Do real names really make people nicer online?

Mathew Ingram at GigaOm has an interesting look at some Livefyre research suggesting that if you force people to use their real names to comment on your site, the vast majority will just stop commenting.

Most of those surveyed said that they responded anonymously (or pseudonymously) because they didn’t want their opinions to impact their work or professional life by being attached to their real names, or when they wanted the point of their comment to be the focus rather than their identity or background. And close to 80 percent of those surveyed said that if a site forced them to login with their offline identity, they would choose not to comment at all.

…

The bottom line is that by requiring real names, sites may decrease the potential for bad behavior, but they also significantly decrease the likelihood that many of their readers will comment.

That led me to an interesting question this morning: do real names really reduce the potential for bad behaviour in comments? It’s a popularly held belief, but there doesn’t seem to be a great deal of evidence out there to support the idea that meatspace identities are any more useful than persistent pseudonyms when it comes to holding people accountable for their actions online. On Twitter, Martin Belam points out:

@newsmary @mathewi at the Guardian we seemed to definitely prove that named staff appearing early in a thread improved user behaviour

— Martin Belam (@MartinBelam) August 27, 2014

…but there’s a big gap between “named staff members early in comment threads” and “real names for everyone”.

I can find some evidence that persistent pseudonymity is a positive thing. Disqus did a very large study in 2012 on their comments database; though their methodology is opaque, their results showed pseudonymous commenters posted both the largest number and the highest quality comments across their network. Pseudonyms make people more collaborative and more talkative in learning environments; they positively influence information sharing; they encourage people to share more about themselves in the relative safety of an identity disconnected from meatspace.

There’s also evidence that anonymity – a complete lack of identifiers and nothing to chain your interactions together to form a persona, as distinct from pseudonymity, where you pick your own identity then stick with it – is a negative influence on the civility of debate, and that it engenders more adversarial conversations in which fewer people’s minds are changed.

There’s a Czech study into the differences between anonymity and pseudonymity and how to design for reduced aggression, apparently, but the actual link 404s and the abstract isn’t detailed on the difference. There’s interesting research into the social cost of cheap, easily replaceable pseudonyms, which allow effective anonymity through the evasion of reputation consequences; Reddit and Twitter are great examples of communities where you can see this behaviour in action.

Investigating this issue isn’t helped by the fact that researchers have in the past conflated anonymous and pseudonymous behaviours, but there’s increasing awareness now that the two engender big community differences; it’s also skewed by places like 4chan and Reddit being prominently discussed while less adversarial communities like Tumblr or fandom communities are less often scrutinised. (Male-dominated communities are covered more than female-dominated communities, and widely seen as more typical: so it goes.)

I’ve found one study that suggests anonymity makes for less civil comments, but I can’t access the full study to find out what it says about pseudonymity. There are suggestions, like Martin’s, that moderation helps keep things civil; there’s evidence that genuine consequences for poor behaviour helps too. And there’s this:

“While evidence from South Korea, China, and Facebook is insufficient to draw conclusions about the long-term impacts of real name registration, the cases do provide insight into the formidable difficulties of implementing a real name system.”

But where’s the proof Facebook is right when it claims its real name policy is vital for civility on the site? Where’s the evidence that Facebook comments are more civil than a news site’s because of the identity attached, rather than because the news site goes largely unmoderated and Facebook’s comment plugin broadcasts your words to everyone you know on there? Or because most people just don’t comment any more, so arguments die faster? I am struggling to find one study that demonstrates a causal link.

So this is an open request: if you know of more studies that indicate that denying pseudonymity improves comment quality, let me know on Twitter @newsmary or share them in the comments here. If I’m wrong, and it makes a big difference, it’d be great to be better informed. And if I’m right, and it’s community norms, moderation and reputational consequences that matter, it’d be great to put the idea that real names are a magic bullet for community issues to bed once and for all.