I am conducting a factor analysis of questions for a satisfaction survey.
The sample population consists of persons who have received employment
services. Naturally, some close in employment, and others do not. Our
sample is biased, because in our respondent sample, those closed in
employment outnumber those who closed without employment by about 2:1,
which is a very different ratio than in the service population as a whole.
There are good explanations for this model (e.g., those closed without
employment are more likely to have moved recently, so that invitations sent
to service recipients who closed without employment are more likely to be
returned by the post office as "undeliverable.")
In addition, I have reason to suspect that those closed without employment
will probably be more discriminating in their responses than those closed
in employment, because they may feel aggrieved by one aspect, or only a few
aspects, of their service experience, so that rather than feeling globally
unsatisfied, they will be more specific about their dissatisfaction.
Consequently, their responses may have more value in determining which
questions do the best job of ascertaining their concerns.
This suggests to me that I should weight each case according to whether the
individual closed in employment or not, giving a higher weight to the
latter, and a lower weight to the former. However, I have not weighted
cases in my prior analyses, so I am not familiar with the "in's" and "out's".
Is there anything wrong with this procedure?
If not, what is the best way to do it?
Thanks,
Bob Schacht
Robert M. Schacht, Ph.D. <
[hidden email]>
Pacific Basin Rehabilitation Research & Training Center
1268 Young Street, Suite #204
Research Center, University of Hawaii
Honolulu, HI 96814
=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Locked
