>Date: Tue, 27 May 1997 14:38:48 -0400 >From: "Susan D. Hamburger" <[log in to unmask]> >Subject: diabetes survey > I agree with what Charles and Ida have written about the >validity of >the survey regarding the incidence of diabetes in Parkinson's >families. >There is a very strong selection bias present because those with diabetic >relatives would be much more likely to respond (probably because they are >more interested). It's important to know what the rate of diabetes is in >the general population and also to have a control group to see if the rate >of diabetes is different in a non-Parkinson's group of subjects of similar > age, weight, and gender. > The right way to do it would be to take a (random) sample of those >with Parkinson's Disease and a random sample of similarly matched non-PWP's >and ask both groups if there is diabetes in their families. Comparing that >result might suggest that there is or is not a relationship between PD and >diabetes. This is awfully hard to do, which is why survey's are usually done >at research facilities, with money, manpower and expertise. I think what we >have might be 'interesting', but not scientifically valid. > Thanks for all your effort Ivan. At least it kept you off the >streets and out of trouble. > Susan Hamburger, cg for Stan 62/16 Susan's got it on the nose. However, we might still be able to obtain some useful info. It seems to me that our spouses would make a pretty reasonable control group. If we got counts of PWP with blood relatives with diabetes and compared the counts obtained for our spouses we should get a pretty reasonable picture. Of course it would take a much larger sample to get a statistically significant result, but it would be a start- Ivan- perhaps you can poll your 100 respondants and get equivalent counts for their spouses. Ron. --------------------------------------------------------- Get Your *Web-Based* Free Email at http://www.hotmail.com ---------------------------------------------------------