That's sad to hear.  I don't doubt that the effect they identified can occurred.  But it's unfortunate that they couldn't find any solution to address it satisfactorily.  Of course, this is an old problem, and there are many solutions that have been tried:

* Straight up moderation (e.g. by an employee/employees of the newspaper, or a small list of trusted people) before posting appears (fails, mainly due to the time required and resulting lag).
* Real-name comments (my understanding is people are less likely to be rude if using their real name, though I don't think it magically solves the whole problem)
* Meta-moderation - Let readers and active participants moderate, and let/require other active participants check whether they're doing a good job

The public policy problem they discussed (everyday people suddenly thinking they're expert scientists and that real scientists don't know what they're talking about) is also real and worrisome.