Method that eliminates data that stands out the most?

Question

Is there an official name for getting a statistic used by calculation the statistic on a sample and eliminating the data that stands out the most?

 

Lets say I'm trying to get the mean on 1,4,6,4,100. I would eliminate 100 from my data set.

 

I would know how to do that, but I would like to research various methods that do these kind of process. 

Answer 1

 What you are asking for is called outlier. It depends on the circumstance and how large of a sample there is, how often they appear, and whether the samples are randomized, or  rightfully belong to the outside of the confidence interval.