Good question and I think there are. Here are my reasons:
1. Even if all info about pwME on this site, when prepared for statistical analysis, is properly anonymized e.g. by replacing their names by a number, it is likely that one who would want to and is savvy in the relevant ways, may link up the symptoms of patient nr N on the statistical survey with the stories of the patient who calls him or herself Y on the forums, in some cases at least.
It is also difficult to make this impossible, but then the reason would be that Y did say rather a lot about Y on the forums and filled out the statistics for Y (that anonymized ended up as patient nr N) conforming to Y's utterances on the forums.
2. In general, I don't believe the internet is safe - and I am computersavvy.
This being said, I also think the PR forums offer a hitherto not existing chance to gather and analyse data from a lot of pwME, and I do think it should happen.
If you were to ask why (I think not, but anyway):
E.g. because: The group is huge, compared to standard groups used in so-called scientific research; it's highly probable that the vast majority of the members do satisfy the Canada criterions; their is a wide spreading in the group geographically and as regards health-problems (who suffers from which symptoms in which intensities); while good statistical analyses of the data one may get, e.g. by a good questionnaire, that also involves measures to assure anonymity coupled to exhaustiveness of data, is easy with modern computers.
But as I said, privacy-issues ARE important here, especially for pwME in England, the US and Holland (where being known to "the authorities" - e.g. of the dole - as such-and-such on the PR forums may create major problems for one) and therefore should be carefully considered.
OTOH: I would not be amazed if analysing the data - when properly assembled - of the pwME at PR may help clarify or answer several problems around ME/CFS, at least, and I myself would be quite excited, qua scientific and logical psychologist, to have such data to analyse.
Indeed, my guess is that Real Scientific Researchers would be interested too and personally - if they exist and are bona fides i.e. not associated with KCL or CDC etc. - I'd prefer leaving the more intricate statistical analyses, such a ANOVAs (analyses of variance i.e. essentially what - group of - symptoms correlates with which - group of - symptoms) to them (essentially because (i) this is tricky business that (ii) requires good statistical programs that (iii) the academic pros have anyway).