Nase

Rating scales are a very complicated topic.

Wider scales allow for greater granularity and stronger effect sizes, but at a cost of reliability and speed. A wider scale can be very useful when you expect someone to rate lots of things and compare them, especially if you expect unbalanced ratings overall.

Also, even-numbered ratings scales make ambivalence impossible, which is handy when you're trying to force distinctions between things (although you have to label them properly, or else people think a 5 on a 1-10 scale is "average").

For a more fun fact, there a confounding factors involved in how people rate things. For instance, the lower your IQ, SES, and education level, the more likely you are to rate things more extremely. I haven't seen a study to this effect, but I would wager that if you make people write reviews before rating things, it will cause their ratings to become more moderate because they're putting more consideration into it.