home | vita | CV (pdf) | picture | papers | talks | research
I am currently working on a project awarded by the Nederlandse Organisatie voor Wetenschappelijk Onderzoek (NWO), that is, the Dutch Organization for Scientific Research. This is an individual research project worth € 250.000 -- it does not pay any Ph.D. students or postdocs, but it allows me to buy myself out of most teaching duties and to travel freely.
Abstract: Statistical modeling is not a remote mathematical activity: using model-based predictions in public policy is nowadays common practice, as witnessed by the IPCC report on climate change, decisions on the economic exploitation of nature reserves, and many other examples. However, if statistical model analysis is supposed to be taken seriously in public decision-making, it has to be objective, reliable and robust. Can we meet these constraints? Nowadays, data-driven, "bottom up" inference techniques - as opposed to theory-based, "top-down" methods - are increasingly employed, and they promise to make statistical inference more robust, objective and policy-relevant. I investigate how far those methods can carry us, and where they can replace theory-based scientific reasoning. Thereby, the project does not only bring a central issue of modern science into focus, it also closes gaps in the current debate: Both the impact of statistical procedures on public policy and the interplay of theory- and data-based models in science have been neglected so far.
Connected to this project are in particular my stay at ACERA in Melbourne (spring 2011), as well as my work on data-based inference, precautionary risk assessment (PSA Symposium 2010), and Bayesian modeling in climate science (EPSA Symposium 2011).