Quote:
Originally Posted by Andrew H.
If reviews were distributed at random, 40% of reviews would be 1 and 2 star, so I don't particularly see an issue with the distribution.
|
I think this would be true if books were evenly distributed in "enjoyability", and people also chose to read and rate them randomly. You'd think that people would choose better rated books and thus push towards higher scores on average, but it certainly doesn't seem to be happening with the rater in question. Although, as others have pointed out here, I guess that many people tend to only put in the effort of rating a book if they really liked or disliked it.
I have no idea about the actual distribution of ratings over all books. Are there really as many 1-star as 3-star as 5-star books? I would be amazed if the distribution did turn out to be flat because there are a lot of non-random factors involved. The whole topic is fascinating to me just from a statistical point of view.