Date: Fri, 26 May 2006 08:02:41 -0400
Reply-To: Peter Flom <flom@NDRI.ORG>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Peter Flom <flom@NDRI.ORG>
Subject: Re: Problem with Mean Percentage change
Content-Type: text/plain; charset=US-ASCII
>>> Madan Gopal Kundu <Madan.Kundu@RANBAXY.COM> 05/26/06 3:21 AM >>>
wrote
<<<
Difference and Mean difference have same sign. So we can expect that
mean difference and mean % difference should have same sign. But in the
above case these two have different signs.
How one should explain this contrasting nature?
Or there is some restriction about using % mean difference.?
Or we should use any other alternative methods instead of Mean
percentage difference?
Please put your valuable suggestions.
>>>
There is no reason that mean dif. and mean % diff should have the same
sign. They
ask different questions. When you ask different questions, you get
different answers.
There are two problems with comparing numbers and %: 1) Your numbers
differ a lot in scale, with
some having base near 1, some near 4, some much higher. % eliminates
these differences.
2) Numbers are symmetric around 0, % are not (that is, the opposite of a
gain of 10 is a loss of 10; but the
opposite of a 100% gain is a 50% loss)
Which question you should ask depends on what you want to know.
If you write back to SAS-L, describing what you are trying to find out,
then someone
will probably be able to help
HTH
Peter