Hello All,

I have a file of survey questions which are mostly categorical responses (i.e. Do you smoke? Daily, Occasionally or Not at all). We currently have a combination of syntax and script files which produce tables of frequencies and graphs and convert them to html for posting on a web site. I am wondering if there is a way to produce confidence intervals for these estimates. In the past I have generally taken the output from the frequency tables and used another software like PEPI to come up with confidence intervals. I'm guessing that to do this in SPSS you would need syntax that reads numbers from the output. Does that sound right? Is this possible? I have found automated processes in SPSS for producing confidence intervals but only for means. I am also aware that this can be done using bootstrapping but am looking for a more basic approach initially. I may return later with a bootstrapping question. Any help would be much appreciated.

Thanks, Sherri

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Sherri Ennis, Epidemiologist Central East Health Information Partnership 17250 Yonge St., Box 234 Newmarket, Ontario, Canada, L3Y 4X1 tel. 905-764-6346, ext. 1212 toll-free 1-877-464-9675, ext. 1212 fax 905-895-0848 e-mail ennis@cehip.org <http://www.cehip.org> <http://www.healthinformation.on.ca> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^