If you want to improve your knowledge in quantile regression and analysis of covariance, we invite you to our one-day Learning Session on Quantile-Based Covariate Adjustment. The Learning Session will serve as an opportunity to study mathematical foundations of quantile-based methods in-depth. We will deal with quantile regression as well as generalized analysis of covariance (ANCOVA) models. On both topics, experts will teach about the mathematical foundations of the current state of research.

Background

Covariate adjustment is essential for improving the power of statistical group comparisons. In recent guidelines on clinical trials, for instance, the European Medicines Agency recommends adjustment for covariates associated with the outcome measure. A standard technique for covariate adjusted group comparisons is parametric ANCOVA. However, ANCOVA uses means as estimands, which are not always well interpretable, and additionally, ANCOVA relies on restrictive parametric assumptions, which are hard to verify in practice. Hence, parametric ANCOVA is inapplicable in many situations. On the other hand, quantiles are intuitive estimands for group comparisons and with QANOVA and QMANOVA, quantile-based methods are available for use in general factorial designs. Extending these methods by options for covariate adjustment is of high interest on the path towards general models with relaxed assumptions.

Speakers

The following distinguished experts will serve as main speakers of the learning session:

• Prof Paulo Serra, junior professor at Vrije Universiteit Amsterdam, who published several recognized works on quantile regression such as this recent one.
• Dr Georg Zimmermann from University of Salzburg, who proposed a semiparametric ANCOVA and dealt with several existing methods of covariate adjustment.

Additionally, there will be smaller talks by ourselves: Marléne Baumeister, a PhD student at TU Dortmund University who investigates quantile-based methods; Paavo Sattler, a postdoc at TU Dortmund University, who works on statistical methods in factorial designs; and Konstantin Emil Thiel, a PhD student at University of Salzburg and researcher in nonparametric statistics.

Schedule

The Learning Session will be a series of interactive presentations held on November 28th, 2024, at TU Dortmund University.

Registration is free of charge but mandatory: quantile.regressionstatistik.tu-dortmundde

We look forward to your participation.

Kind regards,

Marléne Baumeister, Paavo Sattler, Konstantin Emil Thiel

This event is funded by DMV Fachgruppe Stochastik and Research Center Trustworthy Data Science and Security.