If you spend any time around journalists, especially those that write about anything even vaguely political, there's a regular piece of advice you'll hear: don't read the comments.
It's not true of all sites, but for most of the internet the mix on any given thread is one part illuminating criticism to approximately a billion parts angry abuse.
Writers find this impossibly depressing partially because writers, like most humans, don't particularly enjoy personal insults, but mainly because one's faith in humankind becomes eroded by every misspelled rant or furious I-only-read-the-headline-but-not-the-actual-article screed.
However, it's generally accepted that the internet is a cultural and intellectual democracy where everyone's voice is welcome, whether they want to agree, disagree, make thinly-veiled threats or provide QAULITY V1AGRA CANADIAN PHARMACY SECRETS to the broader world.
Whenever the possibility is raised that comment threads are, for the most part, pointless bile-factories, one is immediately accused of censorship. After all, free speech – or, at least, the precious freedom to call someone a “dum fat frigid slut” or a “comunist faggit” – is protected only and entirely by unencumbered access to comment threads.
And, of course, the argument continues, websites benefit because traffic increases when a story about rates of domestic violence is accompanied by a spirited debate on, say, whether all women are bitches. Right?
Well … turns out the answer may be no. And not simply for the boring, making-the-world-a-saner-and-kinder-place-in-which-to-live reasons that you'd assume.
Adam Felder is an American journalist, as well as being the person who studies the web analytics for US news and comment site, The Atlantic. Hence he watched with keen interest when competing site National Journal decided to do away with comments on just about all their stories as a direct result of abusive threads.
Interesting, he thought. Without people laying into one another in the comments, they'd inevitably see a decrease in their traffic. Yet once National Journal killed the comments, page views per visit rose by 10%, page views per visitor rose by 14% and return visits shot up by 20%.
Felder was intrigued. Counterintuitive though it might seem, could it be that people preferred to read articles without hundreds of angry morons yelling at them in the comments? Why were more people more interested and more engaged when the comments were removed?
To get an idea of the reasons behind this result, Felder ran an experiment. He took an article from the National Journal and hired 100 US-based readers via Amazon's Mechanical Turk service (where you get humans to do the sorts of jobs that robots do badly). He gave them the article and asked them to read it. Half of them got the article alone, the other half got the article and a selection of comments.
He then got them to fill out a short questionnaire about the article, starting with some basic questions asking what the article was about, and then how they perceived the quality of the piece. Both groups answered the reading comprehension questions identically, but there was an 8% difference in how the groups rated the quality of the piece. Those that got the comments concluded that the article was worse than those who only got the article itself.
Or, to put it another way: stupid comments don't just scare readers off – they make the article itself look worse.
Now, it's worth point out that this is but a single example – and perhaps it's an argument in favour of moderating comments rather than banning them outright. In fact, moderation is the solution proposed by Justin Ellis at Nieman Lab when he looked at the same issue earlier this year, while Jesse Singal at New York Magazine argues that the Reddit approach of voting comments up or down does a pretty solid job of pushing trolls to the bottom of the pile.
However, I for one am all for preventing people screaming out their idiot opinions and unsolicited insults in comment threads.
After all, that's what Twitter is for.