Date:         Thu, 31 Jan 2002 16:33:53 -0500
Reply-To:     "P. L" <li_9@HOTMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "P. L" <li_9@HOTMAIL.COM>
Subject:      Transforming data

I have a question regarding transforming data in multiple regression
analysis. The dependent variable I'm using is no. of hospital stay, with
about 90% of values being 0 (the other values range from 1 to 31). My
question is: is it necessary to transform the data? If yes, what kind of
method should I use? I tried log, but didn't work well. Any help will be
greatly appreciated.

P. L