Romania: Less hate, more speech

The project "Less hate, more speech" aims at stimulating more civilized discussion with less aggressive speech on the Romanian online media, as one of the authors explains

"Less hate, more speech", started in 2014. It is an ambitious attempt to study the online user behavior when commenting on important Romanian web sites and to understand what interaction stimulates a more constructive approach, while at the same time lessening the cases of hate speech and the necessity for moderating comments. We spoke to Roxana Bodea, one of the researchers working on the project.

What are the project's main objectives?
The main objective is to determine under what conditions those who comment online have a more tolerant attitude towards other users and towards the moderators. Both researchers and journalists have been involved in the project. The journalists' objective is mainly to understand how to maintain the commentators' interest, at the same time lessening the need for moderating, thanks to a lower level of hate speech and intolerance within the site. It was interesting for the researchers to study the conditions necessary to produce this type of result.

How many and which media did you involve in the project?
We began with two web outlets: Gazeta Sporturilor, the online version of the only daily sports newspaper available in Romania. We then added the blog Tolo.ro, produced by Catalin Tolontan, editor in chief of the Gazeta Sporturilor but also one of the most respected investigative journalists in Romania. In fact his blog has a different target from the daily, because mainly in the last year and a half it has concentrated on social investigations, for example, health issues.
At the beginning of 2016 we added two minor websites: Blogsport.ro, mainly focused on comments and analysis on sport, and Paginademedia.ro dedicated to the community of media and marketing. We had already worked with Gazeta Sporturilor in the past and that's why, when in 2014 the project was started, we decided to include them. We then also included sites which spoke about media because journalists were the target group with whom we were interested in communicating and interacting.

Most of the portals involved deal with sports…

Certainly, the sports theme caught our interest in particular, both because of its widespread interest (Gazeta Sporturilor has about 3.5 million accesses alone a month) and for the strong passion which often risks turning into aggressive behavior from fans.

How did you establish the work program for the project?

Once we had established the work group which includes the University of Bergen (Norway), the Central European University of Budapest (Hungary) and the media partners, we developed both the purely theoretical part and a work group to develop the management of the comments.

The management strategy was the result of a collaborative effort between the subjects of the project: we hired five moderator journalists, who together with the rest of the team developed a set of rules for the comments management. In a second phase we gathered all the data collected over a 14-month period in order to rework it from a theoretical point of view. To develop a coherent strategy, we met weekly to discuss with the moderators concrete cases they encountered daily. By working on these real cases we have drawn up a series of guidelines to be used in moderating comments.

What type of strategy have you applied for the moderation?

We naturally decided to moderate hate-speech, aggressiveness and incitement to violence, but without censuring the entertainment, or at least the passion of the comments, unless these go over what we consider to be the minimum threshold of respect and tolerance towards other users. At the same time it was our idea to moderate the comment and not the commentator. In other words, never to interpret what the commentator had in mind, but instead to focus on what was actually said, taking into account the context.

What was the readers’ reaction to this new strategy?

Analysis of the data is still ongoing and it is too early to attempt to better understand the effect our efforts have had. Certainly some elements are already visible. For example, in one experiment in which we sent alerts to the users when a comment would have been subjected to moderating, often the author himself would voluntarily return to correct it.

What other experiments for the management of comments have been used?

Among the experimental methods used we have inserted stars in the place of offensive comments or else we have published them anyway but used a very small font. At the same time the moderator has highlighted in another color those comments which, for one reason or another, are considered particularly positive.

Which of these strategies has been seen to have the best results?

It is difficult to single out or identify one strategy which is better than another. What we have been able to see is that if the readers are involved through a system of moderating in which the rules are known to both sides, there is a notable reduction in the number of comments that require modifying. Perhaps then the key is to involve the readers and to show respect both on one side and on the other.

Is verbal intolerance a widespread problem in the Romanian media?

Certainly, with respect to the comments there is a strong feeling of concern in the journalistic community in Romania. According to information we have collected, 84% of journalists we approached consider that the comment section is mainly dominated by intolerance and verbal violence. However the data collected during the project seems to at least partially disprove this perception. For example: on the Gazeta Sporturilor, between May 2015 and May 2016, about 21% of comments needed to be moderated, while on other sites the percentage was even lower (on Tolo.ro about 14%).

Are there ethnic or political groups particularly subject to verbal aggression on the internet?

Predictably we have noticed that some groups, as for example, the Rom or the Hungarian community in Romania, are often victims of verbal attacks in the comments online. At the same time we noticed that the moderators had different sensibilities towards different groups. For the Rom and the Hungarians, often subjects of intolerance, there was a greater response with the moderators than for other groups less typically attacked, as, for example, the Asians, towards whom evidently there was less sensitivity.

How to define the subtle red line between moderating and censorship?

For the project we started by studying the rules already applied in a series of international online media, searching and underlining all common features, then adding our own consideration for our local area.

Naturally there is always a risk that the moderation can transfer into open censorship. However I believe that with our policy of moderation we have been able to find a sustainable compromise between the right of free expression of thought and that of being protected against aggression and discrimination, excluding openly racist comments or those which incite to violence, without for this arriving at a rigid and sensitive approach to moderating.