French Journal for Media Research

Laura Calabrese et Jérémy Jenard

Talking about News. A Comparison of readers’ comments on Facebook and news websites


Cette étude de cas aborde la question de la consommation de l’information dans deux environnements sociotechniques différents (sites d’information et Facebook) et la manière dont s’y façonnent les pratiques de commentaires de l’actualité en ligne. Une analyse quantitative et qualitative de commentaires confirmera que Facebook encourage les internautes à se concentrer davantage sur les interactions entre usagers que sur l’actualité.


This case study addresses the question of media consumption in two different online environments (news websites and Facebook) and how everyday commenting practices are shaped by these environments. Users’ comments were quantitatively and qualitatively analysed to confirm that Facebook encourages users to focus on interactions rather than on the news per se.

Texte intégral

1This paper addresses the question of media consumption in two different online environments: news websites and Facebook. The latter has become an unavoidable partner for media companies struggling to find a profitable business model, and has provided Internet users with an extra platform to express their opinions, in an era of audience empowerment. But the arrival of Facebook to the online media economy did not put an end to the comments section in online newspapers. On the contrary, most news websites still keep their comments board1, which means that both devices are still considered useful by media editors, allowing them to attract a wider audience and to redirect the traffic to the online newspaper. Moreover, both devices have found their public: while the comments board on online newspapers is more often reserved to regular readers or even subscribers, their Facebook pages meet the demands of a larger public accustomed to commenting the news alongside other public and private stories.

2The goal of this research is to observe how the contemporary practice of commenting the news has been shaped differently in two different technical and discursive environments. As other scholars have pointed out, “Considering the importance of news comment sections as spaces shown to influence how audiences evaluate the news […], it is essential to examine whether distinct platforms would perform comparably in terms of news commenters’ perceptions and behaviour” (Kim et al. 2018, p. 2). In order to observe how these two virtual settings shape readers’ discourse and online behaviour, a case study was conducted, based upon a bilingual corpus of comments originating from two mainstream European media. Readers’ comments were analysed with a quantitative and a qualitative methodology, in order to prove the following hypothesis: contrary to news websites, that incite readers to bring their attention to current events, Facebook encourages users to focus on interactions rather than on the news per se. Should this hypothesis proved to be true, then Facebook, such as other social networks (SNS), would be less suitable than the website comments section to discuss or comment on current events. Far from being deterministic, this hypothesis does not imply that it is the technical features that automatically determine the news consumption, but the equation between the latter and readers’ usages. SNS are then seen not only as technical devices but as socio-technical devices, where the social and the technical dimensions completely intertwine (Proulx, Millerand & Rueff, 2010). If the technical features of each device afford a set of practices (Bucher & Helmond, 2018, p. 3), these have been adopted by the public, tracing a divide between two different ways of commenting the news. The concept of imagined affordances clearly expresses this malleability of digital practices, as “technologies and user’s perceptions of their characteristics” (Kim et al. 2018, p. 3) are constantly evolving. Following this line of reasoning, the usage might continue to evolve if, for instance, moderation becomes stricter on Facebook, or if the new algorithm of the social network makes news less visible.

Commenting on news online

3A few years before the arrival of social networks, when mainstream media opened up to readers, journalists considered comments as a “necessary evil” (Reich, 2011). With time, every traditional media outlet followed, as online audiences seemed to progressively take the web 2.0 under control. With the arrival of SNS, media companies have found a way to fight the loss of regular readers and promote their name. Nevertheless, this partnership poses too many challenges, such as a loss of legitimacy as well as of journalistic quality (Noblet & Cheynel, 2010; Costera Meijer, 2012). As an infomediary2, Facebook gathers and publishes news but does not produce them. In other words, it remains a communication media (Mercier, Ouakrat & Pignard-Cheynel, 2018) initially conceived to foster everyday sociability (Le Caroff, 2018), and not a media company built to bring news to people. Nevertheless, in recent years Facebook has occupied an ambiguous position in the media landscape, by stating that the platform is “not a traditional media company”3, and thus declining responsibility for the spread of content but taking advantage of being one of the first sources of information in many Western countries4.

4As many other SNS, Facebook is focused on the identity of the user, as “the narrative presentation of the Timeline gives each member page the look and feel of a magazine – a slick publication, with you as the protagonist” (Van Dijck, 2013, p. 55). Moreover, users are regularly “accidentally exposed” to news on Facebook; in this context, news content has become a “byproduct of other online activities” (Tewksbury, 2001). This means that people are not particularly looking to comment on current events or politics on Facebook, which as other SNS is mainly used to exchange with friends (Jouët & Rieffel, 2013; see also Kim et al, 2018), and thus presents a curious mix of news consumption and entertainment, where the latter prevails (Mercier et al., 2018). According to Cardon (2012), the mix of private and public speech on Facebook results in a peculiar form of enunciation, where current events are apprehended in a very personal and familiar manner. Rowe (2015a) further strengthened this assertion, putting forward Facebook users’ proneness to imagine that they are having a conversation with their relatives when commenting the news; whereas online news commenters tend to believe that their comments would have an influential power on decision-makers and/or editors and journalists.  This way of addressing public issues can be described as “contextual”: context-dependant, filled with emotions, assumptions and terms of address, whereas “distant speech” has the objectivity of public statements. These conclusions are consistent with the findings of Jouët & Le Caroff (2013), who studied forms of commenting in different platforms. According to the authors (who conducted a netnography as well as interviews with readers), Facebook encourages users who don’t belong to the traditional news media audience to comment, producing different commenting patterns than the traditional audience, which are “less legitimate” and deploy on a “conversational mode”. Hille & Baker also conducted a study comparing comments on Dutch online newspapers platforms and their Facebook page, to conclude that “commenters provide more elaborate comments on the news websites compared to the more personalised comments on Facebook” (2014, p. 569).

5The abovementioned studies support the conception of Facebook as a communication and not an information media, an ego-based ecosystem that fosters the construction of one’s digital persona and not discussion around news5. Van Dijck also shows how Facebook shifted from connectedness to connectivity, by focusing on the latter, encouraging users to produce and consume data in order to support its business model (Van Dijck, 2013). It can also be added that if SNS are one of the first sources of information, they are mostly accessed through mobile devices, which suggests a superficial and heterogeneous grasp of current events6. According to these findings, Facebook is not a natural environment to comment on the news, or at least a less legitimate one than newspapers platforms. It can help editors filter abusive comments, as anonymity is impossible on Facebook (contrary to pseudonymity, which is quite common) (Hille & Bakker, 2014, see also Orr, 2011), but the quality of comments seem to be poorer than in the online platform of newspapers. The specific question to be addressed in this case study is whether Facebook audiences focus on interaction rather than on the news. This research question is based on previous research showing that the comments board is used by media audiences, particularly regular readers with a subscription, to establish an imaginary dialogue with journalists, criticising their professional routines as well as their writing (Calabrese 2014a and 2016; Craft, Vos & Wolfgang, 2015). This paper’s goal is then not to find out whether SNS promote incivility or even hate speech, but to observe the specific enunciation of Facebook audiences.

The sociotechnical devices

6As underlined by Jouët & Le Caroff (2017), conducting an analysis of digital traces of interaction is “closely intertwined with the analysis of the sociotechnical device” in which said data can be found, echoing the need to “take into account the weight of sociotechnical devices in the framing of the interactions and comments” (Jeanne-Perrier, 2010, p. 127, our translation). The following table summarises how the two platforms (website and Facebook page) aim at enhancing users’ participation and interaction – be it with the content of the article or between them. The various criteria selected were based on a tailored version of López Díaz’s work (2015, p. 209-211). It is important to highlight that this chart is but a mirror of the state of things at the time this research was conducted (2017-2018) and that, given the dynamic nature of the web, the two platforms under scrutiny may face transformations in the future.


Facebook page


It is mandatory to sign up for both websites to comment.

Accessible to all providing that they have a personal Facebook account.


No explicit call-to-action.

“Write a comment”.

Quantification of interactions

Number of comments displayed on top of the page (header or near the lead paragraph).

Number of comments and likes/reactions displayed underneath the post.

Cross-channel dissemination

It is possible to share a comment on different social media platforms.

It is not possible to share a comment, neither on a personal Facebook profile, nor on a different social media platform.

Terms and Conditions

When a reader signs up he/she must explicitly specify that he/she complies with the newspaper’s Terms and Conditions.

When a user signs up he/she must explicitly specify that he/she complies with the Facebook’s Terms and Conditions.


Chronological order automatically pre-set with the possibility to rearrange the comments in a “conversation mode”.

Comments are classified according to Facebook’s EdgeRank7 relevance criteria (“top comments”). It is possible to rearrange them chronologically (most recent first).

Interactions with the comments

Replies, upvotes and downvotes, reporting.

Likes, reactions and replies. It is not possible to report an inappropriate comment. Facebook also notifies a commenter if another comment was left or if someone tagged him/her.

Editorial input

Clearly indicated a posteriori moderation.

Supposed moderation either by the page owner, community manager or by Facebook itself (blacklist or profanity filter).

Methodology and corpus

7The corpus is made up of comments related to two articles from two online newspapers: The Guardian (TG) and El País (EP). The motivation behind this choice is threefold. First, they both enjoy a large readership (TG is the third most-read online newspaper in the world, whose readership amounts to 150 million visitors per month; and EP is the daily national newspaper with the biggest circulation in Spain, with an average circulation of 391.816 copies sold every day). Second, both are considered quality newspapers, and third, both have a large and active community on their different social media and especially on their Facebook page, with respectively 7.8 million and 3.9 million likes, indicating that the content disseminated on this platform will reach a large audience, potentially generating a considerable engagement from their audience (direct audience and also reached through the repeated exposure8). The comments were extracted from two articles: “Feminism gave women a glimpse of a happy world. But for young girls that promise is being dashed”9 (TG) and « Machismo en la red: “me llaman feminazi” »10 (EP).

8Both articles address a common theme: feminism. However, in the framework of this paper it was deemed that, although interesting and worth delving further into, it was of little relevance to conduct an analysis on the production of spontaneous discourse on feminism. Articles dealing with feminism were selected for two reasons: first to collect a homogeneous corpus, and second because, as a societal matter, this subject constitutes a debate hub that mobilises online audiences and generates various vivid discussions. The choice of collecting a multilingual corpus was motivated by a comparative perspective, that would allow to highlight common reading practices across countries beyond the variations of discourses on feminism.

9In the present research, collecting comments from the newspapers websites and from their Facebook page proved more complex than expected, as data mining and extraction tools were of no use and had to be discarded from our methodology. With respect to the extraction of Facebook comments, it soon became clear that the Facebook Graph API Explorer tool was ineffective, since its newest update (v.2.12) made it impossible to automatically extract data from a page without the consent of its owners and its identification token – which was somewhat easily bypassed in the previous releases of the tool merely requiring to convert the obtained data in a specific software, such as OpenRefine or FacePager. This is why the comments were manually extracted from both sources. This process, although more tedious and time-consuming, allowed for a more thorough collection of 1.200 comments11:

The Guardian

El País




Facebook post



10In order to verify our hypothesis (according to which the Facebook communicative environment encourages readers to focus on interaction rather than on news), our sub-corpora were manually annotated and coded according to eight variables:

11(1) The engagement generated by a comment (reactions, up- and down-votes, reports, etc.).

12(2) The type of comment. Comments were grouped into two categories: Original posters (OP) and replies, as a means to highlight a potential structural hierarchy in the comments thread, as evidence of the formation of a conversation between users.

13(3) The discourse object and its relevance to the article’s main topic. Paskin (2010) poses that “rather than addressing the story being reported, many readers posted comments on other news within the paper or news reported in other sources, or, even worse, posted comments solely attacking other readers personally” (2010, p. 5). This variable seeks to assess whether this phenomenon is more prominent on Facebook.

14(4) The linguistic markers that directly or indirectly identified or pointed to another commenter or to the journalists, here referred to as “address markers”.

15(5) The linguistic markers that denoted commenter’s subjectivity (disjuncts, interjections, etc.).

16(6) Any direct quotation of either an excerpt from the article or from another commenter’s comment.

17(7) Traces of metadiscourse on journalistic discourse and comments on journalism routines (references to the journalist, to the editorial board, to the newspaper, etc.).

18(8) Traces of references to the communication environment (references to the moderator and to the newspaper’s moderation policy as well as to the platform itself – comments, threads, conversation, message, forum, etc.).

Results analysis

19Before introducing the results12, it is interesting to have a look at the engagement rates of the audience to see how they interact with the content. Engagement includes comments, reactions (likes, reactions, up- and down-votes, etc.) and shares on different platforms or media (personal Facebook profile, twitter, mail, etc.). The following chart displays the results for each newspaper and platform:

The Guardian

El País





















20It is clear that the number of reactions greatly exceeds the number of comments, as users can interact with the content through effortless means (liking, sharing) compared to writing a comment. If writing a comment does not mean that the user has read the article, it requires nevertheless a greater effort than just clicking. This is confirmed by the fact that Facebook ranks comments higher than weaker forms of engagement (Bucher, 2012, p. 1169).  

21In order to annotate the comments, they were divided into two categories, ‘original post’ (i.e. a comment that is not a reply to another one) and ‘replies’. It is easy to identify them in the way they are displayed on both platforms, as they draw a descending hierarchical structure – or conversation thread (as shown in fig. 1 and fig. 2).

Figure 1: Conversatiopn thread on EP's Facebook post


Figure 2: Conversation thread on TG's website


22The analysis of the corpus shows that interactions between users were far more numerous on Facebook for both newspapers. Indeed, 91.3% (n=273) and 96.3% (n=289) of comments on Facebook for TG and EP, respectively, were replies to comments, against 66% (n=198) and 57% (n=171) of the comments on the website. The high percentage of replies can be partly explained by the fact that Facebook implemented an automatic tag when a user replies to someone’s comment. Automatic citation is a technical feature that inserts a user’s name into someone else’s reply, thus fostering interaction between users. So far, these observations confirm the idea that the social network facilitates interaction much more than the newspapers platform, even if it is under the weak form of a like.

Discourse object

23Comments were then categorised according to their relevance to the topic of the article, resulting in a classification of on-topic and off-topic comments. The Guardian’s article addressed a report from Children’s society on the state-of-play in young girls’ mental health. The core message of both the report and the article was that mental health amongst young girls deteriorated over the past years, mainly because of harassment. El País’s article dealt with online feminism and anti-feminist/misogynistic discourses – mostly translated as insults towards women, had they explicitly mentioned they were feminists or not.

24The analysis of TG comments put forward a clear gap between the two platforms. Indeed, most Facebook comments were deemed off-topic as their discourse object often dealt with broader themes, sometimes at odds with those of the article, which was not the case for website comments. The reason behind this may lie in the Facebook post itself, as it was not explicitly referring to the article’s main contents and was but a small excerpt of it (direct quotation from a paragraph):

25Facebook post: “While last year’s report highlighted how girls were unhappy about their appearance, this year’s shows that they are also frightened. Getting followed by ‘a stranger’ is something girls fear. They talk of being scared by men blowing kisses at them in the street and being terrified at 13”.

26Title: “Feminism gave women a glimpse of a happy world. But for young girls that promise is being dashed”.

27The post seems to only focus on young women being terrified of men, and that their being unhappy with their looks is the result of men’s behaviour. Several comments focus on attacking feminism/feminists or defending men rather than on women’s mental health:

28(1) There are little boys that are being sexually assaulted by their female teachers but there is hardly any outrage by feminists or any other female group because as usual it is assumed by society that men are the only predators out there (Facebook)

29This kind of comment was also featured in the newspapers’ website, but interestingly, most of the times other readers tried to get back to the original topic, as shown in example (2). This furthers Rowe’s assertion that commenters on the newspaper’s website are more drawn to seek topicness and more “elaborate and balanced” responses (Rowe 2015b).

(2) – Let’s face it; it’s a terrifying world. Being a parent to girls brings all sorts of worry and I think it’s a really healthy thing to be discussing the issues and trying to find ways of making things safer for them (…)

- Being parents to boys also brings worries.  (Reply to (2))

-Yes, it does; but that wasn’t really the topic of the article, was it? I’m sure there are lots of issues that boys face that are also worthy of discussion. Was there any particular issue you wanted to raise? (Reply to the reply, Web)

30A particular feature of this corpus that needs further research is the way conversation invariably deviates from the original theme to tackle the definition of feminism. Although observed in both sub-corpora, Facebook commenters showed a heavy tendency to wander in that direction:

(3) Feminism is not what it used to be in the past. Present day feminism goes hand in hand with Marxism and feminists are so militant and extreme. (Facebook)

(4) I mean, some branches of feminism have been tied with socialism and Marxism for a long time, and both ideologies have benefited from it. Luxemburg, Kollontai, Eleanor Marx to name a few. (Facebook, reply to (3), Facebook)

31Although present in web comments off-topicness was less significant, as a majority of comments (OP and replies) were found to be related to the article’s themes with, in some cases, direct quotations from it – as most reported discourse found in the subcorpora originated from other users’ comments:

(5) In order to find a solution, you need to identify the problem. You have given a list of symptoms, but no specific causes. The conditions mentioned and the lack of hope do not help. “Sexual harassment kicks in at around 11” - by whom? In what form? “Concerted effort to turn back the clock” - by whom? ‘culture’ or ‘society’ is far too vague an answer. “There exists an ongoing assumption that women’s rights have been ‘won” - liberties have to be defended and in changing environment, reclaimed. If feminism is going to have successes in the 21st century, it will need to define its goals. (Web)

32On both platforms, several comments go well beyond the main topic of the article, dealing with a general criticism of society (anti-capitalism, criticism of the British conservative government, etc.):

(6) The Tory government is the root of all evil in today’s society. (Web)

(7) We are victims of society (conditioning, marketed, single-minded, unable to think). (Web)

33With regards to EP, the web conversation was most often related to the theme of the article and most comments were about the gap between women and men and how the former is –supposedly– trying to take over the latter’s place through feminism (8), the persisting inequalities between genders (9), all the different ways to define feminism, and how sexism and feminism are two diametrically opposed notions (Reply to (9)).

(8) Si el feminismo es la igualdad entre mujeres y hombres me parece perfecto! (Web) 13

(9) El machismo Mata, el feminismo No!! Dejad de confundir los términos, ignorantes! (Facebook) 14

El feminismo también puede matar...15 (Facebook, reply to (9))

34Most off-topic comments were found on Facebook, as it was the case for TG, and were directly targeting both feminism and feminists:

(10) Es lo mismo otra feminazi todos los extremos son malos anda que no hay feministas que no matan. Las mujeres tambien MATAN.16(Facebook)

(11) El feminismo mata a las mujeres, cuando infecta la mente de una tenemos un mamarracho lésbico y frustrado.17 (Facebook)

35These observations shed light to a clear fluctuation in comments’ pertaining to the themes tackled in the articles themselves. As highlighted before, conversations were most abundant on Facebook comments, even though more research is needed to explain this phenomenon.

Terms of address

36This category encompasses words or phrases used to address or refer to someone. The percentage of comments containing one or more terms of address was calculated and revealed that terms of address were more frequent on Facebook: 64.6% (n=194) of TG’s comments and 39.7% (n=119) of the comments left under EP’s post contained at least one term of address. Although most of these terms were mostly pronouns (2nd person singular), they also took the shape of direct interpellation – either by means of a user’s Facebook name or of their username on the website. As mentioned earlier, this results from the mobilisation of an automatic citation on both Facebook and the website; when replying to a comment, the username automatically appears and, when it comes to Facebook, this user will be notified that someone replied to their comment:

(12) So, Jonpurrtree [username], you don’t see where this comes from? You’re not looking, are you? Society has changed / is changing drastically in the last few decades and young women and men feel under enormous pressure – as never before – to LOOK GOOD. (…) (Web)

(13) Tatiana Vladi Happy for you, I’m as well never felt harassed getting compliments. It’s really dependence how those been told. !!! But some people just can’t understand simple interaction and taking it like sexual assault.  (Facebook)

37A third category of terms of address as identified as insults, mostly present on Facebook:

(14) Las feministas no matan, payaso.18 (Facebook)

(15) Why don’t you piss off, moron. (Facebook)

38The frequent usage of these linguistic markers in Facebook comments clearly shows that users interact much more on this platform than on the comments’ board, fostering conversation between them, whether it is under the form non-aggressive statements or violent confrontation. Interaction between users is further triggered through the use of direct quotations from other commenters – in fact, most traces of reported speech found in the corpus originate from other users’ contributions.

Subjectivity markers

39It was already highlighted that online environments allow both a distant and emotionless speech, closer to an expert enunciation, as well as familiar and emotional statements, that focuses on the individual stance of the commenter. Subjectivity markers encompass (1) emotion markers (interjections or emotion words), (2) evaluation markers where the user ascribes a value to what he/she was exposed to, and (3) modalisation, where the user expresses his/her opinion on the discourse’s veracity (Todorov 2006). The latter can, in our case, refer to the article or to another user’s comment. Examples of the three were found in our sample:

(16) Emotion: Jajajaja si como no. Si el feminismo es igual machismo, es puro odio excusado en un discurso muy manipulado y engañoso de igualdad.19  (Facebook)

Emotion: uhhh...probably beside the point here but why only show what women and girls go through? Men and boys have received sexism and hate as well [...] (Facebook)

(17) Evaluation: I agree, I was arguing for feminism in the 90’s and you wouldn’t believe some of the conversations I had to have but I learned pretty quickly that most men who didn’t like feminism just didn’t understand what all the fuss was about (…) (Web)

Evaluation: Es oficialmente un panfleto feminisfa...20 (Web).

(18) Modalisation: [...] Lo siento, si no viene a cuento, pero una persona que no respeta que otra no tenga porque aguantar, ni escuchar, ni gustarle las galanterias de un desconocido, pues la verdad no me parece nada respetuoso.21 (Web)

Modalisation: “On the whole, girls are getting more miserable while boys are not” So there is a bright side at least. On a more serious note, maybe ya’ll made the mistake of thinking “more opportunities = more happiness” or “more freedom = more happiness”. From what I’ve heard, women are less happy now than they were before the Women's Rights Movement won all sorts of important victories for women. That does not mean feminism was in any way wrong, or that women's rights is a bad thing - obviously equality is a good thing - it's simply no guarantee of happiness. [...] (Facebook)

40The analysis of the corpus highlighted that the commenters’ relation with the information was often of an emotional nature. This emotional relation to information is, as defined by Pynson (2011, p. 68, our translation), “characterised by very subjective, spontaneous reactions that may stray from the article’s core theme”. This being said, a few comments were more objective, providing additional information and sourcing their arguments, as shown in the following examples:

(19) RAE22: 1. Feminismo, m. Ideología que defiende que las mujeres deben tener los mismos derechos que los hombres.23 (Web)

(20) Thanks, I appreciate your question. One thing we could address is the influence of social media. A lot of parents seem to be concerned about its effect on their children’s development. See: We could also give more thought to our relationships in general: (Web)

41These examples illustrate the fluctuation of specialised and non-specialised discourse in readers’ comments (Calabrese 2014b), as well as the mix of private and public speech, pointed out by Cardon (2012).

42Although numerous in both platforms, the analysis of markers of subjectivity did not allow for a tangible quantitative analysis. The only obvious difference between the two platforms is that interjections are overrepresented on Facebook and emotions can also be conveyed through emojis, showing that the technical constraints of the social network facilitate exchanges under the form of conversation.

Metadiscourse on journalism

43Regular commenters on newspapers comments platforms usually talk about journalistic routines, mentioning the newspaper, its editorial choices and guidelines, journalists and so on. Even if the lack of data made it impossible to yield conclusive results and calls for further research and for a more comprehensive corpus, it is interesting to point out that results were different for TG and EP. Indeed, in the British newspaper references to the journalistic practice were more frequent in website comments (12% (n=36)) than on Facebook (2.67% (n=8)), confirming previous research on the online behaviour of newspapers’ readership (Craft, Vos & Wolfgang, 2015; Calabrese, 2016). These references feature, for the most part, users appealing to the journalist (Suzanne Moore) as though they wanted to engage in a conversation with her. We observed that these mentions of the journalist were either implicit (use of pronouns) or explicit (Suzanne Moore, Moore, Suzanne, “the author”, “the writer”, etc.).

(21) Tories, Suzanne… Stop enabling them, m’kay? (Web)

(22) She [the journalist] didn’t completely dismiss it. Disagree with her [with the journalist] (I did), but at least recognise the arguments made, rather than suggest she [the journalist]’s making a point she[the journalist]’s not. (Web)

44We have also observed several criticisms of TG and of its editorial standards (be it as a whole or solely for this article):

(23) And who makes good money by pounding 'you are not safe, evil men will get you' message into heads of girls and women every day? Yes, it is Suzanne and her friends from Guardian. (Web)

(24) The Guardian could do its little bit by promoting intelligent feminists instead of ridiculous “empowerment” figures like Beyoncé and Swift. (…) (Web)

45Concerning EP, only 4.73% (n=14) of website comments and 5.33% (n=16) of Facebook comments contained a reference to the journalistic routine. One could thus infer that different reading practices exist across countries, and that EP commenters are more prone to comment on the article’s content or interact with other users.

References to the communicational environment

46Users’ apprehension and understanding of the sociotechnical devices was at the core of this research, as means to assess the devices’ potential impact on their discourse. The quantitative analysis of references to the communicational environment comprised all the comments mentioning other comments, commenters, the forum of discussion, the thread of comments, moderation policies, and moderators. This led to contrastive results, with a clear divide between the two newspapers under scrutiny. On the one hand, most TG comments referring to the sociotechnical environment were found on Facebook (11% (n=33)), whereas only 5% (n=15) of website comments did. On the other hand, an equal presence of references to the sociotechnical environment was observed on both comment modules for EP (10% (n=30) (Facebook) and 10.6% (n=32) (Web)).

47The data showcased an additional divide. First, TG comments mentioning the sociotechnical environment were mostly relating to comments and to the conversation that was being built up:

(25) The only comment with any insight in this morass of a page! I second - good comment! (Web)

(26) Funny, I got what she said. It really does not take much to read between the lines. She saying a lot of feminists chose not to have kiddies. (Web)

(27) This thread is what happens when you get your definition of feminism from someone who really fucking hates feminists. (Facebook)

(28) (…) Some people commenting on this article don’t seem to understand that. (Facebook)

48Second, in most EP comments referring to the sociotechnical environment, users question the newspaper’s moderation policy and directly call the moderator out, as shown in the following examples:

(29) A mi me censuran comentarios sin insultar a nadie, ayer me los borraron prácticamente todos en el foro en un tema de estos, hoy llevo ya uno. Con este van dos. Han publicado en estos foros mensajes donde me han llamado maltratador. Pero de eso no se habla. No te dejan. Menos mal tenemos a Twitter para denunciar estos casos.24  (Web)

(30) Moderador, por qué no publicas mis comentarios si cumplen con las normas de participación?25 (Web)

(31) Yo a veces, para ahorrarle trabajo al moderador, escribo directamente ‘mensaje borrado’, y así se ahorra el tener que censurarme.26 (Web)

49These results show that moderation, as we know it nowadays, forces the production of metacommunicational speech, and references to moderation can be explained as part of the imaginary dialogue that takes place since the beginning of digital journalism (Calabrese, Domingo & Pereyra, 2015). On the other hand, as Jouët & Le Caroff (2013) suggested, stricter supervision may have led to a shift of where debates around the news are held, as conversations seem to now be mostly engaged on social media. If the collected data does not seem relevant in the frame of this study, as it does not shed light on the way users behave in both platforms, it is still interesting to notice that metadiscourse is all in all a good predictor of interactivity between users, as shown in examples 25 to 28.


50The goal of this article was to prove that the sociotechnical environment of Facebook fosters interactivity between users rather than discussion on current events. The variables that were proven more useful to confirm this hypothesis were on- and off-topic arguments, original post/reply and terms of address, as they revealed a clear divide in reading and writing practices among users.  Further research and analyses on a larger corpus needs conducting, as this case study could not yield conclusive results on the other variables, as some results might explain cultural practices rather than generalised online routines (especially press criticism and the way to address journalists). But if the comparative perspective demands caution in the interpretation of the results, it has the advantage to underline which variables are relevant and which are not. As for the subjectivity markers, the inconclusive results reflect the fact that online speech is largely based on opinion, even if some significant differences exist, such as the presence of insults on Facebook, and the presence of much more emotion markers on Facebook under the form of buttons and emojis, which can be explained by the technical architecture of Facebook.

51The results of the analysis grid show that the sociotechnical environment of Facebook distinctly fosters interactivity, generating multiple conversation threads where terms of address are overrepresented. The more people engage in a conversation, the more they move away from the original topic, which is less common in the digital newspaper comments board. In this corpus, most off-topic comments targeted feminism and feminists, and displayed a good amount of name-calling from every side. To conclude, the results confirm that the communicative dimension prevails on Facebook, whose main features promote interactivity and offer the user the full gamut of possibilities, ranging from effortless clicking to more elaborate commenting and replying. If the sociotechnical environment of the social network does not hinder rational and subjective discourse, results show that it is more difficult for a Facebook commenter to stay on-topic, while displaying their opinion under the form of a fragmented conversation with other users.

52Comments are certainly the most common form of user-generated content nowadays, because the social web is designed to promote and circulate this content in order to foster interactivity, and to produce reusable data about online consumption. But this does not mean that there is one way of commenting; on the contrary, the overall communication environment of a platform (its technical constraints and audience usage in the global ecosystem of the web) affords different enunciative features.

5323  Feminism : Ideology that defends that women should have the same rights as men.


Blanchard, G., Gadras S. & Wojcik, S. (2017). Analyser la participation politique en ligne: des traces numériques aux pratiques sociales. In C. Barats (Dir.), Manuel d’analyse du Web (p. 176-195). Paris: Armand Colin.

Bucher, T. & Helmond, A. (2018). The affordances of social media platforms. In J. Burgess, T. Poell & A. Marwick (Eds..), The SAGE Handbook of Social Media (p. 233-255). London: SAGE.

 Bucher, T. (2012). Want to be on the top? Algorithmic power and the threat of invisibility on Facebook. New Media & Society, 14 (7), 1164-1180.

Calabrese, L. (2014a). Paroles de lecteurs : un objet de recherche hybride en sciences du langage. In A. Nowakowska and F. Perea (dir.), Écritures et genres numériques, Studii de lingvistică, 4, 13-27.

Calabrese, L. (2014b). Rectifier le discours d’information médiatique. Quelle légitimité pour le discours profane dans la presse d’information en ligne ? Les Carnets du Cediscor, 12, 21-34.

Calabrese, L., Domingo, D. & Pereira, F. (2015). Overcoming the Normative Frustrations of Audience Participation Research. Sur le journalisme, 4(2), 4-8.

Calabrese, L. (2016). Le discours (très) prescriptif des internautes sur le journalisme et les journalistes. Une étude des commentaires en ligne, Informer avec internet. Reprises et métamorphoses de l’information (133-147). Presses universitaires de Franche-Comté.

Cardon, D. (2012). Le parler privé-public des réseaux sociaux d’Internet. In S. Proulx, M. Millette & L. Heaton (Eds.), Médias sociaux. Enjeux pour la communication (p. 33-45). Presses de l’Université du Québec.

Costera Meijer, I. (2012). Valuable journalism: A search for quality from the vantage point of the user. Journalism, 14 (6), 754-770.

Craft, S., Vos T. & Wolfgang D. (2015). Reader comments as press criticism: Implications for the journalistic field. Journalism, 17 (6), 677-693.

Hille, S. & Bakker, P. (2014). Engaging the Social News User. Journalism Practice, 8 (5), 563-572.

Jeanne-Perrier, V. (2010). Parler de la télévision sur Twitter: Une ‘réception’ oblique à partir d’une ‘conservation’ médiatique. Communication & Langage, 166, 127-147.

Jouët, J., & Le Carroff, C. (2013). L’actualité politique et la participation en ligne. In J. Jouët & R. Rieffel (dir.), S’informer à l’ère numérique (p. 117-157). Rennes: Presses universitaires de Rennes.

Jouët, J., & Le Carroff, C. (2017). L’observation ethnographique en ligne. In C. Barats (Ed.), Manuel d’analyse du web en sciences humaines et sociales (2nd ed.) (158-171). Paris: Armand-Colin.

Kim, J., Lewis, S. C., & Watson, B. R. (2018). The Imagined Audience for and Perceived Quality of News Comments: Exploring the Perceptions of Commenters on News Sites and on Facebook. Social Media + Society, 4 (1), 1-12.

 Le Caroff, C. (2018). Le partage de l’actualité politique sur les proils personnels de Facebook. In A. Mercier & N. Pignard-Cheynel (Eds.), Commenter et partager l’actualité sur Twitter et Facebook (199-225). Fondation Maison des sciences de l’homme.

Lopez Díaz, R. A. (2014). Les discours en interaction de la presse en ligne. Propositions pour une théorisation de la coproduction journalistique. Thèse de doctorat: Sciences de l’information et de la communication. Université de la Sorbonne nouvelle - Paris III. Français.

Mercier A., Ouakrat, A. & Pignard-Cheynel N. (2018). Facebook pour s’informer ? Actualité et usages de la plateforme par les jeunes. In A. Mercier & N. Pignard-Cheynel (Eds.), Commenter et partager l’actualité sur Twitter et Facebook, (169-197). Fondation Maison des sciences de l’homme.

Noblet, A. & Pignard-Cheynel, N. (2010). L’encadrement des contributions ‘amateurs’ au sein des sites d’information : entre impératifs participatifs et exigences journalistiques. In F. Millerand, S. Proulx & J. Rueff (Dir.), Web social. Mutation de la communication (265-282). Presses de l’université du Québec.

Orr J. (2011). Testing a New System for Online Comments, Los Angeles Times blog, March 15.

Paskin, D. (2010). Say what? Journal of International Communication, 16 (2), 67-83.

Proulx, S., Millerand, F. & Ruff, J. (Dir.) (2010). Web social. Mutation de la communication, Presses de l’Université du Québec.

Pynson, L. (2011). La presse à l’épreuve du web participatif : quand la conversation ne va pas de soi. Communication & langages, 3 (169), 63-75.

Rebillard, F. & Smyrnaios N. (2010). Les infomédiaires, au cœur de la filière de l’information en ligne. Les cas de google, wikio et paperblog. Réseaux, 160-161, 163-194.

Reich, Z. (2011). User Comments. The Transformation of Participatory Space. Participatory Journalism: Guarding Open Gates at Online Newspapers, Online Library.

Rowe, I. (2015a). Civility 2.0: A comparative analysis of incivility in online political discussion. Information, Communication & Society, 18, 121-138.

Rowe, I. (2015b). Deliberation 2.0: Comparing the deliberative quality of online news user comments across platforms. Journal of Broadcasting & Electronic Media, 59, 539-555.

Tewksbury, D. et al. (2001). Accidentally informed: Incidental news exposure on the World Wide Web. Journalism & Mass Communication Quarterly, 78 (3), 533-554.

Todorov, T. (2006). Estilo. In O. Ducrot & T. Todorov (Eds.), Diccionario enciclopédico de las ciencias del lenguaje (344-348). México DF: Siglo XXI.

Van Dijck, J. (2013). The Culture of Connectivity. A Critical History of social Media, Oxford University Press.


1  Even though some, such as the Belgian weekly Le Vif/L’Express, have put an end to it given the aggressive nature of comments. In addition, some news organisations, as Kim, Lewis & Watson have highlighted, “have begun shifting commenting from their websites to Facebook, based on the implicit assumption that commenting on Facebook is an equivalent (or preferred substitute)” (Kim, Lewis & Watson, 2018, p. 1)

2  The role of infomediaries is to match a massive news offer with a specific target, by means of algorithms (Rebillard & Smyrnaios, 2010).


4  Pew Institute of Research.

5  Similarly, some authors claim that SNS do not suit political participation but on the contrary, foster slacktivism (Blanchard, Gadras & Wojcik, 2017).


7  EdgeRank is the name of Facebook’s algorithm that organises the display of content on a user’s timeline. EdgeRank also plays a role in attributing a “pertinence” ranking in comments.

8  Repeated exposure refers to the diffusion of a Facebook post or content by means of sponsoring (that is, paying Facebook to automatically disseminate said post on the timelines of users that meet pre-fixed audience segmentation criteria (age, gender, interests, etc.)) to increase its reach amongst users.

9  Moore, S. (2017, August 30). Feminism gave women a glimpse of a happy world. But for young girls that promise is being dashed. The Guardian. Retrieved September 11, 2017, from

10  Martínez, V., & Casal, J. (2017, July 27). Machismo en la red: « me llaman feminazi". El País. Retrieved September 11, 2017, from

11  More comments were left under both articles (on Facebook and on the website); however, given the repetitive content of these comments, it was decided to analyse the first 300 comments in each platform.

12  Note that translations of Spanish examples will be provided in a footnote.

13  If feminism means the equality between women and men, it is perfect to me !

14  Male chauvinism kills, feminism does not!! Stop confusing both terms, idiots!

15  Feminism can kill, too...

17  Feminism kills women, when it infects the mind of one of them we have a lesbian, frustrated freak.

18  Feminists do not kill, you clown.

19   Hahaha yes, but no. Yes, feminism and misogyny are the same, it is sheer hate hidden in a very manipulated and deceptive discourse on equality.

20  It [the article] is officially a feminist pamphlet...

21  Sorry if this is off-topic, but someone who doesn’t respect the fact that a person doesn't have to tolerate, or listen to, or like flattering comments from a stranger, is not respectful at all.

22  RAE: Reference to the Dictionary of the Royal Spanish Academy.

24   They censor me and I have not insulted anyone. Yesterday, they deleted almost all my comments on a forum [of comments] similar to this one [reference to the theme of the article: feminism]. Today, two of my comments were deleted, too. They [other commenters] have called me an “abuser”, but they [the moderators] do not care. Fortunately we have Twitter to report these cases.

25  Moderator, why don’t you publish my comments if they meet the publication critieria ?

26  Sometines, to save the moderator’s time, I write “Deleted message”, so that they don’t have to censor me.

Pour citer ce document

Laura Calabrese et Jérémy Jenard, «Talking about News. A Comparison of readers’ comments on Facebook and news websites», French Journal for Media Research [en ligne], Full texts/Numéros en texte intégral, 10/2018 Le web 2.0 : lieux de perception des transformations des sociétés, Le web 2.0 : lieux de perception des transformations des sociétés/Web 2.0: Places of perception of the transformations of societies, mis à jour le : 03/07/2018, URL :

Quelques mots à propos de :  Laura Calabrese

Université libre de Bruxelles ReSIC

Quelques mots à propos de :  Jérémy Jenard

Université libre de Bruxelles ReSIC



Licence Creative Commons
Ce(tte) œuvre est mise à disposition selon les termes de la Licence Creative Commons Attribution 4.0 International.