I wonder if it's similar to Amazon ratings, where review distributions seem to skew toward people who either love the thing, or hate it? Presumably there's actual research on this. ... Okay here's something about the J-shaped distribution (mostly 5 stars, some 1 stars, little in between) in reviews [1]. (Edit: here's another, more recent summary [2] of polarity in reviews.)
Would we expect the same in something as intimate as amputating pieces of your own body? Hmmm.
Off-topic, but I think the effect is even more pronounced with a binary rating (thumb up/thumb down) vs a 1-5 star rating. I've noticed that really bad or mediocre games will take advantage of the upvote tendency by advertising "90% positive reviews" in an attempt to trick people into thinking they're good.