Date:         Wed, 9 Oct 2002 10:21:24 +0200
Reply-To:     Erkan Yilmaz <Erkan.Yilmaz@stud.uni-karlsruhe.de>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Erkan Yilmaz <Erkan.Yilmaz@stud.uni-karlsruhe.de>
Subject:      AW: How to get non-linear correlations?
Comments: To: "Jarod_Russell, The_Pretender" <jarodx@hotmail.com>
In-Reply-To:  <F131J77msfi7ldyOIwy00013408@hotmail.com>
Content-Type: text/plain; charset="iso-8859-1"

Hi Jarod,

in Excel you have the possibility.
On the x-axis you draw your data1 and on the y-axis you draw data2.
Then you can raw a trendline (e.g. polynomial, logarithmic,...) with
diagram -> add trendline.

But when you have 100 sets of data-pairs
e.g.
set1: data1 with corresponding data2
set2: data1 with corresponding data2
...
set100: data1 with corresponding data2

so you have to plot every set for itself and then get the trendline.
And you see that is some kind of effort.

Perhaps someone else knows anything better?
I would also be interested in getting this with SPSS!

Thanks in advance!
Erkan


-----Ursprüngliche Nachricht-----
Von: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] Im Auftrag
von Jarod_Russell, The_Pretender
Gesendet: Mittwoch, 9. Oktober 2002 01:28
An: SPSSX-L@LISTSERV.UGA.EDU
Betreff: Q: How to get non-linear correlations?

Hello,


First I want to introduce myself:
I am Jarod and studyin at the moment.

I have data I gathered a long time.
So I did make the spearman correlation for the big amount of values,
(of course I checked the data for not-normal distribution first).
Then I made the average with the Fisher-Transformation.

Then I got the values between -1 and 1. Unfortunately I do get a lot of
values in the range of [-0,2;0,2]. So that does mean that it could be a
non-linear correlation.

I wanted to know if there is a possibilty to get such kind of non-linear
correlation with SPSS!

Jarod

_________________________________________________________________
Chat with friends online, try MSN Messenger: http://messenger.msn.com