I’ve always been torn by the compromises of anonymity, identity and digital engagement. Not just in terms of participation levels but also in terms of comment quality – such as how do real identity, pseudonym and anonymous comments compare? One major criticism of prohibiting anonymous users in online communities is that it limits participation.
Well, the computer science department at Pomona College have just released some good research on this which I have been pouring over.
Previous research by Disqus in 2012 revealed that the average pseudonym user (i.e. that has an associated username) contributes 6.5 times the amount of content as the average anonymous user and 4.7 times as much as the average Identified user (i.e. has a linked Facebook profile). They also claim that 61% of Pseudonym posts are positive, vs. 51% for Identified posts and 34% for Anonymous posts. They also found that, generally speaking, the number of comments and likes is smallest for anonymous users.
The latest study conforms to this. It turns out that comments from the Identified users are more relevant than comments from their more anonymous counterparts (no surprises here?). In other words, more relevant comments are associated with more revealed identity. Interestingly, though, relevance from Email Verified users is considerably lower.
So what about words? Well, the nature of the words used in online dialogues (e.g. their complexity versus the various levels of identity) is relatively constant but the length of sentences is a more volatile feature. That surprises me a bit – I expected that identified persons would use more intelligent words in order to demonstrate knowledge in a debate.
Likewise, groups with more identity associated with their comments user less swear words, less anger words, more affect words, more positive emotion words, and less negative emotion words.
Another interesting finding is that longer discussion threads become increasingly less dominated by Anonymous users. This suggests that when more identity is revealed you get longer discussion threads.
What does all this mean? Well, it would point to a good argument for a balanced approach to identity. In other words, full identity disclosure may stifle participation but modest identity (e.g. pseudonym or registered) strikes the right balance between participation and quality of debate.
When full identity is needed, email verification is clearly a bit of a joke – but we’ve all been there and done that!