Thursday, May 22, 2008

What Inference Can Be Drawn From Amazon Book Reviews?

I find book reviews extremely helpful from Amazon. I sift through them and try and glean which reviewer "is most like me" and analyze what he thought. Language, spelling, insight, and brevity are all used as signals. (Note: I am not ranking myself highly in any one of those categories, but rather looking for similar patterns. For instance, I want someone who will not use "u" instead of "you" but does not have enough time to focus on grammar extensively.)

For academically oriented books, if there are less than 5 reviews and they are all good, I discount them heavily. They are probably his friends or colleagues. If there are less than 5 reviews and they are all bad, unless I get the impression the reviewers are misguided based on their criticisms, I move on to the next book. Less than 5 reviews and none good? Couldn't they even get their friends or colleagues to give it a favorable review?

Please chime in on any methods you use.

No comments: