Comment by fiddlerwoaroof

I'm far from an expert on statistics, but wouldn't the fact that star ratings tend to fall on a bimodal distribution just mean that the ratings that don't fit the bimodal distribution have a higher signal/noise ratio than the ones that do? I'd think you could use a 2-4 rating as an indicator of a thoughtful reviewer. Additionally, it seems like you could also use the fact that a particular reviewer doesn't fit the overall bimodal distribution as a sign that this reviewer's reviews are similarly more thought out.