The German Research Foundation (DFG) is funding LMU scientist Anne-Laure Boulesteix: How can we improve research on statistical methods, asks the biostatistician, and thereby also improve their use?
The German Research Foundation (DFG) has awarded LMU biostatistician Prof. Anne-Laure Boulesteix a Reinhart Koselleck grant worth 750,000 euros. This award has been given out annually since 2002 to just a handful of established scientists at research institutions throughout Germany. It supports exceptionally innovative and promising projects in all areas of scientific endeavor. Not least of all, Reinhart Koselleck Projects give researchers more freedom in the pursuit of “in a positive sense, higher-risk projects” as the DFG puts it.
Anne-Laure Boulesteix is a professor at the Institute for Medical Information Processing, Biometry and Epidemiology at the Faculty of Medicine and an associate member of the Department of Statistics at LMU, where she leads the research group for biometry with a focus on molecular medicine. She is also a PI at the Munich Center for Machine Learning and a founding member and former Scientific Board member of the LMU Open Science Center.
Her new project “Design, interpretation, and reporting of empirical evaluations of statistical methods,” which the DFG is now funding, is situated at the interface of statistics/biometry and metascience. Methodological statistical research – that is to say, the development and investigation of statistical methods – is susceptible to the same problems that are known to impair the reliability of empirical studies in other scientific domains such as medicine. These are problems that can result in overly optimistic conclusions or difficulties in translating research results into practice. They include things like poor study design and sketchy reporting. The superordinate goal of Boulesteix’s new project is therefore to strengthen the validity and utility of methodological research and literature by helping improve the methodology for comparing statistical methods. In pursuing this metascientific approach, Boulesteix plans to draw on a variety of methods, including literature reviews, case studies, simulation studies, and Delphi surveys. The work undertaken in the project will also help improve research quality in the long run in areas of empirical research such as medicine.