A simple method for correcting for the Will Rogers phenomenon with biometrical applications.

In its basic form, the Will Rogers phenomenon takes place when an increase in the average value of each of two sets is achieved by moving an element from one set to another. This leads to the conclusion that there has been an improvement, when in fact essentially nothing has changed. Extended versions of this phenomenon can occur in epidemiological studies, rendering their results unreliable.