David,

These are largely opinion pieces, like most of the articles we post on SWJ, but there is little science behind these opinions, and IMO little logic to support their views.

http://smallwarsjournal.com/blog/behavioural-conflict

They wrote at length about the importance of focusing on group versus individual beliefs to effectively shape behavior. I think this phrase effectively captures one of their arguments, “Just like research which disproves any strong connection between attitudes and behaviors, this (discussion on poverty and violence)is contrary to prevailing wisdom and, if true, has serious implications.”
The authors didn't address social media, but assuming that behavior is "mostly" driven by identity group and group expectations, is it is possible that online communities can form identity groups? Then is it also possible that these online communities can provide a social structure that motivates or constrains specific behaviors? I don't think we're too concerned about thousands of people getting online complaining, but this obviously went a step further.

I have read several studies on radicalization, and one of the salient points about online radicalization that at least one study pointed to was the ability of the radicalizer to bypass the normal social structure that motivates and constrains the targeted individuals behaviors. Parents, friends, etc. are at a loss to explain why he or she became a terrorist, not realizing that person had another virtual identity group that compelled him to change his behavior.

I'm only proposing these as ideas for consideration and do not pretend to know the answer, but I am hesitant to dismiss the potential power of social media at this point.