Dear all,
I have a theoretical issue that's not directly related to SPSS and therefore doesn't strictly belong in this mailing list. However, I noticed lots of theoretical disscussions in here so I hope I'll be excused for this and that someone will be willing to provide me with some clue(s).
Somewhat simplified, we measured student's satisfaction with their Universities on a 10 point Likert scale. Eventually, our goal is an intercomparison between the Universities: which one performs best with regard to this criterion?
However, some data has been collected by telephone and some by a web questionnaire. It is well known that the presence of an interviewer causes respondents to be slightly more positive. Strictly, I guess, this is an instance of Differential Item Functioning (DIF). Now, the percentage of telephone interviews differs between Universities and therefore biases the intercomparison (Universities with high percentages of telephone interviews benefit from this). We'd therefore like to (tentatively) correct the scores for this effect.
We have no respondents available who filled out both the web and telephone questionnaires so the web and telephone samples are approximately independent samples.
Some Universities have been exclusively studied with telephone or web questionnaires so this is probably one of many reasons why weighting the samples may not be a good idea.
The data are still under construction but differences between the telephone and web samples with regard to background variables (social status/age/etcetera) are likely to further complicate this issue.
However, since the web sample is expected to greatly outnumber the telephone sample, we could perhaps draw a web subsample with unequal inclusion probabilities that matches the telephone sample on these background variables as a first step.
Does anybody please have any idea how to go ahead with this? Any references are also greatly appreciated.
Thanks a lot in advance!
Ruben van den Berg
P.s. I'll post this problem in SEMNET as well.
Express yourself instantly with MSN Messenger!
MSN Messenger
Locked
