Supposethesampledatadistributionof {yi} = {y1, ...,yn} is veryhighlyskewedtothe right,andwetakelogsandanalyze {xi = log(yi)}.

(a) Is ¯x = log(¯y)? Whyorwhynot?

(b) Ismedian({xi}) = log[median({yi})]? Whyorwhynot?

(c) Tosummarize {yi}, wefind ¯x and thenuse exp(¯x). Showthat exp(¯x) = (Πi yi)1~n, called the geometricmean of {yi}.