1. 5%
2. 5%
3. 95%
4. probably generally the same - you might be more or less extreme.
5. The sources are the two sites you linked to, plus the wikipedia article plus my impression that 1 man killing a girl in a sexual assault is a priori much more likely than three relatively recent acquaintances doing same.

NB - 5% and 95% are basically just proxies for 'pretty sure it's not true' and 'pretty sure it's true'. I wouldn't claim that they're anything like well-calibrated. I don't think I form opinions on stories like this often enough to be well-calibrated.

Oh, and I second the opinion that this is exactly the sort of thing that LW needs more of - people actually attempting to see how good they are at being right about the world (although maybe issues in which it's easier to be sure of the "right answer" would be better tests).