...one thing that comes out clearly is that a lot of different individuals have contributed to the thread. If there is any such fraud, it's either not widespread (i.e. superfluous) or it is rather sophisticated.
@abgd, thanks for the pointers to Stylometry, most interesting.
I should think that some correlation with IP numbers of Poster's, and other such detail, would be valuable in terms of providing additional authority to your thesis. While I imagine this information is available to site admins it may not be of sufficient importance to employ here, although I can think of other threads and/or erstwhile users of the site that might bear a closer look...
FP.