics. His passion is Statistical education, and his research interest includes causal inference, causal mediation analysis with application in designing novel statistical models for biomedical data. 

James McVittie is an Assistant Professor in the Department of Mathematics and Statistics at the University of Regina. He has taught statistics courses at all levels of undergraduate studies and has supervised students at both the undergraduate and graduate levels. His research focuses primarily on survival analysis as well as problems related to missing data.  

Gracia Dong is an Assistant Professor, teaching stream, joint between the Human Biology Program and the Department of Statistical Sciences at the University of Toronto. She received a PhD, MMath, and BMath in Statistics from the University of Waterloo. Her research interests involve healthcare access equity, vulnerable population modelling, statistics education, and quasi-random number generation.