Is this a "How to lie with statistics" paper, where investigator-selected parameters for the weighted average determine the result?