Ivan, I am glad that you do not appear to be dispirited by the comments from group members pointing out the flaws in the method of data collection on the subject. You have put in your own time on this project and I congratulate you for having a go. The members who have written in pointing out the flaws are quite correct. They have said what I would have said. I also have diabetes in my family, but resisted e-mailing you, because I felt I would add to the "biased results" that others have mentioned. However, I feel that you should not give up, but carry it forward in a way which WILL yield some valid data. I will therefore come up with some constructive suggestions which I hope you and others may feel free to take up, modify, or ignore as you wish. 1. Before we proceed with the survey, is it possible to determine if any data already exists on a possible connection between diabetes and PD? There is no point wasting time on a survey if such data already exists but is being ignored by the medical researchers for whatever reason. 2. Can we obtain data which shows the incidence of diabetes in the population at large and, for example, how many people would be expected to have parents or grandparents with diabetes? 3. Assuming the above works out, I would encourage you to proceed with your survey in the following way: a. Go for very clear and simple data - too much data can leave the picture muddy and needs larger numbers to be valid. Suggestion, ask only for PWPD who have parents or grandparents with diabetes. b. Go for a large target - my suggestion, go for the whole list!! METHOD. 1. Ask for volunteers to help you with the survey. 2. Allocate each of your volunteers a letter of the alphabet and make them responsible for data from members with surnames commencing with that letter. 3. Post a message to the group asking for your data, and requesting that members send their repies DIRECT to the e-mail address of your helper who is dealing with their name. They should NOT send it via the group because we do not want the group clogged with data which would be of little interest in itself. 4. Ask your volunteers to obtain a server return of group members and, after a suitable interval and reminder via the group, chase up the data from the stragglers. Your volunteers should then send in their data in the form you have asked so that collating it should be easy. 6. You could use the caregiver, as long as they are not a blood relative, as a control? This method, or a modified version, could give us a response of at least 1000 PWPD and a similar figure for the controls. I know we have statisticians in our group, perhaps they would help/advise? I would be happy to take "One letter" should you wish. The data is there Ivan, - Go for it! Ernie Peters <[log in to unmask]>