I agree that what music someone listens to can be indicative of who they are, but it's important not to overstate the significance of the actual music versus the social context it's in. A large part of the capability to predict personalities based on music isn't rooted in the music itself, but in some other factors certain kinds of music are correlated with.
Preference of pop music or hip-hop can't in itself be a strong indicator of negative personality traits. It has more to do with the fact that those kinds of music are mainstream, and there's a high probability that a person who listens to that kind of music does so mostly because they are more apt to follow others rather than form their own opinions (or because of some other variety of conformity).
You can apply the same idea to any genre of music to any of the social subgroups. Some people like rock because it helps them fit in with some people by "rejecting" the mainstream.
If more people were attached to certain genres of music because of the musical content, there would be enough data to draw conclusions about the musical natures of individual genres. In the present, though, our data is way too noisy because of social factors.
Is there any way you guys know of to isolate the social factors?
I'm looking for someone to help me out with some WordPress web design for my blog - PM for details, will pay :)