I might be mis-understanding you question(s), but let me describe what we did:

We had a questionaire in response to a project which contained some 35 questions.

Each question belonged to one of five groupings: Integrity, satisfaction, ...

The questionaires were responded to by differing groups of medical professionals.

We aggregated the groupings (5 sets instead of 35), and compared each class of respondents to the aggregated grouping (minus the respondents). Because I was using Access, I used Fisher's Exact, but that is a minor detail, Chi-Square will give you the same results.

Hope this helps. e-mail if you need more detail

Robert M. Farmer Data Engineer Alabama Quality Assurance Foundation 205-907-1600 x 3441 1-800-760-4550 x 3441 rfarmer@aqaf.com

>>> Nancy Crowell <nac2@GEORGETOWN.EDU> 06/26 8:56 AM >>> I am a relatively new SPSS user. (Using SPSS 10.1 Base) I am analyzing a survey that has a number of multiple response questions. I have succeeded in getting frequencies and crosstabs, but want to be able to test for significant differences in responses for different groups of respondents. I cannot seem to find a way to do that in SPSS. Is there a way to get chi squares for multiple response crosstabs in SPSS? Alternately, is there a way to transform the multiple response set into a variable on which chi squares can be generated?

////////////////////////////// Nancy A. Crowell Georgetown University School of Nursing and Health Studies 3700 Reservoir Road NW, 139C St. Mary's Hall Washington, DC 20007 202-687-0220 202-687-5553 (fax) nac2@georgetown. edu