Early comments on stories affect what later readers believe, and what they say

A recent scientific experiment demonstrated the importance of intervening in comment sections to cultivate constructive discussion, particularly just after publication.

Scientific American Blog Editor Bora Zivkovic writes about the results, which showed that the tone of pre-existing comments on a story affected subsequent readers.

An article about nanotechnology, a topic most people know very little about and usually have no a priori biases for or against, was presented to the test subjects. Half the people saw the article with (invented) polite, civil and constructive comments. The other half was given the same article but with uncivil comments – essentially a flame-war in the fake commenting thread. The result is that readers of the second version quickly developed affinity for one side of the argument and strongly took that side, which affected the way they understood and trusted the original article (text of which was unaltered). The nasty comment thread polarized the opinion of readers, leading them to misunderstand the original article.

The Guardian saw a similar lesson when it tried two commenting systems simultaneously — Facebook comments within its Facebook app, and traditional comments on Web pages. Former user experience chief Martin Belam writes:

I had rather hoped that by opening two commenting threads underneath each article — one on Facebook, and one on the Guardian site — we’d be able to prove once and for all whether one or [the] other led to better interaction. In the end, it appeared that actually the tone set early on in a comment thread looked like it influenced comments much more than anything intrinsic about the format or identity system used.

Journalists who have written off comment sections as forsaken wastelands should still be concerned with this problem — because rancid comments also spoil the perception and potential impact of your content.

So how do you get the kinds of comments necessary to seed good discussions and avoid meltdowns cheap jordans from china?

That seems more difficult than ever, unfortunately.

Technology is not enough

The act of publishing is now so democratized and social media so pervasive that most everyone whose musings are worth hearing probably has found their own personal avenues of expression.

Smart people with something constructive to say about your article may be posting their thoughts to their Twitter or Facebook or Tumblr. Your comments section could be left as a second-class wasteland suitable only for logical fallacies and trolling.

Major publishers like Politico and TechCrunch recently announced they were dropping Facebook-powered comments and switching to other platforms (Disqus and Livefyre). That renewed debates about which platform produces better discussions.

But most people with experience in the field seem to believe, as Belam says, that “software design and features do influence community behaviours, but not as much as decent community management and personal engagement from journalists does.”

Dan Gillmor recently shared some thoughts about how that might work:

If I could design a comment system, it would put all anonymous comments at the thread’s end, and give the site owner an easy way to move good comments higher. I’d also give users a way to make anonymous comments invisible. Most sites, at this point, require a working email address and let users post under pseudonyms. This, too, can be abused by a troll, but it injects an element of accountability.

In the end, accountability is up to the site owner. Whether you are a lone blogger or a big news organization, comment threads are a platform you make available to others. The thread is your living room, where you’re hosting a conversation. You invite people into your home, and you make the rules on how they should behave.

Maybe “better comments” is the wrong goal. Maybe we need something “better than comments.”

Fresh approaches

The Huffington Post — which received well over 70 million comments last year — is launching a new comment-highlighting tool called “Conversations.”

It plucks discrete discussion threads out of the sea of comments and elevates them to their own Web pages. PaidContent’s Jeff John Roberts has the details:

The new set-up should make it easier to jump in on a given debate about the story that’s of interest. In the Benghazi story, for example, groups of people can find each other to discuss specific facets of the story — whether the US should be in Libya; whether the incident was Hillary’s fault; whether Hillary is actually a Muslim agent sent from Mars to destroy America and so on.

The fact that the “Conversations” will now have their own URL also makes it easier for people to share them and invite others into the discussion.

Gawker has been pushing its comments in a new direction too, with a focus on creating distinct, focused conversations and giving the person who started each conversation control over the responses.

Others are arguing that new systems of “social annotation” will replace commenting forms. One startup to watch is Hypothes.is — an open-source platform for annotating content across the Web. It will act as an overlay that participating users see on top of content as they browse, so individual website owners will have no control over it. But the notion is intriguing.

Reuters’ Felix Salmon looks at another annotations system used by the Rap Genius website to crowdsource understanding of rap music lyrics. The site’s users annotate each line of song lyrics with explanations.

Salmon is enthusiastic about the idea’s potential to spread:

If this takes off, it could be a significant evolution in the way that we talk about Web content. Right now, for instance, if I want to link to something somebody said on a Web page, I’ll normally just end up linking from Twitter to an undifferentiated page, rather than to the specific thing being said. And more generally, the conversation around things like blog posts tends to happen mostly on Twitter and Facebook, where it’s easy to miss and almost impossible to archive.

It would be amazing if annotation could change all that, helping to make comments more on-point and also providing a centralized archive of the conversation around any given story. … Internet comments are more of a bug than a feature these days, and I do think that annotation is a very promising way of potentially addressing the problems they have.

Related: Ben Smith: “It’s crazy that people still read, much less write about, blog comments” | Monday was Community Manager Appreciation Day

We have made it easy to comment on posts, however we require civility and encourage full names to that end (first initial, last name is OK). Please read our guidelines here before commenting.

  • http://www.facebook.com/profile.php?id=1317999050 Kellee Leamy

    It’s not something I’ve ever really thought about before. Now reading this article and thinking about the comments I’ve read and responded too, this does make a lot of sense.

  • http://rocrastination.com/ Ro Gupta

    Thanks for the roundup, Jeff. We’ve been looking at these pieces closely at Disqus.

    Reposting a bit from a comment on Martin’s write-up…we often say ourselves that software is maximum 50% of the equation. Looking across Disqus, we see that the best communities tend to set norms and a distinct tone upfront, and reinforce them through active, ongoing author engagement. Self-policing by commenters is also key, which tends to happen more when the publisher’s staff is consistently present. i.e. The ‘broken windows theory’ and base human psychology are absolutely the core factors in all of this.

    That said, we have been encouraged that there are still a number of things software can really help with. e.g. The newest Disqus has a more sophisticated voting and sorting system and incorporates built-in checks for improper use or distorting the results. Upon moving to the new system last year, abuse reports decreased 80% and moderator workload decreased 25%, so we felt we should keep working on these areas where technology can make a difference. Visibility management and use of quality and reputation signals are the big ones for us.

  • ClaytonBurns

    Typically, the link to Ta-Nehisi Coates is a fragile one (“More” on Coates does not come up).

    The nature of journalism discourages good comment. Far too many journalists are just churning various happenings without thinking about what they are seeing. (The NYT is heavily biased towards externals in its front news section.)

    It might cross an editor’s mind that the ashland Tom Buchanan Yale-style state of commenting tells us something profound about human cognition and education, but focusing that is out of the question.

    Let’s take an example, if we are not too distracted to consider it for even two minutes:

    If you work with students for 40,000 hours, you will notice the power of latency. It is the most important and least understood factor in education.

    If I take an extremely good text such as Mark Ashcraft’s “Cognition” and move it back to high school so that students take five- or six-week summer schools for three summers at the ends of grades 10, 11, and 12 on just this text as split up into coherent units, then I will build latency so that in two or three years many students will come to a global awareness of some subjects in cognition. That is how the human mind works.

    When we remember, we don’t recall a pick-up-sticks set of fragments and then assemble them laboriously. The memory images emerge globally and vividly. Hectic SAT tests are psychologically unreal because there is no curriculum, and thus, in comparison with a true curriculum, no latency at all. None.

    A journalist would never discover this elementary opportunity cost in 1000 years. Nor would an economist, who should easily be able to scan reader comments and grasp that downstream opportunity cost effects are killing education.

    It is a waste of time to comment anywhere. I thought there would be some value in commenting at Guardian Education, but the readers turn out to be hopelessly ignorant, and the education editor is as much of a twit as the distracted fools who edit The New Yorker.

    If there is no follow up, there is no point in commenting.

    There should be real names comment streams where you would have to have something to say, and the website would not ignore you.

  • http://twitter.com/westseattleblog West Seattle Blog

    Absolutely true. Have seen this time and time again without formal study. And your point that the comment platform isn’t what matters should also be heard loud and clear. A couple years back, there was a fad of rushing to Facebook comments because somehow that would solve the problems, no one would troll, etc. We declined to join the rush, noting that we had seen people be just as nasty on our FB page with their real name as some had tried to be on our site with a “handle.” A la TechCrunch’s recent admission, that change hasn’t done much for many except suppress participation. Even if you are a one-person band … you can and MUST get involved in the discussion. And if your site purports to have rules, somebody must enforce them.

  • dkiesow

    The only problem with comments on media sites is that they are often ‘set it and forget it.’ Like any community comments need to be tended to.

    Check Ta-Nehisi Coates at The Atlantic for how he interacts in the threads. Good example: http://www.theatlantic.com/politics/archive/2013/01/chris-christie-on-the-nra/267317/

    Technology can help but if you can’t ‘afford’ the human touch in your comments, don’t have comments.