for themselves.
That's crazy! A 5 star book dropped from one review? I didn't know Amazon wasn't averaging some of their ratings/reviews.
Another thing that bugs me is it always seems that the lowest reviews/reviews seem to rise to the top--particularly with Goodreads where major bestsellers have a lot of 1 stars at the top.
I have a love/ hate relationship with Goodreads. I even blog announcements and have garnered followers. But I'm, on average, 0.5 stars lower with all my books there. So I'm on the side of not migrating reviews to Amazon.
It is crazy, but the issue is not that Amazon omits some ratings or reviews from the average. It includes them all, but it weights them differently. For instance, reviews with the verified tag (purchased on Amazon) weigh more. Reviews are also weighted by how recent they are (which might make sense for tech but doesn't as much for books).
I learned to live with the idea that the average wasn't really an arithmetic average. What irks me is that on the bar chart that's supposed to reflect the number of reviews at each rating point, Amazon distorts the numbers to make it appear that the average
is arithmetic. A customer who bothered to look closely at the math would see something was wrong with it but not be able to decipher what--or care.
The book in question has nine reviews and one rating (the one star). The correct percentages would be 30% five-star, 70% four-star, and 10% one star. But Amazon displays it as 24% five-star, 24% four-star, and 51% one-star--which I just realized doesn't even add up to a 100%. A strict arithmetic average would be 4.4. Amazon's average is 2.7, which is roughly the result one would get using the (incorrect) percentages that Amazon provides. Having a weighted average is one thing. Misrepresenting the number of reviews at each rating point to make the weighted average look arithmetic is an entirely different thing.
The problem exists with all books to some extent, but it isn't as obvious with books where the number of reviews is larger.
As far as GR reviews are concerned, it occurs to me that I don't know whether the average is arithmetic or not. But I do know the scale is different, which would be enough in my mind to suggest that they shouldn't be put side by side with Amazon reviews unless there's an explanatory note, which I'm sure there wouldn't be.