[catpolice]

I think some of the math going on here is interesting and probably has some interesting consequences for people's expectations about means. But I'm not sure about that paper...

I want to be generous here and assume I'm misunderstanding, but it does seem a bit like the argument begs the question a bit.

The intended conclusion is that we should consider non-earth-like (i.e. non-earth-sized) planets as just as likely to be inhabited as earth-like planets. Which is to say that we shouldn't expect that population density is strongly correlated with planet size.

And this is shown starting from a model where "mean population density is invariant to planet size". Hmm...