The end is nigh for Internet trolls


Internet trolls, listen up. Your days are numbered.

Researchers from Cornell and Stanford Universities have created an algorithm that can identify Internet trolls by simply looking at the way they write.

The algorithm is so advanced, it can predict whether or not a person needs banning from studying between five to ten of their previous forum posts or social media comments.

The study entitled Antisocial Behavior in Online Discussion Communities, was part funded by Google and looks at the behaviour of Internet trolls to try and determine what it was that resulted in the being banned from an online community.

Trolls are just looking for trouble

Internet troll

The study found that trolls tend to post more frequently than regular members and that they normally exhibit poor spelling and grammar throughout their posts.

The study, which took place over an 18 month period, looked at the online communities of three websites: CNN, the political site Breitbart and gaming site IGN.

The three sites have a combined total of 1.7 million members with the research team observing more than 40 million posts to the comments section of each site.

The team then compared posts from users who had been permanently banned with users who had never been banned. Posts from users who had been banned temporarily were disregarded.

Studying the evolution of these users from the moment they join a community up to when they get banned, we find that not only do they write worse than other users over time, but they also become increasingly less tolerated by the community.

Interestingly, the study found that users who had been permanently banned from the the sites demonstrated very clear and similar behavioral characteristics.

These included their the quality of their posts or comments, which were much worse than that of other users, and which continued to get worse over time.

The study found that trolls used “language that may stir conflict” and that their posts were “harder to understand according to standard readability metrics.”

Don’t feed the troll

don't feed the trolls

Don’t feed the trolls – it only makes things worse!


Trolls were also found to behave differently with regards to the way they interacted with other members of an online community. Trolls who posted on CNN were found, on average to post 264 times before being eventually banned, whereas regular members, were found to post around 22 times during the same time period.

It was also found that tough or overzealous moderating, as well as harsh feedback directed at trolls from fellow community members was found to only making things worse and could result in an increase in anti social behaviour within the online community, which led those involved in the study to “that communities may play a part in incubating anti social behaviour”.

Are you a troll?

Let me know in the comments section below.


Comments are closed.