LONDON – So the proprietor of the Huffington Post has decided to ban anonymous commenting from the site, starting in mid-September. Speaking to reporters after a conference in Boston, Massachusetts, Arianna Huffington said: “Trolls are just getting more and more aggressive and uglier and I just came from London where there are rape and death threats. I feel that freedom of expression is given to people who stand up for what they say and [are] not hiding behind anonymity. We need to evolve a platform to meet the needs of the grown-up Internet.”
Quite so. I can see heads nodding in agreement. After all, much anonymous online commenting seems to be stupid, nasty, vicious and ignorant. And that’s just the stuff that isn’t tangential to the topic of the article being commented on. If people have to take responsibility for what they say in public, then they will surely behave better.
That seems like common sense. Whether it is supported by evidence is, however, uncertain because at the moment there isn’t much research, and what there is seems to be mostly anecdotal. The most striking study I’ve come across is an experiment conducted by the (South) Korea Communications Commission from July 2007. From that month onwards, anyone wanting to comment on any of the 146 Korean websites with more than 100,000 members was required by law to submit resident-registration or credit-card details.
The hypothesis behind the requirement was that people would behave better online if they were easily identifiable. But it didn’t turn out that way. At any rate, the commission announced recently that it was withdrawing the registration requirement because it had been ineffective in preventing people from posting abusive messages or spreading false rumors. When the regulation was introduced in 2007, malicious comments accounted for 13.9 percent of all messages posted on Internet threads. But a year later, bad behavior had reduced by less than 1 percent.
Without knowing the details of the Korean experiment, the metrics used to measure bad behavior, etc, it’s difficult to know how seriously to take this result. Personally, I’m skeptical, if only because much of the pathological behavior that we’ve observed online — from violent threats, grotesque abuse, “lynch” mobs and bullying to gossip, slander and character assassination — corresponds pretty closely to what psychologists and historians have learned over the centuries about human behavior offline. Put simply: If people believe that they will not be held accountable for their behavior, then they tend to behave badly.
This is no consolation to those who, like Huffington, want to run websites that seek to encourage user-engagement without having to spend large sums of money paying moderators to keep order and enforce civility. But obliging commenters to use their real names will have costs as well as benefits. Chief among the former will be a significant reduction in the volume of commenting, at least in the short term, which in turn will adversely affect the site’s advertising revenues. The hit might be worth taking, however, if in the longer term the quality of the discussion on the site visibly improves.
For site owners who fear that real-name commenting represents a risk too far, there is a halfway house: pseudonymity, i.e., a system in which commenters are allowed to choose their own user names, but have to associate them with a valid email address (which in the last analysis could be used as an identifier). Last year, Disqus, a company that provides an online discussion and commenting service for websites, published some useful research on the behavior of different types of site user.
The purpose of the research was to address the criticism that there is no substantive difference between pseudonymous and anonymous commenters. Disqus begged to differ, arguing that whereas real names are for authentication (establishing who you are), pseudonyms are for expression: they imply “a choice of identity” and therefore have a communicative function. The company looked through its database of 60 million users and 500 million comments to see how commenting practice differed between users who were completely anonymous, those who used pseudonyms and those who had registered via Facebook (and therefore were obliged to use their real names).
The metrics used in the Disqus study are crude but the results are nevertheless intriguing. Pseudonymous users produced significantly more “quality” comments (assessed by numbers of times a comment was “liked” and replied to) than either anonymous or real-name users. And they were much more prolific commenters: 61 percent of the 500 million comments came from pseudonymous users, compared with 35 percent from anonymous users and a mere 4 percent from the Facebook-registered crowd.
And the moral of the story? Think twice, Huffington, before insisting on real names.