Quote:
Originally Posted by ReadingManiac
Always have to be careful of those. If the reviews don't reference something actually in the book, it could be a personal vendetta. On the other hand if a book really hasn't been properly edited or is written badly, that can get a lot of one star reviews.
|
Quote:
Originally Posted by issybird
As far as the personal element goes, I'm most suspicious of a disproportionate amount of five-star reviews, which seem most likely to be driven by F&F and generally account for those 4+ averages. I think the most useful reviews are the two-star ones, which tend to focus on the substantive, but for an absolutely terrible book there is no alternative to one star.
|
I confess that sometimes I have been a hurry, and just considered the average score at Amazon (and hopefully the score at GoodReads, too) before purchasing. That might have been a mistake, although I can't point to any particular cases where that went wrong.
Issybird's calculas for judging the validity of ratings is interesting, to say the least, though; I certainly will give it consideration!
Amazon recognizes the issue of whether or not the average rating give a true picture of how good a book really is. As I mentioned in post #1670, Amazon states that they do not use raw data to come up with an average score. They use a "machine model" to come up with what they think is a more accurate average score. I suppose that that means that they use a Bayesian approach, algorithms, something(s) like that, (a) thing(s) which are way above my mathematical head.
That might make me more confident to pull the trigger on books at Amazon which have a high composite rating. When necessary.