 |
 |
Ivan M Suzman wrote: > > Dear listmembers and friends, > > 1. DIABETES SURVEY APPROACHES GOAL!! > > Thank you for being so RESPONSIVE to the DIABETES SURVEY. As of this > morning, 84 responses are in. I am storing them in a diabetes folder in > my computer, so that they can be easily retrieved when it comes time > (probably later this week) to try to analyze the information at hand.. > > THE GOAL IS 100 responses. We need just 16 more listmembers to > answer the survey. Could 16 (or more) survey responses please be sent? > > 2. PLEASE SEND YOUR RESPONSE--IT'S NOT TOO LATE!! > > PLEASE send in your brief message, stating whether you have a > DIABETIC RELATIVE (genetic, not by marriage). Please also indicate, if > it is your grandparent, aunt or uncle, whether it is a maternal or > paternal relative. > > If it is a niece or nephew, please state if it's YOUR BROTHER'S OR > SISTER'S > child. > > 3. IS THERE ANYONE OUT THERE WHO MIGHT LIKE TO ADVISE ME?? > > I'm not a geneticist, so those of you who have ideas of what you > would look for in our diabetes survey, SEND IN YOUR IDEAS!! > > UNTIL LATER, > > Ivan Suzman > Portland, Maine is rainy and 39 deg. today Ivan, It is certainly interesting to see so many of us with PD have Diabetic relatives. (my paternal Grandmother was my only close relative with Diabetes) Unfortunately, Diabetes is a very common illness and I think it would be difficult to make much out of the results in terms of statistics. The problem is that you have a skewed self selected sample. You have gotten responses from people who the question struck a chord with. Almost certainly you have received more responses from people who have diabetic relatives than from those who don't. Without a truly random sample there are no conclusions that you can safely draw. You might be able to get some indication as to whether the PD tends to be on the maternal or paternal side but I can't think of any other analysis that would be useful from the data. The problem with surveys of the list population have this problem in general unless the event (e.g. toxin exposure, genetic history etc. is so rare as to jump out at you when you get a cluster. For instance if 20 people on the net discovered that they had a close relative with PKU, then because of the rarity of that illness one might want to follow up with a random survey. Also, list members almost certainly tend to be younger and come from a higher educational and socio-economic group. The most famous example of mis-sampling was in 1948 when the Chicago Trib used a telephone survey to predict the presidential election not taking into account that more affluent people tended to have phones while lower income people didn't. If you question the significance of this the ask President Dewey. I think the usefulness of the data is quite limited but if anybody has any ideas you certainly should entertain them. Good try in any case. Charlie -- ********************************************************** CHARLES T. MEYER, M.D. MADISON, WISCONSIN [log in to unmask] **********************************************************