Exactly - how on earth do you compute 'teen email use'? If it is self-reported it is likely to be inaccurate. If it's based on some type of network analysis, the amount of spam would skew statistics, because spam is only sorted generally at the point of arrival.

Anyways, color me skeptical.

Even if it was Facebook messaging taking over from email, that's hardly a problem, because in the eyes of the user, facebook messages are email. You have an inbox, sent items, new message, etc. If/when the teens get to a job or study, they aren't going to be stumped by viewing a messaging client instead of a facebook page.