Wicher’s current research is focused on statistical modelling and testing using reproducing kernels and (empirical) Bayes techniques. In this context, he developed the I-prior methodology for parametric and nonparametric regression models, which is often simpler and better performing than competing methods. He is particularly interested in graphical models and conditional independence testing. Wicher is known for his work in categorical data analysis, in particular marginal models which arise when there are nuisance dependencies in the data. With Angelos Dassios he developed the tau-star test, a scale invariant consistent test of independence.
Wicher joined LSE as a Lecturer in 2005, after completing postdoctoral fellowships at Eurandom in Eindhoven and Tilburg University. Prior to that, he studied Mathematics at Leiden University, did his PhD in Social Statistics at Tilburg University, and worked for one year as an Assistant Professor in Statistics at the Central European University in Budapest, Hungary. He has experience in the application of statistical methods in industry, in particular through projects with GlaxoSmithKline on clinical trials, and with Flextronics and Oce on fault detection in photocopying machines.