On Fri, 8 Sep 2006, in microsoft.public.excel.charting,
Ahab said:

In some cases a speech curve will experience what we call "rollover" i.e.
the ability of the patient to hear speech will deterioriate (rather than
increase) as the intensity of speech sounds are raised. This effect is
characteristic of neural lesions and back in the day before MRI used to be
used as part of diagnostic assessments of brain tumors amongst other things.
Its very important then that any trendline that is drawn from the data will
accurately display this rollover... My question is then, is there a way I can
use these functions or any others to get the results from the x axis with a
different sort of trendline that can reflect the above mentioned rollover
effect?

A polynomial function might do it to your satisfaction, but check the
literature to see if there are models for the resulting curves in
hearing acuity. It's best to follow a model if there's one available.

In general there is a way to use Excel to get a curve fitted for any
function you can think of. I confess I never have used the TREND() or
LINEST() functions myself; I always preferred to use a roll-your-own
version which I fitted using the Goal Seek add-in under Tools. This may
mean I have unrealistic expectations of the built-in Excel curve-fitting
functions.

--
Del Cotter
NB Personal replies to this post will send email to ,
which goes to a spam folder-- please send your email to del3 instead.