Posted by
sr-3 on
Aug 18, 2006; 5:39pm
URL: http://spssx-discussion.165.s1.nabble.com/multidimensional-scaling-with-rank-data-tp1070390.html
Hello Spss newsgroup,
I have a problem with the multidimensional scaling option in spss
(version12).
My data: 222 individuals who ranked 6 objects corresponding to their
personal preferences.
I want to produce a figure that shows the individuals and the objects
in a 2 dimensional space so that closeness of an individual to an
object means that the person likes the object more than other objects
that are in greater distance to the individual point.
Besides I wish to produce iso preference contours within the figure
so that (for example) two objects that are close to the same iso
preference contour are given almost the same ranks by an individual
close to that iso preference contour.
The technique to get this figure is unfolding according to the ideal
point model. I used the ASCAL algorithm (I think PROXSCAL is NOT
suitable, correct?) of SPSS but I got an error: too few cases...??
Syntax:
ALSCAL
VARIABLES= object1 object2 object3 object4 object5 object6
/SHAPE=RECTANGULAR /INPUT ROWS(222)
/LEVEL=ORDINAL
/CONDITION=ROW
/MODEL=EUCLID
/CRITERIA=CONVERGE(.001) STRESSMIN(.005) ITER(30) CUTOFF(0) DIMENS(2,2)
/PLOT=DEFAULT .
Does anybody know how to solve the problem?
Or is it impossible to do unfolding of ranking data with SPSS?
I found shareware in the internet that seemed to be suitable for
unfolding (permap) but I won't start with another software unless i'm
shure that it's impossible to solve the problem with SPSS.
Besides I would like to know whether it's possible to do unfolding
(with SPSS) with aggregated data because another aim of my data
analysis is to compare groups of individuals and their preferences.
Thanks for help!
Susan
--
Der GMX SmartSurfer hilft bis zu 70% Ihrer Onlinekosten zu sparen!
Ideal für Modem und ISDN:
http://www.gmx.net/de/go/smartsurfer