Posted by
statisticsdoc on
Feb 19, 2007; 12:32am
URL: http://spssx-discussion.165.s1.nabble.com/To-Convert-Variable-width-as-per-datamap-tp1073930p1073933.html
Sonia,
What transformations did you ultimately decide to apply and why?
A fairly common form of centering in multiple regression involves
transforming the raw scores into z-scores, and then using the z-scores in
the computation of the interaction cross-product. This procedure is helpful
in terms of placing the main effect and interaction weights on a similar and
easy to interpret metric. However, if you have already decided to transform
the independent variables, I am not sure that I would recommend that you
also apply a z-score transformation on top of everything else.
HTH,
Stephen Brand
For personalized and professional consultation in statistics and research
design, visit
www.statisticsdoc.com
-----Original Message-----
From: SPSSX(r) Discussion [mailto:
[hidden email]]On Behalf Of
Sonia Patil
Sent: Sunday, February 18, 2007 3:30 PM
To:
[hidden email]
Subject: Multiple regression centering and transforming variables
To those who answered my previous question regarding how to interpret
interactions when variables were transformed, thank you. The information
provided was very helpful.
I have another question about centering my variables. The readings I have
done recommend that I center my predictor and moderator variables. As I
undesrtand, I should first apply the transformations and then center the
variables to run the regression. Is this correct? Sorry, if this question is
seemingly simple but I want to make sure that I've got this right.
Thanks in advance for your help.
Sonia
_________________________________________________________________
Want a degree but can't afford to quit? Top school degrees online - in as
fast as 1 year
http://forms.nextag.com/goto.jsp?url=/serv/main/buyer/education.jsp?doSearch=n&tm=y&search=education_text_links_88_h288c&s=4079&p=5116