| |
 |
|
|
Science Forum Index » Statistics - Education Forum » multidimensional scaling
Page 1 of 1
|
| Author |
Message |
| Guest |
 Posted: Sun Apr 06, 2008 10:13 am |
|
|
|
|
dear statisticians,
I have a question concerning an MDS-analysis.
Suppose we want to examine the interrelations in an organisation.
we ask the collegeas (n=15) the two following questions:
- number your collegeas from 'I like to work with the most' till 'I
like to work with the least'. The one you like the most you give
number one, than two, and so on.
- number your collegeas from 'has the most skills and knowledge about
the problems concerning the organisation' till 'has the least skills
and knowledge'. The one with the most skills and knowledge receives
the score of one, the second best the score of two, and so on.
To examine the interrelations in the organisation we can use MDS. Now,
the following problem occured:
I can put each question in a 15x15 matrix (collegeas with respect to
other collegeas), with row conditionality and than do the MDS-
analysis. But I'm extremely interested in how those two configurations
coincide. How can I do this? Do I use three-way MDS (because all the
information I have concerning three-way MDS involves situations where
matrices stand for individuals. I want to examine two different
constructs in two different matrices)? Or can I use procrustes
procedures? Or are their other possibilities? Is all this also
possible with SAS?
Thank you in advance,
E. |
|
|
| Back to top |
|
| |
|
Page 1 of 1
All times are GMT - 5 Hours
The time now is Fri Sep 05, 2008 7:03 pm
|
|