Speaker
Torben Flehmke
(Stockholm University)
Description
The likelihood ratio test statistic is guaranteed to be optimal only for simple vs. simple hypotheses. For setting limits, however, we usually deal with compound hypotheses, in which case no such guarantee exists. I apply a focused test statistic to the WIMP limits of the XENONnT SR0 dataset, concentrating the statistical power around a WIMP rate of zero. The resulting limits improve on those from the likelihood ratio by 20–30% across the 20–300 GeV/c² WIMP mass range.
Author
Torben Flehmke
(Stockholm University)