[jameal]

The study is looking specifically at hostility in the context of political discussions, not just general internet misbehavior.

In any case, The Penny Arcade comic seems to depict the mismatch hypothesis which the study found little evidence to support:

> Overall, however, we found little evidence that mismatch-induced processes underlie the hostility gap. We found that people are not more hostile online than offline; ...

And while they're not ruling out the mismatch hypothesis entirely, they do offer an alternative which they call the connectivity hypothesis:

> ...Thus, our findings suggest that the feeling that online interactions are much more hostile than offline interactions emerges because hostile individuals–especially those high in status-driven risk taking –have a significantly larger reach online; they can more easily identify targets and their behavior is more broadly visible.