What mean use when few high represenated values in distribution?

Question

for a project I need to get a mean each time, this mean I then use it to compare it to a distribution under the null hypothesis H0. However sometimes some values (<