www.statisticsdoc.com
Stephen Brand
Paul,
You can do this, with some caveats.
The procedure you describe removes the influence of demographic
characteristics from the teacher evaluation ratings, but if you are going to
relate the residuals to other variables (e.g., years of experience), you
should consider whether or not you want to remove the influence of these
variables from these other variables. You do not have to - it all depends
on your question - it is something you should consider.
This may be stating the obvious, but you should deal with degree level by
dummy coding levels, rather than entering it as a single variable. As you
note, SPSS will allow you to save residual scores from the regression
equation. Studentized residuals are useful for identifying outliers in the
data, but are not what you want for simply computing residualized scores.
Whether you use unstandardized or standardized residuals depends on whether
or not you want to interpret the residual scores in a norm-referenced form.
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
Paul Ginns
Sent: Monday, December 11, 2006 10:40 PM
To:
[hidden email]
Subject: partialling out the influence of demographic variables
Hi all,
Can I check my understanding of a procedure? I want to remove the influence
of 4 demographic variables (gender, age, degree level, and local vs.
international student status) from a set of teaching evaluation scales,
before performing subsequent analyses. Can this be done in SPSS by entering
these variables as predictors of each scale in a multiple regression, and
saving the residuals? If so, is any particular kind of residual to be
preferred (e.g. unstandardised, standardised, Studentized, etc.)?
Thanks in advance,
Paul
Locked
