It's certainly not a direct correlation, but it does help to sort the spam bots and people who created an account but never use it from the rest.

However, I did mention that the karma approach would not account for lurkers which was why I proposed the logged in method. There are many other interesting data points I'd love, but without pg releasing a lot of the info, it's tough (or in some cases impossible) to gather from public data.