In this paper we counter the idea that Facebook is a unified medium and stress the need to analyze the distinct qualities of its pages. We do so by exploring the Black Pete discussion on Facebook that ignited when United Nations investigator Verene Shepherd held a plea for the Dutch government to abolish the Saint Nicholas tradition, which features a black-face character, Black Pete. Both quantitative and qualitative methods are employed to interrogate how Facebook mediates this discussion in terms of divergent formatted spaces of participation. Reflecting on our findings against the notions of deliberation, slacktivism and flaming wars, we consider the particular features and potentials of different types of mediation.
Studying discussions on Facebook: The case of Black Pete
Facebook’s spaces of participation
Administrators as moderators
The Dutch tradition of Saint Nicholas (Sinterklaas), celebrated annually on 5 December, became a hotly debated topic in 2013 when United Nations investigator Verene Shepherd, in an interview, called the Dutch government’s attention to what she considered the racist traits of the tradition. She had just started investigating the matter on behalf of the U.N., in response to a complaint concerning the portrayal of the Black Pete (Zwarte Piet) characters in the celebration. These black-faced companions of the Saint are depicted with curly black hair and thick red lips, and wear colorful attire including a lace collar and earrings. Shepherd’s initiative unleashed until then fairly latent discussion in the Netherlands, and brought to the public’s attention the fact that some had long experienced this representation as inherently racist and as glorifying the Dutch colonial past. Many Dutch, however, denied such claims, arguing among others that the Black Pete’s facial color is due to the residual chimney soot acquired during the delivery of gifts. In addition, they also took issue with the UN ‘meddling’ in ‘Dutch culture’.
Within the Black Pete controversy in the Netherlands, the social media platform Facebook played a rather central role. This is due in large measure to the introduction of Pietitie (literally ‘Pete-ition’), a Facebook page in defense of the current form of the celebration, which soon became the fastest growing and most-liked Facebook page ever launched in the Netherlands (with currently more than two million likes) . As a form of pro-Pete protest, the initiators set out to collect as many likes as the U.N. institutional page had at the time (one million) — a goal they accomplished within just one day. Other Facebook pages central to the Black Pete discussion are Zwarte Piet is Racisme (‘Black Pete is Racism’, ZPR), a page initiated by a group of long-time Black Pete protesters, and Anouk (a Dutch singer who took an explicit stand in the discussion). This paper is based on an investigation of these three pages.
In this article we use quantitative and qualitative methods to examine how the Black Pete discussion unfolded on Facebook. In calling itself a ‘platform’, Facebook strategically invokes connotations of openness, egalitarianism and progressiveness (Gillespie, 2010) in its mediation of discussions. However, several authors have argued that socio-cultural and techno-economical affordances actively shape those discussions and determine the nature and potentialities of deliberation on social media (Wright and Street, 2007; Van Dijck, 2012; Davies and Chandler, 2012; Monnoyer-Smith, 2012; Halpern and Gibbs, 2013). In what follows, we expose the capacity of Facebook to mold public discourse, based on the specific set of social and technological affordances that its various pages offer.
In this paper we approach the Facebook pages under scrutiny as a series of “spaces of participation” . This means in practice that we consider how socio-cultural and techno-economical powers structure the interactions that take place on those pages. We are aware in this context of the “logic” defining a social medium like Facebook: “the processes, principles, practices through which these platforms process information, news, and communication”, and which revolve around “programmability, popularity, connectivity, and datafication” . We argue that, those forces entail that each of the Facebook pages examined mediates the Black Pete discussion in its own way.
Pietitie, on the one hand, is a pro-Black Pete Facebook page, initiated by, and addressing, those who seek to retain the Black Pete character in the Saint Nicholas celebration. The ZPR page, on the other hand, represents the anti-Black Pete movement. It was established in 2011 to inform the public about the racist elements of the ritual, and holds a plea for modifying the Black Pete character so as to prevent it from offending certain groups in society. Although both parties’ positions on the topic are not diametrically opposed, the two pages do represent the two sides to the debate. The final page investigated is that of Dutch celebrity singer-songwriter Anouk Teeuwe (known under her stage name Anouk) — a page often referred to on the other pages. Anouk openly took a stand in the discussion on social media and received a great deal of media attention in the Netherlands for doing so. Based on our analysis, the Pietitie page can be characterized as facilitating slacktivism, the ZPR page as a space for deliberation, and the Anouk page as a space where flaming wars ignited. The meaning of these classifications and their significance for and role in the study will be clarified later.
To investigate the pages qualitatively and quantitatively, we scraped the posts and comments from these three pages between 31 October 2011, when the ZPR page was launched, and 30 January 2014. In doing so, we made use of two tools, Pulse and Netvizz, that generated spreadsheets containing over 130,000 posting users and 183,000 messages, and nearly 1.6 million likes distributed over these messages . The datasets contained the texts of comments per post, accompanied by metadata such as the name or user identification number of the commenter, the date and time when a comment was published and the number of likes. The Netvizz files additionally listed engagements per post in numbers of comments, numbers of likes, and numbers of shares.
We begin with a brief consideration of the academic debate on the role of Facebook as a platform for political deliberation and democratic decision-making. Subsequently, in order to highlight the differences between the aforementioned Facebook pages, we explore their respective spaces of participation. In doing so, we focus, consecutively, on four (interrelated) features of these spaces. First, we scrutinize socio-cultural forces by analyzing the role of administrators in the discussions that take place. Second, we trace the distribution of user attention between posts and comments (through likes and comments) and consider how algorithms, as technological forces — necessarily informed by economic agendas — affect communication flows. We then shift our attention to reflect on how the spaces are used by contributors. In the third section, we consider the features of the comments in an attempt to trace how users partook in the conversation (focusing among others on the length of comments and the relations between them). Finally, in the fourth section, we try to determine whether like-minded publics flock together on these pages, by looking at the comments in terms of the respective positions they take in the discussion. In the conclusion, we return to the question of how the Facebook platform mediates public discourse.
The impact of the Internet on democracy has been assessed both favorably and unfavorably. Whereas some argue that online platforms offer a lower threshold for participation and a more egalitarian space for communication (Kellner, 2001), others bemoan how online discussions erupt into flaming wars (Alonzo and Aiken, 2004). More recently, the idea of “slacktivism” emerged (Morozov, 2009; Christensen, 2011; Gladwell, 2010) as a means to reflect upon and scrutinize social media’s democratizing potential. The term, derogatory in connotation, is used to describe online participation that requires little personal effort, and that is directed at achieving personal satisfaction rather than bringing about soci(et)al change.
The notion of the public sphere is a central concept in thinking about communication and citizen engagement. Jürgen Habermas (1974) proposed that early modern capitalism of the eighteenth century encouraged reasoned and critical political deliberation. Here a public sphere offered a space where public opinion could be formed out of rational public debate, enhancing consent and decision-making. Conversation is therefore considered to facilitate democracy. Habermas’ work is often used as a normative ideal against which communication structures like the Internet are evaluated . However, the suitability of the concept for evaluating communication structures in online and social media was contested by, amongst others, Habermas himself, who drew attention to (their affiliation) with corporate structures and their fragmented nature .
Rather than to work with a normative model based on Habermas’ work, we trace how people interact with and around the Black Pete debate on Facebook. Our study empirically charts how the platform is used and functions as a mediator, and uses these findings to critically reflect on prevailing ideas about online deliberation. Rather than to position the Internet as a new public sphere, we choose to focus on how Facebook’s pages shape interaction in different ways, creating distinct spaces of participation (thus in fact also incorporating Habermas’ observations on fragmentation). In doing so, we can however consider Facebook’s potential to serve Habermas’ revised conception of the public sphere as “a network for communicating information and points of view” .
In order to identify the specific characteristics of the pages’ spaces of participation, we can rely on a combination of quantitative and qualitative methods. In interpreting the patterns they reveal, we shall make use of the notions of deliberation, slacktivism and flaming wars.
Online deliberation is understood here as “online communication that is reasoned, purposeful, and interactive” . “Reasoned and purposeful” communication is taken to involve lengthy sentences . Interactivity in turn entails that people respond to each other’s comments. Patterns for both can be revealed using quantitative and qualitative methods. An important feature of deliberative democracy is the possibility for discussion to affect policy and have direct practical implications .
Slacktivism is of a rather different order than deliberation. The term is used to “describe feel-good online activism that has zero political or social impact” . It gives those who participate an illusion of having a meaningful impact on the world without demanding anything more than, for example, joining a Facebook group. People show support for a cause by simply clicking a like-button, sharing a status or changing their profile picture. Here, political participation is not the result of careful deliberation, but executed through easily performed activities. The effectiveness of slacktivism, in terms of real-life outcome, has been both critiqued  and (re-) valued (Christensen, 2011; Vie, 2014).
Flaming wars, in turn, concern “an overly heated and unthinking series of rants among [the] participants” . It is characterized by the use of profanity, obscenity and insults (Alonzo and Aiken, 2004). Although flaming often also involves short responses, the term draws attention in particular to the quality of the debate. Therefore, reflecting on what exactly is being said is crucial also in the analysis. As will be argued further on, ‘flaming wars’ played a central role in the Black Pete discussions, which were marked by emotional reactions, rooted in a perception that opponents jeopardize Dutch culture and identity.
Facebook is composed of relatively monitored spaces in that it affords page administrators the role of gatekeepers. On Pietitie and ZPR the administrator only can publish new posts, as per the page’s post settings. As administrators have a post monopoly, they play a significant role in determining which issues are discussed and when. By the same token, they constitute a socio-cultural force that actively shapes the spaces of participation of the pages. Facebook users may participate on the pages through liking, sharing and commenting on posts, or liking and replying to comments, but they need to rely on administrators to facilitate these interactions. As we discuss further on, most comments published by users ‘drown’ in the sea of other user comments on the pages.
The visualization of the distribution of messages on the Pietitie page in Figure 1 shows that the administrators’ posts play an important role in the liveliness of the page. Whenever a post was published, a large amount of comments followed. However, this activity was short-lived, and within a two-day period, the commenting activity stagnated. Since the number of posts on Pietitie is limited, the page also offers users restricted opportunity to be heard.
Figure 1: Amount of comments on Pietitie posts (each color representing a unique post).
See also https://public.tableausoftware.com/shared/8B9PGMP2Z?:display_count=no.
Note: Larger version of image available here.
Several posts, primarily those on the Pietitie and Anouk page, received the first one hundred comments within a few minutes — in some cases even within a mere three minutes — after the initial post was published. This raises the question whether the ‘conversations’ taking place here can rightfully be considered deliberative acts, as users in this case had no reasonable chance to read, reflect on, and respond to the majority of the comments made. The fact that the comments show little topical consistency seems to confirm this impression.
In contrast to the situation on the Pietitie and ZPR pages, the Anouk page users can actually write posts. In the design of Facebook, posts take a prominent place on each page — a far more prominent one, in any case, than comments (of which only the most recent ones are visible). In this respect, Anouk followers, theoretically speaking at least, are given the opportunity to set the agenda for the discussion that takes place. As we shall demonstrate further on, this theoretical possibility is not necessarily borne out.
Exploring the content of the (administrators’) posts helps to explain the distinct characteristics of the pages. Using the coding scheme below (Table 1), which identifies the particular speech act performed in each case (Searle, 1975), we can categorize posts and show how administrators, through their posts, help to define how a topic is discussed.
Table 1: Coding scheme for analyzing speech acts and quotations in Facebook status messages (Carr, et al., 2012).
Using this coding scheme reveals that posts on Pietitie, amounting to a total of 40 over the period considered, have a high degree of directives (40 percent) in comparison to the other two Facebook pages. The posts mostly prompted people to like and share items, or to partake in contests (e.g., voting for the logo they favored). On more rare occasions, they were asked to respond to a poll to show their support for the cause. The page administrators herein made perfect use of Facebook’s algorithmic architecture and interface design, the underlying “sharing norm” as well as the integrated “economic values [...] [of] attention and popularity” .
Posts like these invite the sort of ‘low-effort’ participation that is commonly associated with slacktivism, and which marks the space of participation of Pietitie. The page was established as an ad hoc reaction to the threat of the tradition being abolished, and not designed as part of a lasting campaign. Posting on the page was suspended on 6 December 2013, after the Saint Nicholas celebration . The ephemerality of the page corresponds to its main goal (collecting more likes than the U.N. institutional page) and name (an adaptation of the term ‘petition’, i.e., a short-lived initiative to collect signatures of as many individuals as possible).
By contrast, the page administrators of ZPR, with a total of 691 posts, kept posting quotations (20 percent of all contributions) and links to Black Pete news coverage. Apparently, they are determined to keep ‘the public’ informed about developments pertaining the Saint Nicholas celebration. Self-reflective posts that explicitly mention their role in nurturing the debate support this observation. The following post shows that the initiators have very particular ideas about how people should share their opinions on the page:
[...] It’s not productive to tell someone something is racist without explaining why. It’s not productive to want to change something simply to avoid having it be discussed. Both positions consider the conversation partner not as a full, honest and lasting conversation partner. We want to end that. We live in a country where honest and enduring conversations, especially now, are needed. (On 15 February 2012, our translation)
Over time, the administrators repeatedly pointed out that they would remove racist comments or personal attacks, and would block users uttering them from the page. The goal of the campaign, as another post made explicit, was to facilitate a productive discussion. It asked users to act respectfully and to remain open to a variety of opinions. In their enforcement of rules, the administrators assumed the role of debate moderators.
The ZPR page featured new posts throughout the year. This seems indicative of the campaign organizers’ ongoing involvement and commitment to creating awareness that the way Black Pete is portrayed can be perceived as racist. Furthermore, while Pietitie replied to 38 user comments, ZPR responded to its users a total of 2,983 times. These observations underscore that the page administrators of ZPR sought to actively encourage a thoughtful exchange of ideas, rather than low-effort interactions such as likes.
Unlike the slacktivist nature of Pietitie contributions and the deliberative character of ZPR posts, the posts and comments on Anouk’s page make for a characterization in terms of ‘flaming wars’. The discussion on the page became heated between 26 October and 14 November 2013 when Anouk published 21 topic-related posts. Almost 40 percent of these posts were expressive, many constituted attacks on other users/posters. The post that received the most comments (4,546) was one in which Anouk asked those people following her who featured a Black Pete figure in their profile photos to either change the photos, or block her on Facebook. Evening addition, she also posted screenshots of private messages from ‘haters’ on her timeline and provided them with an expressive, and often emotional, caption.
Anouk, as page administrator, played a central role in defining the heated tone of the discussion on her Facebook page. Analyzing the 407 user posts pertaining Black Pete in the aforementioned timespan, we mainly find expressive comments (40 percent), and directives (34 percent). Posts were emotional in tone; nearly half of them (45 percent) contained personal attacks, whereas the other half (46 percent) contained declarations of sympathy. In terms of deliberation, most of these posts cannot be characterized as ‘solution-directed.’
As Dahlberg (2001a) argues, flaming is seen as the result of the disinhibiting effects of computer-mediated communication, and there is therefore an inherent need for online discussions to be moderated. However, moderation requires a careful balance between the desire to create civility and freedom of expression (Dahlberg, 2001b). In fact we found that on all three Facebook pages the discussion was moderated in the sense of excluding certain voices: Anouk threatened to remove people from her page with Black Pete profile pictures; Pietitie momentarily suspended posting probably to let the discussion cool down, depriving users from a chance to be heard by means of commenting on recent posts ; and the administrators of the ZPR page laid out rules of conduct that, if not followed, would result in a comment being removed and/or a user banned from the page.
In order to explore the attention distribution on the pages, it is not sufficient to just look at human actors. By assigning agency to mediators like Facebook itself (or its pages), we depart from the idea that social network sites are “networked publics”, where the imagined collective is the result of the intersection between technologies, people and practice . The concept of networked publics further highlights that the possibilities for interaction of the spaces under scrutiny are shaped by networked technology, which relies on programmability (or programmed content flow) and popularity “measured mostly in quantified terms” . Both principles in turn are “conditioned by [...] algorithmic and socio-economic components” . In this respect, Facebook’s spaces of participation should be considered formatted spaces of participation.
Relying on insights from software studies, we consider also the algorithms that calculate how content is ranked and subsequently made visible. Facebook, Bucher (2012) has argued, is prototypical of the “algorithmic power” at work on social media. Consequently, we should take into account how the platform’s news feed is the result of processed user data. Moreover, others have observed that collaborative filtering results in the amplification of popular opinion at the cost of the visibility of opposing arguments .
As we mentioned earlier, the administrators of the Pietitie and ZPR pages are in a position of power because only they can publish posts. Due to the interface design of Facebook, posts are also positioned prominently on a page, making them far more visible, which in turn invites liking and commenting on them. As it turns out, posts did indeed receive more likes than comments, signifying an uneven distribution of attention. This is true especially for Pietitie, where the 40 posts (0.03 percent of all messages) received 89 percent of all likes (see Table 2).
Table 2: Two pages by numbers. In the ZPR column the numbers in brackets are those concerning the period since 22 October 2013, the date Pietitie was founded. (Source: Pulse).
Most user comments here (83 percent) failed to receive any likes. This means that a small number of comments managed to receive a large amount of likes. At the time the data were scraped, Facebook algorithms would automatically highlight two popular comments for each post in a page overview . In the event that a comment was highlighted, this acted as a popularity catalyst: the comments went on to receive (even) more likes. It turns out that the 104 comments that were highlighted on Pietitie (0.084 percent of the total amount of comments) altogether received 55 percent of all the likes granted to comments on the page.
Although the Facebook algorithms are ‘black-boxes’ — in that users, including researchers, do not know how exactly they work — there is one factor that clearly affects whether a comment gets highlighted: the speed at which a comment was published after the initial post. This is illustrated by a comment posted through the Facebook account of a user named Sinterklaas on 23 October 2013 (at 16:55):
The Petes are at this moment busy packing almost 1 million extra presents for you! On behalf of all Petes, thanks for your effort! Keep it up!
(by Sinterklaas, http://facebook.com/Sinterklaas, our translation)
The same comment was posted on the Pietitie page nine times, but in response to different posts. Filtering the comments chronologically, we discover that the first time the comment was posted, it was the seventeenth comment on a post. In this position, it received 5,474 likes — and as such, it was one of the most popular comments to be found on Pietitie. In the eight attempts Sinterklaas made to equate this success, s/he posted the comment later (relative to the original post). With the exception of one of them (which received six likes), they all failed to get any likes. In further support of this insight, we can observe that 91 percent of the highlighted comments were written within an hour of the post. Being one of the first 100 users to comment, however, is no guarantee for success. The popularity of a comment depends on numerous factors, including its content and formal aspects, the day of the week and the time of day when it was published, and if it resonated with the popular discourse.
The user comments on ZPR, in contrast to those on Pietitie, obtained 59 percent of all likes, despite the fact that the same page design applied. Such results match the overall deliberative appearance of the page. Moreover it shows how thinking of these technologies only in terms of their potential or limitations is problematic, as it ignores the reality of how they are used . To understand the impact of these technologies on public debate requires looking at if and how these platforms are mobilized for the purpose of debate.
As for attention distribution on the Anouk page — here users can create posts (not only the initiator/administrators) — 34,408 comments pertaining the Black Pete discussion were shared within a timeframe of three weeks . The 22 posts published by Anouk — 21 of them topic-related — received 96 percent of these comments and 93 percent of all likes . While it is true that the design of the page displays administrators’ posts more prominently, we contend that users are able to comment on and like posts/comments of their choosing. Although the interface strongly influences what we call attention distribution, we find that these percentages reflect the attitude of the users and help to characterize the discussion. The 34,408 comments received 34,457 likes (meaning that the average comment received one like). Apparently, then, individual users had limited sway in the discussion on the page and the situation was more like the Pietitie page than the ZPR page.
The Facebook pages do not offer all active users equal opportunity to be heard. Attention inequality was the rule rather than the exception. This inequality on social media has been broadly documented (Barabási, 2007). The discussion was shaped by the page administrators as well as by the ranking algorithms that distributed attention. However, it is remarkable also that although likes on posts were the most common form of participation on all pages, ZPR featured significantly more likes on comments than Anouk’s page and Pietitie. The prominence of likes on Pietitie is further emphasized when we compare the number of page likers to the numbers of active users on Pietitie (101,527 commenting users and 2,097,799 page likers) and ZPR (12,806 commenting users and 13,878 page likers) respectively . Again we find that the Pietitie page users seem to favor low-effort participation.
Shifting our attention away from how these spaces were shaped, we now address how they were used in order to get a sense of how the Black Pete discussion unfolded. Quantifying the comments made, we try to get an impression of the broader patterns in terms of participation. To do so, we calculate the statistical measures of mean, mode, range, standard deviation and skewness of the characters and words found in each comment. Doing this reveals telling differences between the three pages.
Examining and comparing the averages of characters and words used in the comments made (see Table 3) leads us to the conclusion that interactions on ZPR have features expected from more deliberative expression. With an average number of 42 words per comment (and 246 characters), users on the ZPR page posted the more elaborative comments. In contrast, the comments on Pietitie are half as long, averaging only 20 words . Moreover, the mode, the number most often repeated (in terms of the amount of words and characters), in comments on Pietitie, highlights how short comments are. Considering their length, it is very unlikely that these comments contain motivated arguments. The Anouk Facebook page generated comments with features of word count and character count that lie close to those of Pietitie.
Table 3: Numbers of comments, characters, and words, as well as averages, standard deviation, and skewness calculated from these values .
In quantifying the comments, we interpret what the average values such as mean and mode might signify. In this process, we detect an asymmetry in the distribution of the length of user comments revealed by the measure of skewness that we applied to our datasets. The data concerning comments from all three of the pages demonstrate what is called a “right skewness”  meaning that the tail of distribution is on the right side. This is indicated by the positive values of skewness (see Table 3) implying that on all three Facebook pages, longer comments are the exception rather than the rule. However, while the range of the length of comments on all three pages is comparable in its dimensions, we find the highest number of longer comments on the ZPR page. This is also indicated by relatively lower measures of skewness, in comparison to the other two pages. Based on these findings, deliberation, in the sense of reasoned and purposeful interaction, seem most likely on the ZPR page. It once again reaffirms previously formulated hypotheses.
A qualitative encoding of a sample of comments from the three pages confirms the assumptions formulated above . We encode whether comments were posted as reactions to the initial post or the comment of another user. Furthermore, we identify whether the comments expressed agreement or disagreement with the messages they referred to. We then establish whether or not this (dis)agreement was also motivated with supporting argumentation.
We found most motivated expressions in comments on the ZPR page, which we interpret as signs of deliberation. On this page users referred to comments of other users more frequently. Furthermore, the comments here most often expressed motivated disagreement — as opposed to the other pages, which more often expressed unmotivated agreement. The agreement and disagreement between users leads us to question whether any of these pages constituted enclaves of like-minded publics. This question is of particular importance to our endeavors here, as political theorists have long argued that the exposure to opposing arguments is beneficial for democracy (Sach, et al., 2009).
With algorithms, social media platforms like Facebook tailor their news feeds to individual users. A common fear has been that as a result, users become isolated in a “filter bubble” (Pariser, 2011), since they are only exposed to content they are expected to like. As pointed out by Lev-On and Manin (2009):
When diversity of views is combined with freedom of speech and association, and especially with enhanced abilities to locate like-minded others and filter out opposing views, the result may be enclaves of like-minded people talking to one another, even in a context of a wide multiplicity of views. 
If exposure to opposing arguments is seen as beneficial to democratic interaction, the implication of algorithmic filtering is worrying.
To determine if the Pietitie and ZPR Facebook pages are ‘enclaves of like-minded people’ we take a random sample (n=1,020) of comments from these pages. We encode the messages to determine the stance of participants in the discussion using the coding scheme presented in Table 4 .
Table 4: Coding scheme used to identify positions on Black Pete.
If a page attracts a like-minded public, we expect to be able to infer this on basis of the fact that there is strong consensus between the participants. Our data suggest that Pietitie does operate as a venue for a like-minded public. Here, from the 276 messages that made their position known, all were in favor of keeping the Black Pete character as part of the celebration. Of those 276, only eight were open to (limited) adaptions.
The situation is different on ZPR. Despite the fact that the page is dedicated to raising awareness on the racist elements of Black Pete, only 67 participants made statements to this effect in their comments. In fact, 119 people commented that the Black Pete character should remain a part of the tradition. In other words, the users of this page cannot be characterized collectively as simply ‘anti-Black Pete’, because some found that the Pete character could remain, albeit with adjustments. Furthermore, unlike on Pietitie, the comments on the ZPR page ranged from what could be classified as superficial expressions to opinions supported by argumentation.
In terms of the overlap between users of the different pages it is remarkable that only about one percent of all users commented (or in the case of Anouk’s page, posted) on all three pages. Compared to the absolute numbers of users active on the three pages — 101,527 on Pietitie, 12,806 on ZPR, and 19,076 on Anouk’s page — there were few users at all who commented on more than one page. The exception to this rule were the 1,500 users who commented on both the ZPR and the Pietitie page. In an attempt to determine the allegiance of these users, we establish on which page they commented first. We found that most of these users first commented on ZPR (68 percent, n=966), suggesting once again that the page can be seen as a place attractive to users more concerned with deliberation.
In conclusion, we can state that all three pages seem to be like-minded bubbles within Facebook. And, among people gravitating towards like-minded users, exposure of users to divergent opinions is unlikely. On the whole, ZPR users presented a bit of an exception to this rule, in that they more often also commented on another page. However, it would be worth to investigating the situation on the Anouk page more closely, as users here are brought together not on the basis of their position in the Black Pete debate, but their interest in the musician/her output. Perhaps this factor too helps to explain the ignition of the page into flaming wars.
This paper has sought to highlight the fact that Facebook offers various spaces of participation for the discussion of Black Pete. To cluster social media, or even Facebook’s pages, and claim that they necessarily impact democracy in similar ways seems to be based on a mistaken assumption. The discussions on all three pages we explored had different characteristics, despite sharing the Facebook platform. We infer from this that different socio-cultural and (even) technological forces were at play on each of these pages, generating a number of distinct spaces of participation.
We have found that the ZPR page does show characteristics expected from a deliberative space, with the administrators acting as moderators of the Black Pete discussion. The rules they enforced actually helped to procure the quality of the debate. They encouraged users to reflect on their position and to use arguments in the comments they made. What we learn here about the use of social media for online deliberation confirms the proposition made by Michael Schudson that democracy served through conversation is best served not by egalitarianism, but by norm-governedness and publicness . In this particular case, then, participation in the debate benefited from the socially formatted character of the page, which generated a (rudimentary) deliberative space. By contrast, Anouk ignited flaming wars on her page and Pietitie actively encouraged slacktivism by prompting low-effort participation (see Table 5 for a schematic overview of main findings).
Table 5: Overview of findings. Note: Larger version of image available here.
It has been discussed how administrators have the capacity to prevent free speech from becoming just noise through regulation , positively influencing the quality of the debate. However, all administrators to some degree moderated the discussion by excluding certain opinions and people from the discussion. This could be said to harm the process of deliberation and personal autonomy found important to deliberative democracy. As mentioned, a careful balance needs to be sought between civility and freedom of expression.
A particular problem the ZPR page faced is the fact that administrators could not procure that both sides of a debate were represented. This is because people voluntarily ‘like’ the page and comment on posts. In the end, although the administrators managed to secure the quality of the debate, it was fairly one-sided in that ZPR attracted those with an intention to campaign for change. The Internet makes it easy for individuals to find groups of people on Facebook that share their point of view.
We have discussed how interface design and algorithms structure the flow of information and communication on Facebook. Although algorithms constitute powerful actors, their exact effects are largely unknown, as they are produced or owned by corporate entities that prefer not to disclose information on how they work. The choice to protect this information directly relates to the economic interests shaping Facebook’s pages, and ties in with its tendency to stimulate interactions in order to collect data for analysis and targeted advertising. The power of Facebook should not be ignored. The interface design of the Facebook pages and the algorithms working beneath the surface configure the power relations between participants in the discussion. Facebook is anything but neutral in that it hinders equal access and opportunity, making only a few users visible (and thus heard) and stimulating the formation of like-minded publics through algorithmic filtering. That it is powerful, but not determining, has been illustrated in that attention on the ZPR page is more equally distributed between posts and comments (although posts remain convincingly more prominent).
As for the societal implications of the analyzed discussions, these are difficult, if not impossible, to measure. However, some tentative conclusions can be drawn as to the outcome of the discussion. Herein we find it important to recognize how the slacktivist tendencies found on the Pietitie page, as well as the flaming wars on the Anouk page fueled mass media attention. The attention of the central media for the manifestation of the discussion on Facebook made it a societal debate with all Dutch citizens being drawn into the discussion. In this way, slacktivism has proven to be important in mobilizing people (Christensen, 2011; Vie, 2014). In late 2014 we see the Black Pete discussion re-surfacing and find the tradition in a state of flux. In particular traditional mass media constitute the arena of actual deliberation, in which both positions are represented. Even supermarkets have made statements in press releases on the issue (and two have even been boycotted by customers for saying that they intend on removing the traditional image from stores and in their commercials).
Of particular interest has been how Jeugdjournaal (a popular news program targeted at children aged 10-12) has dealt with the issue. The program shows clear indications of change as different colored Petes are introduced on television and narratives are offered to explain their emergence. Indeed upon arriving in Gouda (the Netherlands) Saint Nicholas was accompanied by different colored Petes including the ‘Stroopwafel Pete’, in honor of the popular Dutch syrup biscuit, the ‘Cheese Pete’, and the ‘Rainbow Pete’. Although the tides are changing, it is a slow process requiring long-lasting commitment and deliberation, as shown by those of the anti-Black Pete movement who have sought to provoke and engage in discussion, on numerous platforms, for some time.
About the authors
Karin van Es received her Ph.D. from Utrecht University in 2014. She currently works as a lecturer in the Media Studies Department at the University of Amsterdam and is project coordinator at the Utrecht Data School (UDS). Her research concerns television as a medium in transition.
E-mail: k [dot] f [dot] vanes [at] uva [dot] nl
Daniela van Geenen is a Master’s student in New Media and Digital Culture at Utrecht University (UU), with a background in Cultural Studies. She is a junior researcher and teaching assistant at UDS. Her main research interest concerns the intersection between data-driven research, the impact of algorithms on and the role of images in the production of knowledge.
E-mail: daniela [at] dataschool [dot] nl
Thomas Boeschoten is founder of and researcher at UDS. UDS is a research platform at Utrecht University, which, among other things, offers students the opportunity to conduct data-driven research on behalf of companies, governmental and non-governmental organizations.
E-mail: thomas [at] dataschool [dot] nl
The authors would like to acknowledge the Institute for Cultural Inquiry at Utrecht University for funding this research.
1. Various media, both national and international, informed about and reflected on this milestone. See e.g., http://www.nrc.nl/nieuws/2013/10/23/pietitie-heeft-een-miljoen-facebook-likes-snelstgroeiende-pagina-in-nederland/, http://www.volkskrant.nl/binnenland/zwartepietpetitie-op-facebook-naar-1-miljoen-likes~a3531937/, or http://www.theguardian.com/world/2013/dec/06/black-pete-dutch-christmas-zwarte-piet.
2. Müller, 2009, p. 60.
3. Van Dijck and Poell, 2013, p. 5.
4. Pulse was provided to us by IT company Ordina B.V. Netvizz was developed by media scholar Bernhard Rieder as part of the Digital Methods Initiative, situated at the University of Amsterdam (Rieder, 2013). Using both tools allowed us to compare the gathered data in order to overcome possible flaws in crawling methods; moreover, they provided us with different metadata, which made it possible to analyze a broader range of page characteristics.
5. Chadwick, 2012, p. 47.
6. Habermas, 2006, referenced in van Dijck, 2012, p. 163.
7. Habermas, 1996, p. 360.
8. Davies, 2009, p. 2.
9. As opposed to single words or phrases expressing mere agreement or disagreement; see Halpern and Gibbs, 2013, pp. 1,160-1,163.
10. Cavalier, et al., 2009, p. 71.
11. Morozov, 2009, n.p.
13. Poor, 2005, n.p.
14. Van Dijck, 2013, pp. 62, 65.
15. In the last phase of revising this paper (November 2014), just before the Saint Nicholas season of 2014, revived activity on Pietitie is noticeable — which once again demonstrates the occasional character of activism on the page.
16. The administrators suspended their activity on the page on 31 October (until 15 November) 2013, shortly after pro-Black Pete protests on the Malieveld (The Hague) escalated, and probably because users on Pietitie reacted by means of emotional comments on those developments.
17. boyd, 2011, p. 39.
18. Van Dijck and Poell, 2013, pp. 6-7.
20. Lev-On and Manin, 2009, p. 111.
21. The scraping for this paper was done in May 2014. Users looking at Pietitie and ZPR pages at that time could opt to see the comments thread sorted either by activity type (i.e., liking, sharing, commenting), as “Top Comments” (which was the default and dominant setting), or chronologically, as “Recent Activity”. Such a feature is in constant flux, as Facebook perpetually updates algorithms and the ways in which people can interact. Since March 2013, for example, it is possible to reply to comments made to a post directly and comments are now filtered according to weight rather than chronology (Lavrusik, 2013).
22. Couldry, 2000, p. 190.
23. Since 26 October 2013. In that same period, ZPR generated 23,525 comments.
24. In this period, 412 user posts related to the discussion were produced; five of them did not explicitly mention the topic, but other users commented on them referring to the controversy.
25. Facebook does not provide the possibility to determine whether active users are page likers as well.
26. Here we leave out the more than 50,000 comments the earlier mentioned logo-voting post received, since most of the comments contained no more than one character: the letter or number assigned to the chosen logo.
27. Values in this table for mean, standard deviation and skewness are rounded off. All measures are calculated in Excel. To keep things comparable and compact, we choose to display here measures calculated based on data from comments produced in a similar time frame. For the comments on the ZPR page, the measures do not deviate significantly from measures calculated for all comments published within the whole time span the page existed.
28. Kallner, 2013, p. 8.
29. Specifically, we looked at the first 100 comments published as reactions to two posts per page, dealing with similar topics (n=600). These six posts were published on 26 October 2013, the first day Anouk was concerned with the topic of Black Pete on her Facebook page, and the day after. The posts on Pietitie and ZPR dealt, amongst others, with Anouk’s role in the debate and with demonstrations related to the topic that ended in riots.
30. Lev-On and Manin, 2009, p. 107.
31. Two encoders worked on the encodings separately, after which their results were compared. If the encodings did not correspond, a third encoder assessed the message, and if uncertainty remained, it was marked as ‘not encoded’. Our gratitude extends to Heidi de Mare from Stichting IVMV (Instituut voor Maatschappelijke Verbeelding) and Tilburg University students Desirée van Bakel, Veerle Hendriks, Jaap Peeters, Jan van der Welle, Celine de Vries, and their supervisor Prof. Dr. Gabriël van den Brink, who put tremendous effort into encoding the comments sample and shared their impressions and ideas about the research.
32. Schudson, 1997, p. 297.
33. Blumler and Coleman, 2001, pp. 17-18.
Mei Alonzo and Milam Aiken, 2004. “Flaming in electronic communication,” Decision Support Systems, volume 36, number 3, pp. 205–213.
doi: http://dx.doi.org/10.1016/S0167-9236(02)00190-2, accessed 26 November 2014.
Albert-László Barabási, 2009. “Scale-free networks: A decade and beyond,” Science, volume 325, number 5939 (24 July), pp. 412-413.
doi: http://dx.doi.org/10.1126/science.1173299, accessed 26 November 2014.
Jay G. Blumler and Stephen Coleman, 2001. Realising democracy online: A civic commons in cyberspace. London: Institute for Public Policy Research, and at http://www.ippr.org/publications/realising-democracy-online-a-civic-commons-in-cyberspace, accessed 26 November 2014.
Taina Bucher, 2012. “Want to be on the top? Algorithmic power and the threat of invisibility on Facebook,” New Media & Society, volume 14, number 7, pp. 1,164-1,180.
doi: http://dx.doi.org/10.1177/1461444812440159, accessed 26 November 2014.
danah boyd, 2011. “Social network sites as networked publics: Affordances, dynamics, and implications,” in: Zizi Papacharissi (editor). Networked self: Identity, community, and culture on social network sites. London: Routledge, pp. 39-58.
Robert Cavalier with Miso Kim and Zachary Sam Zaiss, 2009. “Deliberative democracy, online discussion, and Project PICOLA (Public, Informed Citizen Online Assembly),” In: Todd Davies and Seeta Peña Gangadharan (editors). Online deliberation: Design, research, and practice. Stanford, Calif.: Center for the Study of Language and Information (CSLI), pp. 71-79, and at http://odbook.stanford.edu/, accessed 26 November 2014.
Andrew Chadwick, 2012. “Web 2.0: New challenges for the study of e-democracy in an era of informational exuberance,” In: Stephen Coleman and Peter M. Shane (editors). Connecting democracy: Online consultation and the flow of political communication. Cambridge, Mass.: MIT Press, pp. 45-74.
Henrik Serup Christensen, 2011. “Political activities on the Internet: Slacktivism or political participation by other means?” First Monday, volume 16, number 2, at http://firstmonday.org/article/view/3336/2767, accessed 1 May 2014.
Nick Couldry, 2000. The place of media power: Pilgrims and witnesses of the media age. London: Routledge.
Lincoln Dahlberg, 2001a. “Computer-mediated communication and the public sphere: A critical analysis,” Journal of Computer-Mediated Communication, volume 7, number 1, at http://onlinelibrary.wiley.com/doi/10.1111/j.1083-6101.2001.tb00137.x/full, accessed 1 November 2014.
doi: http://dx.doi.org/10.1111/j.1083-6101.2001.tb00137.x, accessed 26 November 2014.
Lincoln Dahlberg, 2001b, “Extending the public sphere through cyberspace: The case of Minnesota E-Democracy,” First Monday, volume 6, number 3, at http://firstmonday.org/article/view/838/747, accessed 9 November 2014.
Todd Davies, 2009. “The blossoming field of online deliberation,” Todd Davies and Seeta Peña Gangadharan (editors). Online deliberation: Design, research, and practice. Stanford, Calif.: Center for the Study of Language and Information (CSLI), pp. 1-19, and at http://odbook.stanford.edu/, accessed 26 November 2014.
Todd Davies and Reid Chandler, 2012. “Online deliberation design: Choices, criteria, and evidence,” In: Tina Nabatchi, Michael Weiksner, John Gastil and Matt Leighninger (editors). Democracy in motion: Evaluating the practice and impact of deliberative civic engagement. Oxford: Oxford University Press, pp. 103-134.
José van Dijck, 2013. The culture of connectivity. A critical history of social media. Oxford: Oxford University Press.
José van Dijck, 2012. “Facebook as a tool for producing sociality and connectivity,” Television & New Media, volume 13, number 2, pp. 160–176.
doi: http://dx.doi.org/10.1177/1527476411415291, accessed 26 November 2014.
José van Dijck and Thomas Poell, 2013. “Understanding social media logic,” Media and Communication, volume 1, number 1, pp. 2-14, at http://www.librelloph.com/mediaandcommunication/article/view/MaC-1.1.2, accessed 26 November 2014.
doi: http://dx.doi.org/10.12924/mac2013.01010002, accessed 26 November 2014.
Malcom Gladwell, 2010. “Small change: Why the revolution will not be tweeted,” New Yorker (4 October), at http://www.newyorker.com/reporting/2010/10/04/101004fa_fact_gladwell, accessed 1 May 2014.
Jürgen Habermas 2006. “Political communication in media society: Does democracy still enjoy an epistemic dimension? The impact of normative theory on empirical research,” Communication Theory, volume 16, number 4, pp. 411–426.
doi: http://dx.doi.org/10.1111/j.1468-2885.2006.00280.x, accessed 26 November 2014.
Jürgen Habermas, 1996. Between facts and norms: Contributions to a discourse theory of law and democracy. Translated by William Rehg. Cambridge, Mass.: MIT Press.
Jürgen Habermas, 1974. “The public sphere: An encyclopedia article (1964),” New German Critique, number 3, pp. 49-55.
doi: http://dx.doi.org/10.2307/487737, accessed 26 November 2014.
Daniel Halpern and Jennifer Gibbs, 2013. “Social media as a catalyst for online deliberation? Exploring the affordances of Facebook and YouTube for political expression,” Computers in Human Behavior, volume 29, number 3, pp. 1,159-1,168.
doi: http://dx.doi.org/10.1016/j.chb.2012.10.008, accessed 26 November 2014.
Andreas Kallner, 2013. Laboratory statistics: Handbook of formulas and terms. Amsterdam: Elsevier.
Douglas Kellner, 2001. “Techno-politics, new technologies, and the new public spheres,” illuminations, at http://www.uta.edu/huma/illuminations/kell32.htm, accessed 10 September 2014.
Vadim Lavrusik, 2013. “Improving conversations on Facebook with replies,” Journalists on Facebook (25 March), at https://www.facebook.com/notes/facebook-journalists/improving-conversations-on-facebook-with-replies/578890718789613, accessed 10 September 2014.
Azi Lev-On and Bernard Manin, 2009. “Happy accidents: Deliberation and online exposure to opposing views,” In: Todd Davies and Seeta Peña Gangadharan (editors). Online deliberation: Design, research, and practice. Stanford, Calif.: Center for the Study of Language and Information (CSLI), pp. 105-122.
Laurence Monnoyer-Smith, 2012. “The technological dimension of deliberation: A comparison between online and offline participation,” In: Stephen Coleman and Peter M. Shane (editors). Connecting democracy: Online consultation and the flow of political communication. Cambridge, Mass.: MIT Press, pp. 191-208.
Evgeny Morozov, 2009. “The brave new world of slacktivism,” Foreign Policy (19 May), at http://neteffect.foreignpolicy.com/posts/2009/05/19/the_brave_new_world_of_slacktivism, accessed 1 May 2014.
Eggo Müller, 2009. “Formatted spaces of participation: Interactive television and the changing relationship between production and consumption,” In: Marianne van den Boomen, Sybille Lammes, Ann-Sophie Lehmann, Joost Raessens and Mirko Tobias Schäfer (editors). Digital material: Tracing new media in everyday life and technology. Amsterdam: Amsterdam University Press, pp. 47-61.
Eli Pariser, 2011. The filter bubble: What the Internet is hiding from you. New York: Penguin Press.
Nathaniel Poor, 2005. “Mechanisms of an online public sphere: The Website Slashdot,” Journal of Computer-Mediated Communication, volume 10, number 2, at http://onlinelibrary.wiley.com/doi/10.1111/j.1083-6101.2005.tb00241.x/full, accessed 1 November 2014.
doi: http://dx.doi.org/10.1111/j.1083-6101.2005.tb00241.x, accessed 26 November 2014.
Bernhard Rieder, 2013. “Studying Facebook via data extraction: The Netvizz application,” Web Sci ’13: Proceedings of the Fifth Annual ACM Web Science Conference, pp. 346-355.
doi: http://dx.doi.org/10.1145/2464464.2464475, accessed 26 November 2014.
Warren Sach, John Kelly and Michael Dale, 2009. “Searching the Net for differences of opinion,” In: TTodd Davies and Seeta Peña Gangadharan (editors). Online deliberation: Design, research, and practice. Stanford, Calif.: Center for the Study of Language and Information (CSLI), pp. 95-103.
John R. Searle, 1975. “A taxonomy of illocutionary acts,” In: Keith Günderson (editor). Language, mind, and knowledge. Minneapolis: University of Minneapolis Press, pp. 344-369.
Michael Schudson, 1997. “Why conversation is not the soul of democracy,” Critical Studies in Mass Communication, volume 14, number 4, pp. 297-309.
doi: http://dx.doi.org/10.1080/15295039709367020, accessed 26 November 2014.
Stephanie Vie, 2014. “In defense of ‘slacktivism’: The Human Rights Campaign Facebook Logo as digital activism,” First Monday, volume 19, number 4, at http://firstmonday.org/article/view/4961/3868, accessed 1 November 2014.
doi: http://dx.doi.org/10.5210/fm.v19i4.4961, accessed 26 November 2014.
Scott Wright and John Street, 2007. “Democracy, deliberation and design: The case of online discussion forums,” New Media & Society, volume 9, number 5, pp. 849–869.
doi: http://dx.doi.org/10.1177/1461444807081230, accessed 26 November 2014.
Received 21 October 2014; revised 17 November 2014; revised 20 November 2014; revised 25 November 2014; accepted 25 November 2014.
“Mediating the Black Pete discussion on Facebook: Slacktivism, flaming wars, and deliberation” by Karin van Es, Daniela van Geenen, and Thomas Boeschoten is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Mediating the Black Pete discussion on Facebook: Slacktivism, flaming wars, and deliberation
by Karin van Es, Daniela van Geenen and Thomas Boeschoten.
First Monday, Volume 19, Number 12 - 1 December 2014
A Great Cities Initiative of the University of Illinois at Chicago University Library.
© First Monday, 1995-2016.