Moderating blog comments: Best practices from the Journalism That Matters community

 The following comments were sent to the Journalism That Matters list as the result of a query April 8, 2010 by Bill Densmore, who wrote:  "Can I ask for you please to contribute a little crowdsourced wisdom for to share tomorrow? I'm going to be participating in a one-day event at the Berkman Institute at Harvard Law School on cyberlaw topics. See:  http://www.omln.org/conference/agenda "  The are edited and used by permission.

TWO CASES TO CONSIDER:  The first involving the Cleveland Plain Dealer is elevating once again the challenge of managing anonymous comments:
http://blog.cleveland.com/metro/2010/04/cuyahoga_county_judge_shirley.html

 ALSO: KY. NEWSPAPER COMMENTER TO REMAIN ANONYMOUS FOR NOWHowever, Madison Circuit Judge Jean Logue adopts multipart test that allows for Web poster to be identified if certain criteria are met.http://www.firstamendmentcenter.org/news.aspx?id=22822   

THE QUESTIONS:

1)      On your site, do you allow anonymous comments? If so, do you pre-moderate? Do you remove comments? If so, why?

2)      In your privacy policy or service terms, do you promise not to reveal the identity of an anonymous commentator other than in response to legal process? What if you thought the legal process was SLAPP-like or chilling to a whistleblower? Would your state shield law (if you have one) protect you from revealing the identity?

3)      Is your comments management a manual process? If you use any automation can you say which vendor or software you use? How did you pick? Do you have any handy links to such services or software?

 ANSWERS:

 Jeff VanderClute (http://www.avanoo.com) (jvandercute@gmail.com)  comments:  1. I would say step back from the question. A deeper question is how to create an environment that is respectful. Avanoo does this by stating an intention to be *helpful* to all. If you don't have something helpful to contribute, it's not the place for you. We can implement all sorts of functionality for flagging, burying, categorizing (as a gem, as civil, as opinionated, etc.). Then as you suggested we could allow the user to set the threshold for displaying comments. By default the threshold might be civil. This functionality isn't implemented on Avanoo because we haven't needed it. However, I could write such functionality very quickly.  I'd be happy to explore creating a "Civil Discourse" site that serves the needs of newspapers. The way in which conversation flows on Avanoo might be better suited to discussion than the hierarchical threads offered just about everywhere else.  

Michelle Ferrier,  (Michelle@michelleferrier.com ) of Elon University, formerly of My Topia Cafe in Daytona Beach, and a blogger for Poynter: wrote "pool rules" that shouldbe posted near and  *before* people post comments, read: http://www.poynter.org/column.asp?id=31&aid=137534  

1. Also, our comments were not moderated. I too am of the mind that identities should be authentic...and that the user's identity should post along with the comment. In our privacy/terms of use, we spelled out that users were responsible for all content posted on the site. We did not moderate comments...meaning they went up on the site automatically. They were taken down after the fact if they violated the rules. Also, we had a community red flag mechanism so that if the community found something offensive, they could flag it to the editors. We would take a look and take stuff down. In the two years I ran MyTopiaCafe.com, we maybe deleted six comments, mostly because of advertising messages. Perhaps it was the good news nature of the site, but I also think it was about the tone we set as editors -- being respectful of our members and in turn them being respectful of each other.  

Martin Langeveld, (mlangeveld@gmail.com), CircLabs Inc.:

As Howard has said it's well-nigh impossible to require real-name commenting.

That said, requiring registration with email verification, rather than allowing purely anonymous commenting probably helps improve the dialogue. Facebook ID log-in as well.

But, most importantly, any comment forum is likely to degenerate without some form of moderation. News orgs should have engagement editors focused on this. They should encourage reporters and editors to interact in comment threads by rounding to questions, asking questions, correcting  misinformation, etc., and to maintain their own Tweetstreams and blogs. And news orgs should use their Facebook Pages and Twitter accounts actively to encourage interaction, not just to post a stream of news blurbs.

See this NiemanLab post for a good description of what an "engagement editor" can do: Quote:
      "Imagine if there were an opinion editor who had never heard of what an opinion editor was in a newspaper," says Scott Lewis, Voice of San Diego's CEO. That person would aim to spark discussions. And expand discussions. And guide discussions. And frame discussions.

      That person would also curate the web -- no information overload, only filter failure -- to add depth and breadth to those  discussions. "I'm really sold on this idea of context as the future of news," Lewis told me.  "For so long we had this idea, from  newspapers, that you put a story up for 24 hours, and it did what it needed to do, and then you moved on. Now, though, we're  engaging differently with our news -- and are more in need than ever of people to act as stewards of engagement. That's where this
      new editor will step in.

_________________________

Rob Weir, (WeirR@missouri.edu) Columbia Missourian:  

1. We do not use anonymous commenting, though it can be difficult to verify real names. We do remove comments that don't comply if the user does not respond  to a request to verify his/her name within a reasonable amount of time (usually a day or two). Our editors think that our policy on commenting tracks with  our reluctance, in most cases, to use anonymous sources.   2. No, we don't have anything about that because our default assumption is that their identity is public  3. Yep, it's manual.  

 Jeb Bladine, McMinville,  Oregon, News Register Publishing jbladine@newsregister.com  

1. Comments allowed by registered users -- we have only their user name and e-mail address, and can see an IP address. Wedon't pre-authorize, but we do eliminate those that violate published "Policies and Standards." 2. We explicitly go the other direction, warning posters that they are not protected from inquiry by law enforcement. We say we'll general protect identity,  but not promised. We don't consider these anonymous posters to be "sources" in the sense of us trying to provide protection under our state shield law. 3. Comment function is built into our CMS, which comes from New Jersey-based DPC.  

Howard Saltz, MediaNews Group (HSaltz@medianewsgroup.com) : 

 For whatever it’s worth, commenting is a huge issue for many of our editors. I probably get more questions or concerns about it than any other initiative. 1. In MediaNews Group the decisions about removing comments are made at the local level. Generally, our editors will remove posts that are racist, obscene,  copyright violations, libelous, invade privacy, are spam, or are threatening.   2.  I can’t answer because our sites are too different, and in too many different states. 3.  The majority of our sites use the vendor Topix. Its automation includes: a.       A filter for obscene words and spam. The database of obscene words can be adjusted for each site, because of regional slang. Of course, someone can  create a unique spelling to communicate an obscenity and get past the word filter, but the filter helps somewhat.  b.      Every comment includes a link with which the public can report abuse. When abuse is reported, an email is automatically sent to a pre-determined set  of editors at the local newsroom. The editor(s) then decide whether to remove or allow the post. 

 Art Campbell ( art.campbell@gmail.com ) : 

 1. No anonymous comments; people must register, but they are free to mask his or her name -- essentially the identity may be masked, but there's an email  address that was at least valid at the time they registered as a user. What is "pre-moderate?" We only do one pass, which we call moderation. 3. We use WordPress with plug-ins for comment editing.  

Barr Parr, http://www.coastsider.com   (barry@parr.org)  

1. We don't allow anonymous comments.  All users' identities are subject to verification. Until verification, users' posts are pre-moderated before release. 

2. This doesn't apply to me directly, but for the record, it take an extreme case for me to ignore a valid subpoena. My attitude is that you shouldn't post  anything online expecting to keep your identity secret, unless you take serious measures to do so yourself. If you're not protecting your IP address and/or  email address, I don't see how you can expect me to do it for you. But I might go the extra yard to protect a clueless whistleblower from persecution. 3. I review all messages personally, but we're a small town and I know most of the posters personally.  

Steve Hanson, CruisKeen Consulting (shanson@cruiskeenconsulting.com ) :  

1) He allows anonymous comments, he does pre-moderate and he removes comments if they contain "libelous statements, advertising, or total nonsense in the post. Most of the time it's because they're blatant advertising spam." 

2) Yes, they do make privacy promise. What if you thought the legal process was SLAPP-like or chilling to a whistleblower? That would depend --- though we're not really in a position to take on major financial and legal issues. State shield law "almost certainly would not" protect.  3) Partly manual process. Registered users can comment without moderation until proven otherwise.  Non-registered users get moderated by hand.  However, there are 2 different layers of automated filtering ahead of that, which kicks out most of the spam etc. For technology,  I use http://mollum.com  a Drupal spam module. http://drupal.org/project/spam   

Gail Robinson of the Gotham Gazette (grobinson@gothamgazette.com ) :  

1. Yes. We don't premoderate but we do monitor comments and remove ones that are SPAM, would be libelous if those rules applied or blatantly racist, sexist or otherwise offensive. Our policy, though, is to err on the side of allowing comments. 2. WE should have a policy like this but don't and I;m not sure what NY's law is on this. I should probably find out. 3. We have two separate comment systems. On our articles in Gotham Gazette we use Disqus. Disqus automatically sends notification of all comments to me and I delete those that need to come down. In general, we have been happy with Disqus because it's user friendly (no registration required), and we want to encourage comments and discussion. It has greatly increased the number of comments on our side. Our Wonkster blog is on Wordpress. We do not get any notification of comments but I try to go to the dashboard regularly to check whether new comments have come in and read those that have. In general, we get more comments on our Gotham Gazette articles than on the Wonkster.  

Robert Brown  Swift Communications (rbrown@swiftcom.com) :  

1. We do not verify identities of our online community members, but we do require them to create an online identity. Most users choose to anonymize this identity and it includes very little information that would identify them in the offline world. So, our anonymous members do have online identities, but they are not necessarily representative of true identities and never are verified unless there is concern over someone acting as an impostor. By default, new community members are allowed to post without pre-moderation. If our site moderators identify a troublesome user, mainly a user who breaks terms of use more than a few times, the ability to post unmoderated comments is removed. Generally, we treat this pre-moderated tier as a temporary probation, a penalty box of sorts. It's a very visible thing to the user and they tend to not like it. In almost all situations, though, we try to contact the user before we use the pre-moderation settings on their account, via either online direct messaging through the site or e-mail addresses provided during registration. Directly contacting troublesome users has been fairly successful in toning down a heated conversation and avoids the need to use harsher tactics. We have a policy of removing comments only if they violate our site's Terms of Use. Our TOU is well written and comprehensive and is a good standard for site operation. We may also remove comments that were in response to removed comments, posted before the initial comment was removed. A response to a non-existent post is neither good for the reader nor fair to the responder. 2.  Our site's privacy policy states that we will only share information if it is necessary to comply with legal requirements. We have an internal policy to pass any request for site members' information first through a Swift Communications executive level officer who will consult with legal council on how to proceed. 3. We have partnered with Pluck, a Demand Media company, to provide a very actively manageable online community. Much of the online community is automated, with minimal staff hours required to moderate, unless there is a very contentious issue being discussed. We also give the online community some moderation roles, both in reporting abuse, and if the 'crowd' thinks something is abusive, a contribution will be temporarily hidden if it reaches a certain threshold of abuses. Pluck was chosen in a fairly thorough vendor selection method used in Swift Communications. Pluck was the best match for our publishing group's pre-determined requirements and have remained an active member in moderating our online communities as efficiently as possible. http://www.demandmedia.com/products/pluck/  

Aldon Hynes, a Connecticut community blogger (Aldon.Hynes@Orient-Lodge.com) :

My general approach to comments is similar to that of comments at a town meeting.  If you want to make a comment, you need to stand up and identify yourself. 1. I run a bunch of different sites, and the rules vary by site.  For a more blog oriented site, I often do allow anoymous comments.  I do not pre-moderate, but I do use various forms of software that filters out spam. Spam can be a real big problem. 2. I do not believe that I've ever promised to not to reveal the identity of anonymous commentators on my websites.  Personally, I have little use for people who are not willing to put their name to their comments. In terms of whistleblowing, I provide opportunities for people contact me directly, without putting up a content.  I don't think whistleblowers should use anonymous comments to raise an issue.  That just seems too dangerous. It makes it both more difficult to check the source as well as more difficult to protect the source. 3. I've also used Mollum for Drupal sites and have been pleased with it.  I also like to use Disqus, http://www.disqus.com for comments on more blog orient sites, since they handle spam fairlly well, and add a social networking component to the comments.  I believe that this social networking component also helps increase the quality of comments.  

Tammy Daniels, from iBerkshires.com / Boxcar Media, North Adams, Mass.,  comments (tdaniels@iberkshires.com) :

1. Yes, we allow anonymous comments and, yes, we premoderate. It wasn't much of a problem when I started but our profile has risen sharply, as has "toxic Topix."   I delete out of hand comments that are obviously libelous, slanderous and just plain nasty:  "Mary Smith can't do her job because she's too busy snorting coke."  "Jim Bob is a fat slob." Sometimes I delete parts of comments, indicated where the editing took place and then explain why: "We don't use that language." I have left comments that are borderline and explained why I won't allow anymore. I have also shut down rants on such hot topics as immigration and abortion. They're usually trolls that are just repeating the same thing.

I have given greater latitude to criticism of higher officials. The reasoning is that 1) we don't want to become toxic Topix in which namecalling and bullying are the norm and 2) we want to encourage constructive, civil dialogue.   If someone comments that a legislator is lousy, I'll immediate ask them why. We're trying to encourage not just opinions but facts. (And because it's our Web site; don't like it, get your own Web site. :)

2. No. We don't have a spelled out policy (though we should and it's on my to-do list) other than we give members who have to provide addresses the option to opt out of receiving e-mail from third parties affiliated with us, such as our advertisers. The comment section is in no way linked to our membership; members get to post events, class reunions, weddings, etc., all of which also must be approved.

No shield law in Massachusetts.

3. Pretty much everything at Boxcar is homegrown -- I can't imagine we would ever use a outside content manager. Everything is manual, although I have urged the use of memberships to allow instant posting with the threat of suspension or banishment for violating rules. 

 Joe Dwinell  senior city desk editor and web editor, Boston Heraldjdwinell@heraldinteractive.com 

We do allow anonymous comments. You must register with us first before you can leave a comment. This requires a real e-mail address and we can capture your IP address. We police comments by asking all readers to hit a "report abuse" button. We then review all comments that are reported as abusive. We can and do scan comments for any violations of our policy. That, as you can imagine, is more time consuming. I just know which stories to keep an eye on. We do our best to keep an open dialogue. We also reach out to those who leave a comment if it is newsworthy. So begins the newsgathering process with all its complicated issues - including anonymity.

From G. Patton (“Pat”) Hughes, Paulding.com publisher@paulding.com  

1. As a practical matter, I don't know everyone who posts on the site and most have handle-type email addresses and member names that would establish anonymity.  I don't pre-moderate but when things get out of hand, when comments are getting personal, when personal identities are disclosed in a manner that might be libelous for the member posting, I, or a group of moderator volunteers, do set those comments invisible.   Obviously, one reason we set most comments invisible is because they have crossed the line that might, in general judgment subject themselves and by extension us, to litigation.  (We set comments that are demeaning, racially over the top and or personal attacks on others invisible.  We've also been known to set legitimate news invisible as when a next door neighbor identifies a teen-ager who died in an automobile accident by name before the authorities have had a chance to notify next of kin.)

I have set up a forum where guests were informed how to use an anonymizer for their IP and were allowed to post as a guest but the comments would be reviewed before publication.  I think the forum may yet be there but it was never really used by the public.

2. My privacy policy sets forth that I will not, as a general rule, disclose the identity of posters  unless or until ordered by a court of competent jurisdiction to disclose said personal identifying matter ...  I don't require full name or any identifying information other than a working email and a date/time/ip address sequence attached to each post.

My privacy policy does give the right to break that promise by acknowledging  that I also reserve the right to cooperate with authorities and even proactively expose what I know to authorities if I feel that a poster (who might be using the site to sell methamphetamine or breaking the law ) is endangering other members or making me an accomplice or party to a crime.

The only instance in which I invoked that was when a person in an adjoining county was publicly threatening suicide in which case I informed authorities in that county of that threat and the information I had that would identify the individual.

In contrast, I have not invoked my option in the case  of a political poster who happens to be a federal employee in blatant violation of the Hatch Act.  What I know from privacy-protected information is  this individual has made over 4,000 posts from his government work many of which were partisan political in nature.  Using a 10-minute per-post factor, this equates to about 16.5 weeks of work time (40/hour weeks) posting primarily partisan political and religious commentary.  This is in addition to posting from home and other locations which is legal for an employee of his grade.  I should add that the person disclosed his federal employment on the board himself. The only privacy policy information I'm privy to is the IP from which this member posts which is identified as a  federal work location and the actual number of posts associated with that account at the federal work location.

Regarding a SLAPP like inquiry, I would resist the subpoena. I have indeed resisted a subpoena for news data that is detailed on the Citizen Media Law Project site.  I did not disclose the information claiming coverage under Georgia's shield law.  The subpoena became moot when the civil suit was dropped before the judge rendered an order either confirming my coverage under the shield law or compelling the production of the information sought. (Basically the edited out-takes of video shot at a public meeting.)

3. Comments management is a manual process with some automation for reporting.  A central moderator panel lists all posts/topics that are reported by any member from a button appearing adjacent to every post.  A handful of volunteer moderators review these reports.  Typically, I act as the person to which a moderated member can appeal.  I have occasionally overridden moderation that I felt was too high handed.  Most moderation is done by moving topics and posts from a widely public forum to one where a member has to be registered to enter a password - typically these less public forums are optional and can be turned "ON" with "ON" being the password.