Maybe you go a little too far. I agree that > 19% of kids encountering "unwanted sex talk" on the internet might be acceptable. But I wouldn't say that 1/1501 of kids actually having sex with adults they meet on the internet is per se an acceptable number, as you suggest when you say 'People who cite the study can't have their cake and eat it too, taking the "1 in 5" number as accurate but discounting the fact that none of the teens surveyed reported a sexual relationship with an adult they met online.' To me the margin for that particular category ought to be so small that this sample doesn't really measure it.