Comments & Platforms

Explanatory factors

Research question 2: What technological, psychological and social factors can explain the differences between comments on a news website and Facebook?

It is common to explain online behavior with anonymity. But other explanations will be suggested for the observed differences between comments on news articles on vg.no and its Facebook page. This research has undcovered four types of differences: Level of public engagement, meta data, level of discussion, and categorical differences. Technological, psychological and social factors have been suggested as possible explanations to these differences.

Public engagement

There are far more comments on VG’s Facebook page than on vg.no. And when also taking into account the ability to like, share and react to comments on Facebook, the number of interactions on Facebook is about 10 times higher than on vg.no. And each of these interactions increases spreadability of an article. On VG’s comment section, a commenter must actively choose to share comments on his or her Facebook page. Because this option is turned off by default, there are probably very few interactions with the comment sections on vg.no that increases the spreadability of an article. This may help explain the motivation for news sites to close comment sections and focus their attention to user interactions on Facebook, since higher spreadability means more user clicks and ad revenue. Even though combating uncivil comments and spam is the most cited reason for closing comment sections, news sites do not deny that higher user engagement on Facebook is a at least a part of their motivation – as was the case with Dagbladet (Ramnefjell 2016).

The reason for the higher number of comments and reactions on Facebook is partly technical. The design of Facebook as a social platform, where users are automatically presented with the activity of friends and contacts, makes any content posted there spreadable by default. When a Facebook user interacts with a post, he does not have to actively choose to share it for it to be visible for his friends – although he can, and it is generally implied that interactions on Facebook results in automatic sharing.

Another explanation for the higher number of comments on Facebook is the psychological concepts of affordance and cost. Because of the design of the platform, and users actively using it on a daily basis, articles on Facebook are easily available and easy to comment on – requiring very little from the user. On VG’s comment section the user has to go to vg.no, read an article, click on the box to show comments and allow commenting. This means that commenting on vg.no is a much costlier activity than on VG’s articles on Facebook.

Meta data

Quantitative analysis has found that the average number of words of comments on vg.no is about three times higher than on Facebook. This can be explained technologically with an input box design that discourages longer comments on Facebook when using mobile, and encourages longer comments on vg.no with a wider design. I have also speculated that users on Facebook feel more at home on the familiar platform, and that venturing to vg.no may feel like going into a public space. This would encourage commenters on vg.no to be more articulate, and emphasizing their opinions and argumentation more than they would among their friends “back home” on Facebook. Another explanation for longer number of words can be found in theories about social influence. If there is a general trend towards longer comments, this trend may be strengthened by new commenters conforming by writing longer comments themselves.

Another observed difference is the higher number of emoticons used when commenting on Facebook. Just as with number of words, this can be explained with social influence, and the users’ feelings of being at home when on Facebook making them more informal in their communication. The users on Facebook are also exposed to a lot of emoticons, and not only from other users. The article posts on Facebook display emoticons next to the react-buttons. Users on Facebook may be primed to use emoticons. Commenting on Facebook is also done for different motives than on vg.no. The informal environment and the use of emoticons suggests a more socio-emotional view on commenting, as opposed to a task-oriented.

Level of discussion

Debate, whether public or not, is not possible without some sort of interaction. In comment sections, interaction can be measured by looking at how many replies there are to previous comments. VG’s comment sections have more replies than their Facebook page. There are also longer strings of replies, which based on the qualitative analysis of them suggests more and longer conversations.

From a technical point of view, Facebook discourages conversational engagement. Users on Facebook can comment on an article without ever reading previous comments. Reading previous comments requires active clicking by the user, and reading replies to comments requires an extra click. Affordance and cost, and the user’s motivation, also influences the number of replies. If users on Facebook are motivated by a socio-emotional view, because of the low cost of commenting on Facebook, they are more likely to comment for the sake of sharing an opinion, not engaging in debate. commenters on vg.no are actively engaging in commenting, are more task-oriented, and are more likely to have the time and interest to engage in discussion. A higher level of discussion can also lead to more derogatory and uncivil commenting, as uncivility between commenters is dependent upon conversations and discussions.

Categorical differences

This research resulted in the creation of a coding system for comments with 12 categories. The system was developed through a heuristic process where the data shaped the categories, it has been tested and refined for a high reliability score, and it has worked sufficiently for this research. All categories, except those that were grammatically or contextually meaningless and labeled as arbitrary, fitted into the 12-category system. Using this system, it became possible to measure the full width of comments and compare differences between Facebook and vg.no.

On vg.no there were more questions, suggestions, informative, argumentative and derogatory comments. These are comments that I associate with higher interest and higher engagement with other commenters. In one way or another they represent a willingness to read previous comments, to engage with other commenters, and to expect responses. These are all signs that these commenters have the time and interest to invest in their commenting.

I have found that the higher number of informative, argumentative and derogatory comments can be partly explained by the user’s feeling of being in a public space, although derogatory comments can also be explained by introjection and perceived anonymity and invisibility. Social influence might also explain these differences, as the first commenters on vg.no are more likely to be genuinely interested in the topic, and more argumentative and informative.

On VG’s Facebook page, there are three categories of comments that are more common than on vg.no. Tagging comments are a way for a person on Facebook to direct the attention of someone else to an article, and is not found at all on vg.no. Reactive and supportive comments were found to be more frequent on Facebook. These comments are rhetorically similar, as supportive comments are often reactive and short, but are defensive or show empathy towards someone. Again, the cost of commenting can be used to explain the higher number of supportive and reactive comments. They are short, easy to write, and generally doesn’t invite many responses. They require little activity from the commenter, and cost very little in spent time. Reactive comments can be explained technologically through labelling, and the design and limitations of the input box. The labels and the size of the input box discourages longer comments. And there is also a competition with the much easier to use like / react-button.

References

To top ↑