The bit I find intriguing is how the paid subscription packages aim to improve your diagnosis by actually reducing the number of eligible participants.
That and (according to the FAQ) the unusual design decision of making flagging public. I wonder if the "try homeopathic magic pills" suggestions are efficiently eliminated, and whether reciprocal flagging behaviour has ever been an issue.
The more expensive packages only allow higher-rated Medical Detectives to fully participate, which reduces the quantity but increases the quality of participants. FWIW, we've found that beyond 30 or so Medical Detectives, results quality doesn't really improve anyway.
We've gone back and forth a lot on whether to make flagging public. The advantage of public flags is that users flag more carefully, the disadvantage is the potential for reciprocal flagging behavior.
That and (according to the FAQ) the unusual design decision of making flagging public. I wonder if the "try homeopathic magic pills" suggestions are efficiently eliminated, and whether reciprocal flagging behaviour has ever been an issue.