Rianne de Heide, assistant professor at the University of Twente, has received an ENW-M-2 grant of € 742,708 of the Dutch Research Council for her project Flexible and user-adaptive statistical inference. This research is a collaboration with Jelle Goeman of the Leiden University Medical Center.
The project focuses on the development of a new statistical method that offers researchers more flexibility in their research design and analysis. With these new methods, researchers can adjust their research questions and sampling plans based on incoming data without resulting in more false positives. A false positive result is a result of an experiment that is falsely positive.
A new statistical approach to modern data
In academic research, it is common to calculate the so-called 'p-value'. With this value, researchers aim to separate reality from chance. "The p-value was developed more than 100 years ago. It is suitable for small-scale experiments, such as those that took place at the time, but nowadays we work with much larger amounts of data and often test hundreds to hundreds of thousands of hypotheses at the same time," says De Heide.
That is why De Heide, together with colleagues, developed the so-called 'e-value' in 2019. This fairly new statistical method is much less susceptible to fraud than calculation of the old-fashioned p-value. E-values allow for flexible sampling plan, and are particularly useful when testing multiple hypotheses simultaneously.
About ENW-M-2
In the Open Competition ENW-M (Domain Science – M), researchers can submit research proposals either individually or in collaboration for non-programmed, curiosity-driven fundamental research. ENW-M-2 funding amounts to a maximum of € 800,000, with one main applicant and one co-applicant who has complementary expertise.
The ENW-M-2 grant will be used to appoint two PhD students and an academic programmer, who De Heide and Goeman will supervise jointly. Rianne de Heide: "I am extremely happy with this grant from the Dutch Research Council. This project gives us the opportunity to develop a pioneering statistical method that is desperately needed in today's day and age."
Learn more
Dr Rianne de Heide is an assistant professor in the Mathematics of Operations research department (MOR; Faculty of EEMCS). There, she works on problems in and solutions for machine learning and statistics. She is currently working on her VENI project 'E-values for Multiple Testing'.