Master’s Thesis:
Understanding Incivility in User Comments: A Content Analysis of News Websites Nori Messchaert, 10764380
Graduate School of Communication - Universiteit van Amsterdam June 26, 2015
Author Note
Nori Messchaert, MSc. Political Communication,
Master’s programme Communication Science - Universiteit van Amsterdam Supervisor Dr. Sophie Lecheler
Correspondence concerning this Master’s thesis should be addressed to nori.messchaert@student.uva.nl
Abstract
Academic concerns have arisen about a so-called ‘civility crisis’ in public and political life. This crisis appears to manifest and spread itself online, and literature suggests that this online incivility (1) undermines the fundamental democratic principles of modern societies, (2) entails negative effects on citizens, and (3) discourages the establishment of respectful
conversations in society. Surprisingly, research on the matter has been mostly conducted in an American setting, even though several authors claim it is also a European issue.
Consequently, this Master’s thesis content analyzes (N = 301) political news articles’ user comments of two popular news websites (NU.nl in the Netherlands and HLN.be in Belgium). Findings suggest that even though the amount of incivility is quite substantial, it does contain a lot of meaningful arguments and/or statements. In addition, conflict issues, such as
elections, and interaction between users can predict the amount of incivility in comment threads. The results also show that there is no systematic difference in incivility between the two news websites. The findings could imply that academic concerns about a ‘civility crisis’ might be a bit overstated, and that it is rather a societal concern.
Understanding Incivility in User Comments: A Content Analysis of News Websites Bas van Stokkum, renowned sociologist and philosopher, argues that the Netherlands are under the spell of loutish behavior (2010). Van Stokkum proclaims that Dutch citizens are becoming increasingly aggressive in social situations, for example through the use of
swearwords, threats, or other types of unwanted verbal behavior. He also concludes that rude manners and the stigmatization of minorities have become the rule and not the exception. His concerns are legitimized by academic research showing a so-called ‘civility crisis’ in public and political life, which deeply infects TV, radio, talk, blogs, and other media outlets (e.g., Jamieson, 1999; Mutz, & Reeves, 2005; Rodin, 1996).
Interestingly, this crisis appears to predominantly manifest itself online (e.g.,
Thompsen, & Foulger, 1996). Therefore, it is not surprising that academics have encountered incivility in online discussions, such as in comment threads of political newsgroups, and political blogs (e.g., Papacharissi, 2004; Sobieraj, & Berry, 2011). These discussions and other online platforms enable people to ‘act out’, and appear to have become a breeding place for uncivil behavior, as people swear, use vulgar language, and disrespect other people to get their point across (Papacharissi, 2004).
Research also shows that incivility can be detrimental to democracy, as it has demonstrated that exposure to online uncivil content may cause polarization, and the
formation of negative attitudes towards politicians and governmental institutions (Anderson, Brossard, Scheufele, Xenos, & Ladwig, 2014; Cappella, & Jamieson, 1997; Smith, John, Sturgis, & Nomura, 2009). Moreover, incivility activates feelings of aversion, anger, and dissatisfaction if it targets specific persons or their groups to whom they belong (Gervais, 2015).
However, there are several gaps in the literature that research must address. Firstly, most insights in incivility derive from an exclusively American setting (e.g., Coe, Kenski, &
Rains, 2014; Ksiazek, Peer, & Zivic, 2014; Papacharissi, 2004), even though incivility is also a European issue. Sociologists such as Van Stokkum have noticed the appearance of online uncivil traits in the Netherlands (2010), and others have noted rudeness in EU political discourse. Think of, for instance, when Nigel Farage, a European Parliament member,
claimed that Herman van Rompuy, the former president of the European Council, looked like a ‘low-grade bank clerk’ (Daily Mail Reporter, 2012). Secondly, we are not aware of how much incivility there is in political discourse, as most research has focused on its effects (e.g., Fridkin, & Kenney, 2008). Thirdly, we need to further test if incivility online can demobilize and disengage people (e.g., Ksiazek et al., 2014). Thus, incivility is a societally and
academically relevant topic, as it is hard to draw conclusions on whether incivility is actually that detrimental to society, and if the crisis has emerged in Europe.
Therefore, this Master’s thesis aims to ascertain to what extent the European political landscape is ‘infected’ by incivility, and if in fact incivility is a societal problem. In addition, I investigate whether there are certain relational determinants or factors of incivility present. By means of a quantitative content analysis, I analyze user comments posted to political news articles on the Dutch news website NU.nl, and the Flemish news website HLN.be. According to Neurauter-Kessels (2011), articles that attract user comments may contain controversial data, and thus, may also contain incivility. In addition, studying user comments is important, as the dissemination of thoughts in user comments can polarize people who passively read user comments (Anderson et al., 2014; Smith et al., 2009), and affect journalistic production in the sense that when journalists read user comments they may refuse to write certain pieces to prevent uncivil reactions from occurring online (Reader, 2012).
Overall, the study provides a first European insight of incivility, as it investigates whether two outlets from two countries (Netherlands, Belgium), which are similar in their political background (Hooghe, 2005), differ in incivility. I believe that incivility is not merely
a Dutch issue, but a European one as well, or even a global issue, as it is linked to computer-mediated communication (e.g., Thompsen & Foulger, 1996), and thus may find its way to everyone who has access to the Internet and online discussion forums.
Understanding Political Discussions Online: An Uncivil Landscape?
Incivility, or the dissemination of online hostile interactions, has become a major area of interest for academics (O’Sullivan, & Flanagin, 2003). Research increasingly focuses on the decline in political and public civility, as some promulgate that a ‘civility crisis’ has emerged in the United States (e.g., Jamieson, 1999; Mutz, & Reeves, 2005). This so-called civility crisis is believed to endanger political conversation, as it discourages meaningful participation in public life by weakening political trust, and causing citizens to be dissatisfied with politics (Elving, as cited in Anderson et al., 2014; Mutz, & Reeves, 2005; Schudson, 1997). Academics argue that TV, radio, social and political life are increasingly affected by incivility (Rodin, 1996). However, the recent focus on incivility does not imply that public discourse has always been civil. On the contrary, incivility has always been present, but the 21st century interactive media environment and its narrowcasting character create new
opportunities for incivility to spread more rapidly (Sobieraj, & Berry, 2011). Therefore, more focus and attention is paid to the current spread of online incivility and its effects than before (Coe et al., 2014). Yet, the focus on effects appears troublesome, as there is still a lack of comprehensive understanding of political incivility itself (Sobieraj, & Berry, 2011). Does Incivility Really Endanger Society?
In the 21st century’s interactive media environment, people can consult news, comment on news articles, and interact with other users. In this manner, citizens can (often anonymously) provide other readers and journalists with their opinions about news coverage. They can also discuss political issues online (Dürscheid, 2007; Manosevitch & Walker, 2009), due to the fact that online comments offer interactive audience participation by the
click of a button (Neurauter-Kessels, 2011). These discussions can be civil, but research has shown that they are often also uncivil (e.g., Papacharissi, 2004).
This raises the question of whether we should be worried about the presence of online uncivil content. In general, the literature suggests that online incivility is believed to be a threat, due to the fact that (1) it undermines the fundamental democratic principles of modern societies, (2) entails negative effects on citizens, and because (3) it discourages the
establishment of respectful conversations in society.
Stromer-Galley and Wichowski (2011) proclaim that it “is through political conversation that members of society come to clarify their own views, learn about the opinions of others, and discover what major problems face the collective” (p.169). Their quote emphasizes the importance of the nature of political conversations. In that sense, meaningful participation in public life can be established and maintained, provided that citizens remain civil and respectful in political conversations (Schudson, 1997). Therefore, it appears logical that the dissemination of online uncivil content must be discouraged,
particularly because researchers suggest that exposure can have detrimental effects on citizens (e.g., Cappella, & Jamieson, 1997; Ng, & Detenber, 2005). Although my research is not experimental by nature, it is nonetheless necessary to briefly stress what effects emerge when one is exposed to uncivil content, and why scholars are worried.
Exposure to uncivil content can affect citizens’ political perceptions and behavior. Scholars argue that television coverage of uncivil political discourse causes audiences to be dissatisfied with politics, weakens political trust, and causes audiences to evaluate political institutions, governments and political arguments from the opposition more negatively (Elving, as cited in Anderson et al., 2014; Forgette, & Morris, 2006; Mutz, 2007; Mutz & Reeves, 2005). These negative feelings towards incivility arise as literature suggests that exposure to uncivil opposing views in the media is not in line with established civil norms for
verbal interaction. In turn, television coverage of incivility instantly speaks to one’s emotions, as it mimics real life conversation, and increases these negative feelings (Mutz, 2007; Mutz, & Reeves, 2005).
Furthermore, research demonstrates that uncivil expression negatively influences the credibility of other discussants and their messages (Ng, & Detenber, 2005), and that it may cause polarization. ‘Lurkers’ are people who passively partake in online discussions by only reading along. Consequently, scholars worry that reading uncivil messages makes audiences close-minded towards new information, and causes them to form their opinions on their own preexisting knowledge (Anderson et al., 2014; Smith et al., 2009). An explanation for this close-mindedness is that feelings of anger, aversion, and dissatisfaction may be activated if incivility targets specific persons or a person’s group (Gervais, 2015). In turn, people are unreceptive to new information (MacKuen, Wolak, Keele, & Marcus, 2010)
However, researchers have also encountered positive effects of incivility. Brooks and Geer (2007) argue that incivility in campaign discourse can stimulate the propensity to vote and spark political interest. In addition, uncivil comments are not always meaningless and hurtful to the discussion. On the contrary, Coe et al. (2014) encountered in their content analysis on user comments on a newspaper website that 16.5 percent of uncivil comments contained some sort of evidence. Comments that contain evidence are, according to Coe et al. (2014), comments that use facts, statistics, dollar amounts, and/or percentages to get a point across. An example could be “at the end of 2008, our economy had contracted by -9 percent. Change happened. It reversed and headed upwards by 13 points ending in 2009 with a plus 4. I call that success’’ (p.667). Interestingly, only 13.7 percent of civil comments contained evidence.
All these unwanted uncivil effects that researchers have found, could demolish what we perceive to be important in modern society: to engage in a respectfully civilized manner in
public discourse. According to Papacharissi (2004) and others, civility in public discourse has always had profound importance in democracies, as it enables democracy to function properly in a respectful manner (Herbst, 2010; Papacharissi, 2004; Sapiro, as cited in Coe et al., 2014). Moreover, in order to establish and maintain civil societies, public opinion must be
disseminated (Papacharissi, 2004). Aristotle is believed to be one of the first to come up with the term ‘civil society’ in order to portray “a form of political association referred to as ‘state’ or ‘polis’ ” (Papacharissi, 2004, p. 263). In ancient Athens, a person was considered a citizen, if one was civilized and politically oriented. If this was not the case, then people were
considered ‘barbarians’ (Calhoun, 1992).
In sum, we can notice that incivility can entail both positive and negative effects on society (e.g., Coe et al., 2014; Gervais, 2015). Consequently, this study aims to ascertain which perspective is true.
Defining Incivility Online
Now that we are aware of the effects of incivility, we have yet to answer when a user comment or message is considered uncivil. This question is difficult to answer, as scholars have looked at incivility in different academic domains, media outlets, countries, and via the use of different research methods (e.g., Coe et al., 2014; Mutz, & Reeves, 2005). While this study focuses on user comments, I use definitions stemming from research on news content, and experimentation also, as there is no clear answer to what incivility is. In general, I argue that incivility should comprise the following four aspects, described as follows:
First, some scholars who researched online incivility in interpersonal discussions and discussions in political newsgroups, perceive it as a manner of engaging offensively in a discussion, which is inappropriate in a democracy (Anderson et al, 2014; Papacharissi, 2004). Online uncivil tactics entail rude critiques, name-calling, outrageous claims, and incensed
discussion, but also aspersions, synonyms for liar, hyperbole, pejorative speak, vulgarity and words indicating non-cooperation (Jamieson, 1999; Papacharissi, 2004).
A second factor stems from Brooks and Geer (2007), who studied incivility in negative campaigning, and Sobieraj and Berry (2011), who examined the concept in political blogs, cable news, talk and radio. Both parties go beyond merely rude critiques and name-calling, as they proclaim that vital aspects of incivility are the derision of the opposition, and to provoke a visceral response from other discussants. The parties differ, however, in the sense that Sobieraj and Berry call this aspect outrage and not incivility, which is in their perception considered to be something more than just incivility.
A third characteristic of incivility that was encountered while examining user comments on newspaper websites and televised political disagreement is disrespectfulness. Both Coe et al. (2014) and Mutz and Reeves (2005) state that incivility conveys disrespectful, unnecessary tones towards discussion topics and discussants.
Lastly, O’Sullivan and Flanagin (2003) coin incivility as ‘flaming’, and proclaim that it is an intentional manner of disseminating hostile and aggressive interactions online. Mortensen (1997) makes an important contextual distinction in inappropriate online content that other scholars have not, because he argues that flaming cannot be generally marked as inappropriate, as it is based upon the observer’s perspective.
An explanation as to why so many different perceptions of incivility exist could be due to the abstract nature of the term, which appears to be connected to negativity (Brooks, & Geer, 2007). Political communication scholars have long examined negativity and negative campaigning, trying to ascertain what (detrimental) effects it could have on society (e.g., Ansolabehere, Iyengar, Simon, & Valentino, 1994; Finkel, & Geer, 1998; Lau, 1982; Lau, & Pomper, 2001). Both negative campaigning as well as incivility contain the same basis: to criticize another person (e.g., Lau & Pomper, 2001; Papacharissi, 2004). However, incivility
is different from negativity, in the sense that it is not considered a tactic to mobilize or demobilize voters (e.g., Ansolabehere et al., 1994; Finkel, & Geer, 1998). Negativity can be productive when implemented in political discourse, as it can spark political interest, whereas incivility is hyperbolic and is perceived not to convey any extra information (Brooks, & Geer, 2007).
Due to the many (American) views on this abstract term (e.g., Anderson et al., 2014; Brooks, & Geer, 2007; Coe et al., 2014; Mutz, & Reeves, 2005; O’Sullivan, & Flanagin, 2003; Papacharissi, 2004; Sobieraj, & Berry, 2011), I consider a broader definition of incivility more appropriate and suitable, particularly because we are not certain yet if incivility develops and evolves differently in a European setting. In sum, I perceive online incivility to be any characteristics of discussion through computer-mediated communication, that are deemed by the researcher to be of intentional, disrespectful, and offensive nature towards discussion forums, their topics or their participants, which aim to provoke a visceral response from other discussants (Anderson et al., 2014; Brooks, & Geer, 2007; Coe et al., 2014; Mortensen, 1997; Mutz, & Reeves; 2005; O’Sullivan, & Flanagin, 2003; Papacharissi, 2004; Sobieraj, & Berry, 2011). My definition specifically accounts for online incivility, and not offline incivility, as factors such as anonymity, and asynchronicity of the web are said to cause uncivil behavior by allowing users to disinhibit, and experience the feeling of not being able to be held accountable for their rude behavior (Neurauter-Kessels, 2011; Suler, 2004). Therefore, I cannot simply claim that both offline and online incivility can be defined in the same manner, as online communication has specific factors that cause incivility.
In turn, online civility is commonly absent of uncivil characteristics, and thus can be considered as the counterpart of incivility, which implies a respectful and appropriate form of behavior through computer-mediated communication (Carter, as cited Brooks, & Geer, 2007; Sapiro, as cited in Brooks, & Geer, 2007).
All in all, this chapter has established that (1) there is a lack of comprehensive understanding of incivility, that (2) incivility is mostly studied from an American angle, that (3) we are not certain how much incivility is out there, and (4) if it actually is a problem (e.g., Coe et al., 2014; Gervais, 2015; Sobieraj, & Berry, 2011). Therefore, there are legitimate concerns and motivations to research this concept in Europe by means of a quantitative content analysis and ascertain whether a civility crisis has arrived in Europe as well (e.g., Jamieson, 1999; Mutz, & Reeves, 2005).
Determinants of Incivility in User Comments
Scholars have suggested that incivility is linked to computer-mediated communication (e.g., Thompsen & Foulger, 1996), as several characteristics of computer-mediated communication, such as the lack of filtering of non-verbal cues, cause citizens to behave in an uncivil manner (Siegel, Dubrovsky, Kiesler, & Mcguire, 1986). Consequently, I will now delve deeper into when and why incivility should occur in user comment threads, which connects to how much incivility there is online, and whether it has negative effects on the quality and nature of political discussions.
Structural Reasons for Incivility Online
A number of reasons for incivility can be explained by calling on the so-called online disinhibition effect (Suler, 2004). This effect posits that anonymity online causes people to seek for websites or forums on the Internet of which they would never attend a real life equivalent. This can enable them to be rude, express hatred, or even threaten other users, because they are freer in their expressions and thoughts than they would be in an offline setting (Suler, 2004). The Social Identity and DE-individuation (SIDE) theory combines anonymity with the importance of social context in computer-mediated communication (Postmes, Spears, & Lea, 1998). Due to the lack of non-verbal social cues, such as facial expressions, and due to factors such as anonymity, people are more likely to disseminate
contentious ideas, as an anonymous comment cannot be traced and held accountable (Neurauter-Kessels, 2011; Spears & Lea, 1994). .
Aside from anonymity, the asynchronicity of the web is another important factor of the online disinhibition effect. This factor enables people to postpone their reaction, which in turn, causes people to disinhibit, as they do not have to face someone’s immediate reaction. This is connected to dissociative imagination, a factor that causes people to become aware of the possibility of easily ending a conversation or thread when things get out of hand. Another important aspect is the minimization of status and authority. Online conversations lack social cues that would normally determine which person ranks the highest, and consequently causes all people to be equal online, which in turn enables them to reduce inhibitions during
discussions (Suler, 2004).
Lastly, the Uses and Gratifications Theory (UGT) supplies us with knowledge about incivility from a motivational angle, and posits the idea that “people use certain types of media to satisfy their needs” (Alonzo & Aiken, 2004, p.206). The theory postulates that the use of specific media types is connected with a set of psychological motives (Blumler, 1979; Conway, & Rubin, 1991; Katz, Blumler, & Gurevitch, as cited in Alonzo & Aiken, 2004). When McGuire examined flaming, he discovered that stimulation, tension reduction, expression, and assertion were the most prevalent and relevant human motivation theories, and consequently argued that individuals behave in an uncivil manner to fulfill certain needs. Stimulation, for instance, can imply that individuals aim to escape from a state of boredom, and thus seek for sensation by behaving inappropriately (McGuire, as cited in Alonzo & Aiken, 2004).
All these theories enlighten us by explaining when and why users are uncivil online, and indicate that we can expect a great deal of incivility online. Online news websites’ comment threads contain important factors such as anonymity, asynchronicity, and a lack of
non-verbal social cues (e.g., NUjij, 2015), which can influence or even cause online incivility. Moreover, when Kayany (1998) examined four social Usenet newsgroups, he discovered that political context influenced the presence and amount of incivility in online discussions. Perhaps, this finding can also be extended to comment threads. The theories suggest that there should be incivility in Europe, and as such, incivility will not differ between Europe and the United States. However, American media and politics are more polarized (e.g., Arceneaux, Johnson, & Murphy, 2012; Levendusky, 2013), whereas European media and politics are much more populist (Albertazzi, & McDonnell, 2008; Mazzoleni, 2008), which could suggest that, even though the online environment supports the occurrence of incivility (Suler, 2004), incivility will perhaps differ in the amount or its form, as people and media cope differently with politics. Consequently, my first research question is:
To what extent can we find incivility in news websites’ political comment threads (RQ1)?
Relational Effects of Incivility Online
A second important aspect, apart from the online environment creating incivility, is whether incivility really is nothing more than empty insults, and thus a danger to public discussions. Reasoned arguing is said to be an important aspect in societies, and therefore discussions, as it shapes public opinions (Habermas, as cited in Stromer-Galley, 2007). Reasoned or meaningful arguments contain assertions that “are grounded in empirically verifiable evidence or in shared understanding of moral or normative behavior” (Stomer-Galley, 2007; p.4). Consequently, in order for an uncivil message to be considered
meaningful, it must contain evidence, such as facts, numbers (Coe et al., 2014), reasons, or examples, and not be a mere summation of uncivil words. As such, people should not say “Wilders is an idiot”, but would rather say “Wilders is an idiot, because he destroyed our immigration policy”. Interestingly, Coe et al. (2014) discovered that uncivil comments
contained more evidence than civil comments. This implies that even though people behave in uncivil manners towards one another, the nature and quality of the discussion is not
necessarily endangered by the use of incivility, as people give the impression that they care enough about the discussion to back up their uncivil statements with meaningful
contributions. This is also what De Vreese and Semetko (2002) suggested in their study on political cynicism. Despite the presence of cynical or negative attitudes among citizens, citizens are still aware that it is important to participate in politics, and thus political discussions, in a meaningful manner. Therefore, I believe it is of utmost importance to ascertain whether meaningful arguments occur more often than we think in uncivil discussions. Consequently, my first hypothesis is as follows:
H1: Incivility in user comments will not be meaningless, but will also contain meaningful arguments.
Another point relates to the question, whether the occurrence of incivility will demobilize and disengage citizens in political discussions. Similar to reasoned arguing, engagement is considered an important aspect of discussions and deliberation (Stromer-Galley, 2007), as it indicates whether users exchange thoughts with others. Engagement refers to users engaging in a discussion by commenting underneath articles or videos, or replying to comments of other users (Coe et al., 2014; Ksiazek et al., 2014; Stromer-Galley, 2007). Consequently, scholars, such as Ksiazek et al. (2014), have studied YouTube videos’ comment threads, and questioned the relation between engagement and incivility. They examined whether popular engaging YouTube videos would contain more contentious comment threads, and measured engagement by counting the total number of comments underneath any given video. In addition, they reported the ‘ratio of replies to total comments’ in order to indicate to what extent users exchanged thoughts with one another. Findings suggest that more engaging videos can cause uncivil discussion. One could argue that this is
an unwanted causal relationship, as discussions between users should not spark incivility. However, in the previous section we have established that the presence of incivility is often not meaningless (Coe et al., 20214). Therefore, I do not believe that the presence of a possible causal relationship between engagement and incivility is of major concern to the academic community. This does raise the question, however, whether these findings extend to other online interactions, such as user comment threads on newspaper websites. Given the fact that comment threads underneath YouTube videos and comment threads underneath news
websites’ articles both allow users to comment on one another (e.g., Ksiazek et al., 2014; NUjij, 2015), it would appear logical to expect that engagement can also cause incivility in user comments. Consequently, I hypothesize:
H2: User comments in which users interact with one another, will support user comments displaying incivility.
Online Incivility in Europe: Differences and Similarities
Most of the research on incivility derives from an American setting (e.g., Coe et al., 2014; Ksiazek et al., 2014), which is why this study focuses on incivility from a European standpoint by examining the Netherlands and Belgium. However, can we expect to find incivility in the European online landscape to the same extent as in the United States, and do possible causes of incivility in the United States extend to the European landscape?
Online incivility most likely occurs in Europe as well, as incivility is linked to the presence of computer-mediated communication in general (e.g., Thompsen, & Foulger, 1996). Therefore, we might not even expect any differences between incivility in Europe and the United States, as factors such as anonymity, asynchronicity, and dissociative imagination are part of the online communication environment, and thus are also factors that play a role in European online communication (Suler, 2004).
There are, however, also several other, more macro-level, reasons why the presence of incivility would not differ between the United States and Europe. First, some scholars argue that the Netherlands are adapting to media modes, styles, and cultures of the United States (Brants, & Van Praag, 2006; Mazzoleni, 1987; Pfetsch, 2001; Semetko, Blumler, Gurevitch, & Weaver as cited in Huyn, 2012). This adaptation of European media and their political communication systems to the American setting, could translate into, for instance, the expression of opposing views in uncivil manners. Consequently, similar uncivil outcomes could be encountered in Europe (Mutz, 2007; Mutz, & Reeves, 2005). Second, the spiral of
silence theory posits that in collectivistic cultures, such as China, citizens avoid the violation of social consensus in order to prevent isolation. Therefore, citizens constantly examine what the majority thinks about controversial issues, and then decide whether they disseminate their opinion or not (Hofstede, as cited in Dalisay, 2012; Noelle-Neumann, 1974, 1977). On the contrary, individualistic cultures, such as the United States, and most of Europe (including the Netherlands and Belgium), prefer direct assertions of personal opinions (Dalisay, 2012; Noelle-Neumann, 1974, 1977). This cultural perspective in combination with the fact that American media styles are being adapted in the European landscape (Brants, & Van Praag, 2006), could suggest that both Europe and the United States have no problem expressing divergent views, which could spark uncivil reactions that are disseminated without hesitation (Arceneaux et al., 2012; Dalisay, 2012; Noelle-Neumann, 1974, 1977). Lastly, Walter (2012) argues that political systems do not play an important part in regard to the presence of
negativity, and thus perhaps incivility, as negativity occurs regardless of a countries’ party system.
Even though the previously discussed literature suggests that the presence of incivility will not necessarily differ between Europe and the United States, the antecedents of the presence of online incivility might be different.
First, the United States, partisan media, and American politics are increasingly polarized (Levendusky, 2013), while this is not necessarily the case in Europe (e.g., Keman, & Krouwel, 2006; Mazzoleni, 2008). As a consequence, Arceneaux et al.’s (2012) findings suggest that when American citizens were exposed to counterattitudinal news programs, they were more likely to adopt hostile attitudes and perceptions towards news programs, which may cause polarization of the electorate. Moreover, the combination of these hostile attitudes and factors such as anonymity of the web (Suler, 2004) could stimulate the propensity of citizens to be rude online. Given that Europe is a lot less polarized, we could expect less incivility in Europe. Even so, this argument is weakened by the fact that the European media and political landscape offers more suitable ground for populism, as voters become detached within party systems, and work practices of Western media change (Albertazzi, &
McDonnell, 2008). Populism, in that sense, could perhaps be the European cause of incivility. Over the past decades, far-left and far-right populist parties have gained influence in
European countries, such as the Netherlands and Belgium, and have made provocative issues, such as fear towards foreigners and/or immigrants, steady issues on their political agendas. In turn, these issues can create or reinforce existing feelings of dissatisfaction, political anxiety or distrust among the public, and may spark uncivil online behavior (Arceneaux et al., 2012; De Lange, & Akkerman, 2012; Suler, 2004).
In sum, we can assume that European countries might have different causes for incivility than the United States, but that incivility will occur nonetheless. However, this does raise the important question of whether incivility will differ within European countries themselves. So far, no study exists that begins to explore intra-European differences in online incivility. Consequently, this study examines two political news websites and their comment threads from the Netherlands and Dutch-speaking Belgium (Flanders) (Hooghe, 2005). I have specifically chosen to study two countries that use the same language on their news websites
in order to be certain that possible differences in incivility do not derive from linguistic aspects. Therefore, my last research question is:
To what extent does incivility differ between a Dutch and a Flemish news website and their comment threads? (RQ2).
Method Design
Due to the fact that a vast amount of research focused on effects of incivility in
absence of a proper overview of how much incivility there is in media discourses (e.g., Coe et al., 2014; Fridkin, & Kenney, 2008; Mutz, 2007; Mutz, & Reeves, 2005), it appears a logical choice to conduct a quantitative content analysis in order to address this gap in the literature. The lack of European research on incivility has caused me to examine two news websites of two European countries (NU.nl and HLN.be) that show great similarities in political
communication systems, and consequently ask whether they differ (e.g., Coe et al., 2014; Hooghe, 2005; Ksiazek et al., 2014). Therefore, my units of analyses are political articles that are published on news websites NU.nl and HLN.be. Per political news article, I will examine both the article and the attached user comment threads.
I have chosen to investigate user comments, because these online communication forms can entail interactive audience participation by allowing users to comment, and engage in discussions with other users (Neurauter-Kessels, 2011; Rowe, 2015). According to
Neurauter-Kessels (2011), articles that contain user comments are usually more controversial and thus, may also contain incivility. In addition, user comment threads are very important, as they can polarize people who passively read user comments (Anderson et al., 2014; Smith et al., 2009), and affect journalistic production in the sense that journalists read user comments, and consequently refuse to write certain pieces to prevent uncivil reactions from occurring online (Reader, 2012).
Sample
The sample consists of political news articles and their user comment threads published on the Dutch news website NU.nl (n = 178), and the Belgian newspaper website HLN.be (n = 123). NU.nl is part of Sanoma Media Netherlands Groep and can be characterized as an online news platform, which publishes news on a 24/7 basis (Sanoma Media Netherlands B.V., 2015). HLN.be is an online version of the print newspaper Het Laatste Nieuws, and is part of the Persgroep (De Persgroep, 2015). All articles that contained at least one user comment were accessed and collected between March 5, 2015 and March 26, 2015. Due to the limited time frame of this study, I have chosen to analyze the comment threads for three consecutive weeks, in order to get the best possible insight in incivility. Therefore, my units of analyses (Krippendorff, 2013) are the articles published on NU.nl and HLN.be (N = 301). In order to avoid comparing articles that contained 180 comments to articles that contained 2 comments, the number of times incivility occurs in articles’ comment threads was counted to give an idea of variance. In total, 301 articles and 26,911 user comments were collected. Due to time restrictions, only the first 50 comments of each article were coded and analyzed. Nonetheless, these first 50 should be sufficient, as they still give an appropriate overview of the extent and nature of incivility. Please consult Appendix A for examples of articles and comment threads used in the content analysis.
An online form allows users to submit their comments, which then posts them below current news articles. One important caveat of studying incivility in user comments is that newspapers or news websites establish house rules in order to discourage the submission of uncivil comments (Neurauter-Kessels, 2011). Other manners to prevent online uncivil discourse involve active moderation by digital media personnel, the use of automated screening algorithms, or moderation by users (Ksiazek et al., 2014). Comments are either placed immediately after submission, and are then reviewed by news websites for uncivil content, or the news website first screens the comment for incivility prior to publication
(Neurauter-Kessels, 2011; Soni, 2013). Also, users can be encouraged to report inappropriate comments by ‘flagging’ them (Ksiazek et al., 2014). However, research shows that moderation policies do not necessarily prevent the occurrence of incivility (Coe et al., 2014).
NU.nl and HLN.be both have moderation policies in regard to user comments. In order to comment on articles, one has to give out an email address. Consequently, users on NU.nl most obey guidelines, which basically revolve around using appropriate language that is focused on the subject of the article at hand, and not revealing any personal information when posting comments. Inappropriate behavior can be reported to NU.nl by pressing the ‘X Nuniet’ button (‘Not Now button’) on a given reaction (NUjij, 2011).
Although NU.nl and HLN.be are not exactly the same, they could be considered cross-country equivalents from one another. Similar to NU.nl, HLN.be allows users to comment on articles (e.g., HLN.be, 2015; NUjij, 2015). In addition, HLN.be is one of the most popular newspapers in Flanders (De Persgroep, 2015), as its newspaper and website can count on 1.6 million daily visitors (“HLN bereikt”, 2014), whereas NU.nl is the most popular news site in the Netherlands, which can count on 2.5 million visitors on the website and its apps (Sanoma Media Netherlands B.V., 2015). Its guidelines are also greatly similar to the guidelines from NU.nl (HLN.be, 2015; NUjij, 2011).
Codebook
The codebook was developed by means of quantitative pre-testing which determined which categories and dimensions should be included. Answers were collected in an online questionnaire on Qualtrics, in order to make coding easier and faster. Consequently, the date, news website, headlines, length of article, type of article, number of comments, issues, actors, (type) incivility in article, (type) incivility in comments, frames, number of deleted comments, amount incivility in comments, (amount) evidence in comments, and type and amount of engagement were registered of each unit (Krippendorff, 2013). I will shortly go over how the most important variables were measured per research question or hypothesis.
If incivility was present in the comment thread, the coder had to go over some very crucial variables. The main variables in my thesis concern ‘type of incivility’, ‘amount of incivility’, ‘amount of meaningfulness’, ‘engagement’, ‘conflict frame’, ‘issue(s) in article’, and ‘incivility in article’. Please consult Appendix B for the codebook and all its variables.
For the first research question, which asked to what extent there is incivility, I used several variables. The first variable concerns ‘type of incivility’. Different types of incivility were established in order to examine what kind of incivility the user comments contained. Per type it was indicated whether that type was present (1) or absent (2). Table 1 shows the different types of incivility, and several examples that were found in the sample. It also shows, where in the literature the different aspects of incivility were taken from. These specific types derive from the perspectives and research of the scholars on which my broader definition of incivility was founded upon.
Table 1
Types of incivility and their examples
Type Description Example from NU.nl and HLN.be Based on
Name-calling
Words directed at groups of people, a person, plan, policy, or behavior that are insulting, mean-spirited, or disparaging.
Intussen hebben PVD/CDA/VVD/D’66 alles in het werk gesteld het allemaal naar de knoppen te helpen. En nu er stemmen moeten worden binnen gehaald schreeuwen en liegen ze weer over elkaar heen!
Coe et al. (2014)
Lying Giving the impression that a policy, plan, or idea was deceitful, or dishonest.
Beetje ongeloofwaardig dat het 'CDA' dit opeens wil 3 dagen nadat het in de pers verscheen. De mailtjes die ze daarvoor gekregen hadden zullen wel op zn Ivo
Opstelstens 'verdwenen' zijn, of zich niet meer kunnen 'herinneren' .. Coe et al. (2014) Vulgarity/ obscene language
The use of language or profanity that would not be considered appropriate and/or professional in public and/or professional
discourse.
Verdorie, shit Coe et al.
(2014)
Racist remarks
Disseminating racist or bigoted feelings/sentiments
Asscher heeft gelijk, als er een artikel komt over dat er drie donkere jongens een agent aanvallen, zit hij meteen te twitteren dat ze het land uit moeten, maar als er in de polder een jongen een agent moedwillig aanrijd dan blijft hij stil. Hij zit constant de boel op re ruien. Een vreselijke man.
Santana (2014)
Threats The expression of having the desire or intention to hurt someone physically or mentally.
Hier zijn we weer met het belerende vingertje! Wat ik doe met mijn lichaam is mijn eigen zaak en als iemand ook maar durft mij daarvoor aan de schandpaal te nagelen… ik vuur terug met legale papieren aan de eerste die het probeert.
Santana (2014) Xenophobic language Blaming and/or be scared/anxious (about) foreigners
Als het om Haat-gangers (Jihadisten) gaat ben ik heel rechts. Laat ze verrekken die monsters
Santana (2014) Stereotypes Assigning stereotypes to
another person, or groups of people, in the discussion or in general
Typische reactie van iemand in loondienst die geen idee heeft van de (al bestaande)
verplichtingen van werkgevers. Heel sneu dit.
Papacharissi (2004) & Santana (2014) Capitalized comments Comments in which discussants unnecessarily capitalize part or all of their comment to symbolize shouting or to emphasize a certain aspect of the discussion
Zou je eens moeten kijken hoe lang de OVERHEID doet over betalingen…
Another important variable is ‘amount of incivility’. This variable recorded the portion of incivility that the first 50 comments of an article contained, and thus asked for the overall amount of incivility in an article’s comment thread (Sobieraj, & Berry, 2011). Answer categories varied in steps of 5, ranging from ‘less than 5, 5 - <10, 10 - <15, 15- < 20’ to ‘20 and over’.
The following variables were established in order to ascertain whether incivility correlates with conflicting frames, issues in articles, or incivility in article:
‘Conflict frame’ consisted of two questions, which tapped into the presence of disagreement between parties, individuals, groups or countries in the article (yes/no), and whether a party, individual, group or country reproached one another in the article (yes/no) (De Vreese, 2001). Both questions were computed together as conflict frame (yes/no).
The variable ‘issue(s) in article’ (Bos, Van Der Brug, De Vreese, 2010) asked what the main issue of a given article was, and was consequently entered in an open entry field in Qualtrics. The issues ranged from political topics, such as election campaigns, to social affairs, such as emancipation of females. Please consult Appendix B for the full range of categories.
‘Incivility in article’ questioned whether incivility was present in the article (absent/present).
The variables amount of incivility and ‘amount of meaningfulness’ were used for the first hypothesis. The variable ‘amount of meaningfulness’ (Coe et al., 2014; Sobieraj, & Berry, 2011) asked how many of the uncivil comments contained meaningful arguments and/or statements. This ranged from ‘less than 5, 5 - < 10, 10 - <15, 15 - <20,’ to ’20 and over’. Meaningful arguments and/or statements imply that the user does not merely comment something along the lines of ‘’Wilder is an idiot”, but would rather say “Wilders is an idiot, because he destroyed our integration policy”. Therefore, in a meaningful uncivil argument
and/or statement people give reasons for certain topics, provide thoughts, provide
background, relate to own experiences, or provide facts/figures/examples, as long as it is not a mere summation of uncivil words (Coe et al., 2014; Stromer-Galley, 2007).
For the second hypothesis I used the variables amount of incivility and ‘engagement’ in order to ascertain whether a correlation existed between the two. The variable
‘engagement’ (Coe et al., 2014; Ksiazek et al., 2014; Stromer-Galley, 2007) pertains to users replying to one another. The coder first stated whether engagement was present (1) or absent (2), and then indicated whether the uncivil comment directly replied to someone else’s comment (users reply by using the ‘’@’’ sign, or address another user directly by using his or her name), or whether the comment indirectly replied to another/previous comment without mentioning a name. Per engagement type the coder stated whether that type was present (1) or absent (2).
Lastly, for the second research question, I used the variables amount of incivility and the two news websites in order to examine whether there were differences in incivility between the two.
Intercoder-reliability
A second coder from the Political Communication program was asked to code 10 percent of the sample, in order to calculate intercoder reliability by means of Krippendorff’s alpha. The first test did not yield satisfactory results; therefore the codebook was adapted and another 8 percent of the sample was coded in order to reach sufficient outcomes. Overall, Krippendorff’s alphas varied from 0.62 to 1.00, and were thus satisfactory. Please consult Appendix C for all scores.
Coding Procedure
The coder read the entire article of the particular website and its first 50 comments. Consequently, the coder first answered some identification questions, such as the article ID number, the coder ID, which news website the article and comment thread derived from, the
date, the number of words and the headline of the article, the number of comments, and the type of article (ranging from news agency reports, news reports, editorial, third party opinion article, interview to others). Then, the coder answered one of the most vital questions:
whether any type of incivility occurred in the comment thread. If this was not the case, the coder was given instructions to stop coding.
Analysis
In order to answer all research questions and hypothesis, statistical analyses are conducted using SPSS. For all research questions and hypotheses, basic descriptive statistics are given to examine how often incivility, meaningful arguments, and engagement occurred respectively. Furthermore, a linear regression analysis is conducted to ascertain whether any variables can predict the amount of incivility in user comment threads (RQ1). For H1, a linear regression is executed, in order to examine if the amount of meaningful arguments and/or statements can be predicted by the amount of incivility. H2 also contains a linear regression to ascertain if engagement predicts the amount of incivility in user comments. Finally, for
research question 2, a test of normality was conducted by means of a Shapiro-Wilk test, which revealed that the dependent variable amount of incivility was not normally distributed for both NU.nl and HLN.be, p < .001. Therefore, a Mann-Whitney U test is performed to examine whether the amount of incivility differs between the two news outlets.
Results
Overall, only 11.6 % (n = 35) of the total sample (N = 301) contained incivility in the articles themselves. The most covered topics in these articles revolved around anything directly related to politics, such as election campaigns, disputes between politicians, and the resignation of politicians (n = 71, 23.6%), and security, police, justice, crimes, and national terrorism (n = 41, 13.6%). Politicians (n = 162, 53.8%) and political parties (n = 40, 13.3 %) were the most visible actors in the analyzed news articles.
To What Extent is There Incivility in User Comment Threads?
The first research question this content analysis tries to answer is to what extent there is incivility in user comment threads (RQ1). It appears that out of all political articles and their comment threads (N=301), 92.7 % contained some sort of incivility in their comment threads (n = 279). This is an interestingly high amount, as it implies that almost all comment threads contained at least some type of incivility. Consequently, I examined how much incivility occurs in these comment threads. Table 2 illustrates that out of all articles and their comments threads (N=301), only 26.2 % of the sample contained less than 5 uncivil
comments (n = 79). However, 52 comment threads (17.3 %) contained 20 and more instances of incivility, which suggests a high amount of incivility.
Table 2
Amount of incivility in user comment threads
Total uncivil comments % of Uncivil comments
Less than 5 79 26.2 5 - < 10 49 16.3 10 - < 15 54 17.9 15 - < 20 45 15.0 20 and over 52 17.3 Missing 22 7.3 Total N = 301 100
I was also interested in which types of incivility were most prominent in online user comments. As described in the method section, I tested several types of incivility: name-calling, lying, vulgarity, racist remarks, threats, xenophobic language, stereotypes, and capitalized comments. Consequently, Figure 1 shows that name-calling, words directed at groups of people, a person, plan, policy, or behavior that are insulting, mean-spirited, or disparaging (Coe et al., 2014), was by far the most common type of incivility, as it occurred in 88.7 % of comment threads. An example of name-calling could be “the security system of this country is ridiculous”. The second most prominent incivility was lying (61.1 %, n = 184), which is a type of incivility that gives the impression that a policy, plan, or idea was deceitful or dishonest (Coe et al., 2014). An example of lying is, for instance, “it is very suspicious that
the CDA all of a sudden opts for this plan, three days after it has been released in the media”. The third most prominent type were capitalized comments (49.5 %, n = 149), which are comments in which discussants unnecessarily capitalize part or all of their comments to symbolize shouting or to emphasize a certain aspect of the discussion. An example of this type of incivility is for instance, “no, that groups feels as if it IS OFFENDED!”
Figure 1. Percentages of types of incivility in user comment threads. This figure illustrates the
percentages of which types of incivility were most prominent in political news articles’ user comment threads (N = 301).
I was also interested to find out what predicts the amount of incivility in user comment threads. Therefore, I conducted a linear regression analysis to ascertain if the issue(s) in the coded news article, the presence of a conflict frame, and the presence of incivility in the article(s) could predict whether user comment threads contain more or less incivility. I chose to include issue(s) in article, because Kayany (1998) shows that incivility most often occurred in a political context, which raises the concern that discussions about explicitly political issues
0 50 100 150 200 250 300 88.7% 61.1% 43.5% 1.7% 0.7% 6% 37.9% 49.5% Name-calling Lying Vulgarity/ obscene language Racist remarks Threats Xenophobic language Stereotypes Capitalized comments
Percentages of types of incivility in comment threads
will contain greater amounts of incivility. Therefore, I made two dummy variables out of this variable: a political one (conflict issue), versus a social one (positive issue). The conflict issue that I chose for this analysis is anything directly related to politics and election campaigns, as De Vreese (2001) suggests in his study on framing politics that journalists often framed political news in the Netherlands in terms of conflict. I chose emancipation as a social issue, as Bos, Van Der Brug, and De Vreese (2010) consider emancipation of for instance, women, as a social affairs issue (please consult Appendix B for the codebook). I believe emancipation can serve as a positive opposition to the conflicted political issue, as political issues might spark more conflicted and perhaps more uncivil debates than social issues (Neurauter-Kessels, 2011).
In addition, I chose the variable conflict frame. Conflicting articles might predict the amount of incivility, as articles that attract a lot of user comments are usually more
conflicting, and thus may also contain incivility (Neurauter-Kessels, 2011). Lastly, I included incivility in article, because studies suggest that exposure to uncivil media might translate to incivility in user comments, as users can adopt the uncivil language used in online articles (Gervais, 2014; 2015).
The linear regression model with the amount of incivility in user comment threads as a dependent variable and conflict issue, positive issue, conflict frame, and incivility in article as independent variables yielded significant results, F(4, 274) = 5.35, p < .001. The regression model is suitable to predict the amount of incivility in user comments, however, the fit of the model is weak: 7 % of the variation in the amount of incivility can be predicted based upon issue(s) in article, conflict frame, and incivility in article (R2 = .007). Yet, we have to
acknowledge that the importance of the predictor variables should not be dismissed based on a weak R2, as the model contains few predictor variables and a complex dependent variable. However, conflict frame, b* = .04, t = .58, p = .561, 95% CI [-.92, 1.69], incivility in article
b* = .11, t = 1.84, p = .067, 95% CI [-.03, 1.62], and positive issue, b* = .08, t = 1.35, p = .177, 95% CI [-.40, 2.17], did not add statistically significantly to the prediction. Only conflict issue, b* = .22, t = 3.72, p < .001, 95% CI [.35, 1.14] has a significant, yet weak, correlation with the amount of incivility. If a conflict issue increases with one unit, the amount of incivility will increase with .74.
Tests to see if the data met the assumption of collinearity indicated that
multicollinearity was not a concern (Conflict frame, Tolerance = .95, VIF = 1.05; Incivility in article, Tolerance = .93, VIF = 1.07; Conflict Issue, Tolerance = .97, VIF = 1.03, Positive Issue, Tolerance = .98, VIF = 1.02). However, a Shapiro-Wilk Test ascertained that the dependent variable amount of incivility was often not normally distributed for all levels of my independent variables. Conflict frame: absent, p < .001, present, p = .377, incivility in article: absent, and present p < .001, conflict issue: absent, and politics p < .001, positive issue: absent, p < .001, and emancipation, p = .032, which implies that results must be interpreted with caution.
In sum, the answer to the first research question is that almost all user comments threads contained some type of incivility, the amount of incivility is quite substantial, and that users interestingly use name-calling, lying, and capitalized comments in most instances. In addition, we can most likely predict the amount of incivility in comment threads by conflict issues in articles.
Can Incivility Also Be Meaningful?
My first hypothesis stated that uncivil user comments will not be meaningless, but will also contain meaningful arguments (H1). Out of all articles and user comment threads (N =301), 71.8 % (n = 216) of uncivil comments contained meaningful arguments and/or statements. The majority of times (57.5 %, n = 173), I counted less than 5 meaningful arguments and/or statements in the comment threads. However, 14.3 % (n = 43) contained
between 5 and 10 meaningful arguments and/or statements, which implies that quite a lot of uncivil comments contained meaningful arguments and/or statements. For this hypothesis, a linear regression was conducted to predict the amount of the dependent variable meaningful arguments and/or statements by the independent variable amount of incivility. The analysis revealed that the linear regression model was statistically significant, F(1, 214) = 19.62, p < .001. This implies that the amount of incivility in user comments can predict how many meaningful arguments and/or statements there will appear in comment threads. The fit of the model is weak: 8 % of the variation in the amount of meaningful arguments and/or statements can be predicted on the basis of the amount of incivility (R2 = .08). However, the importance of the predictor variable should not be dismissed, as the model only contains one predictor variable. Amount of incivility, b* = .29, t = 4.43, p < .001, 95% CI [.05, .12], has a weak correlation with the amount of meaningful arguments and/or statements. The amount of meaningful arguments and/or statements increased 0.08 when there was more incivility. Overall, hypothesis 2 is confirmed. However, caution must apply as a Shapiro-Wilk test ascertained that the dependent variable amount of meaningful arguments and/or statements was not normally distributed for all levels of my independent variable amount of incivility, p < .001.
Does Engagement Predict Incivility?
My second hypothesis stated that comments in which users interact with one another will cause incivility. Approximately 50 % (n = 151) of user comments contained some form of engagement between users. Direct engagement, thus replying to others’ user comments by addressing names and/or using the “@”-sign, occurred 41 % of times (n = 124) in uncivil comment threads, whereas indirect engagement, replying without any names or “@”-signs to other users, occurred in about 19 % (n= 56) of threads. Table 3 illustrates that 40 % (n = 119) of the total sample involved engagement that occurred less than 5 times in comment threads.
Table 3
Frequencies and Percentages of the occurrence of interaction between users Total engagement in uncivil comments % of interaction
None 0 0 Less than 5 119 39.5 5 - < 10 30 10 10 - < 15 2 .7 15 - < 20 0 0 20 and over 0 0 Missing 151 49.8 Total N = 301 100
In addition, a linear regression showed that engagement (present/absent) statistically significantly predicted the amount of incivility in user comment threads, F(1, 277) = 94.64, p < .001. This implies that engagement between users can predict how much incivility there will appear in user comment threads, and the fit of the model is relatively strong for only one predictor variable: 26 % of the variation in the amount of incivility can be predicted on the basis of engagement (R2 = .26). Engagement, b* = .51, t = 9.73, p < .001, 95% CI [1.19, 1.79], has a moderate correlation with amount of incivility. The amount of incivility increased 1.49, for each user comment that contained engagement. Overall, hypothesis 2 is confirmed. However, caution must apply as a Shapiro-Wilk test ascertained that the dependent variable amount of incivility was not normally distributed for all levels of my independent variable engagement, p < .001.
Does The Amount of Incivility Differ Between News Websites?
The last research question asked to what extent incivility differs between a Dutch and a Flemish news website and their comment threads (RQ2). Firstly, descriptive statistics showed that 96 % (n = 171) of articles from NU.nl (N =178) contained uncivil user comment threads, whereas 88 % (n = 108) of articles from HLN.be (N = 123) contained uncivil user comment threads. A test of normality was conducted by means of a Shapiro-Wilk test, which revealed that the dependent variable amount of incivility was not normally distributed for both NU.nl and HLN.be, p < .001. Therefore, a Mann-Whitney U test was conducted, which
compared the amount of incivility between the two news websites (NU.nl/HLN.be), and revealed no significant difference between the amount of incivility in NU.nl (mean rank 144.24), and HLN.be (mean rank 133.29), (U = 8509.50, p = .258). Consequently, we can conclude that there is no systematic difference in the amount of incivility in user comments threads between NU.nl and HLN.be
Discussion
This study aimed to shed light on the current state of incivility in the European online political landscape. It investigated the political news articles’ comment threads of the Dutch news website NU.nl, and the Flemish website HLN.be, and ascertained to what extent there is incivility, if there are any relational factors that influence it, and whether there are differences in incivility between the two websites. The results reveal that almost all user comment threads contained some type of incivility, the amount of incivility is quite substantial, and that users interestingly prefer to use name-calling, lying, and capitalized comments when they engage in uncivil behavior. The analysis also showed that conflict issues, which in this case was
anything directly related to politics and election campaigns, predict the amount of incivility in comment threads. The analysis also revealed that the amount of incivility in user comments could be predicted by engagement between users. In addition, the amount of incivility in user comments can predict the amount of meaningful arguments and/or statements in user
comments. Lastly, the analysis for the last research question showed that there is indeed no systematic difference in the amount of incivility in user comment threads between NU.nl and HLN.be.
Implications
The results of this Master’s thesis contribute to the existing literature about incivility in the sense that they provide the first European insight of incivility, they explore whether we should be worried about a civility crisis in Europe, and how incivility is constituted in the
online political landscape. We have ascertained that incivility is indeed a major part of Dutch and Flemish discourse, at least for the news outlets NU.nl and HLN.be. In that sense, this European study differs from its American equivalents (e.g., Coe et al., 2014; Ksiazek et al., 2014), as it encountered much more incivility.
However, the presence of more incivility might not necessarily be a problem, as the most interesting finding of this Master’s thesis is the fact that quite a substantial amount of meaningful arguments and/or statements occurs in uncivil comment threads. This amount is much higher than American findings suggested (e.g., Coe et al., 2014). This implies that when people do ‘act out’, it does not merely revolve around name-calling, and lying to one another. Rather, people use facts, statistics, reasoning and arguments to back up their uncivil thoughts. Moreover, hypothesis 1 ascertained that the amount of incivility could predict the amount of meaningfulness. This illustrates that academic concerns about a so-called ‘civility crisis’ (e.g., Jamieson, 1999; Mutz, & Reeves, 2005) might be a bit overstated. Rather, the results show that we might not have to be that worried. Yes, people do call each other names, and are more likely to be uncivil when they engage with one another, yet often opt to back up their uncivil claims with meaningful arguments and/or statements. Perhaps, we could state that incivility might not be an academic problem, but is moreover a societal problem, as public life desires respectful and civil conversation with one another (Schudson, 1997), and might be blinded by the negative effects of incivility.
Lastly, the findings suggest that the amount of incivility in user comments can be predicted by engagement between users. Although this finding concerns user comment threads underneath news articles, it is still in line with the results of Ksiazek et al. (2014), as they discovered that engaging YouTube videos contain more uncivil discussions.
Limitations and Outlook
Naturally, this study has some limitations. To begin, due to its limited time frame its sample was relatively small. I was only capable of examining two news websites over the course of three consecutive weeks. As such, it is difficult to generalize results, as two news outlets cannot be representative of two countries as a whole. Moreover, the comparative aspect of this thesis is limited, as both countries are quite similar in their political background. Consequently, future research should aim to raise the external validity of these European results on incivility, and should therefore study more news websites from a given country, over a longer period of time. In addition, different news outlets from multiple countries that vary in political (communication) systems should be examined in order to yield more accurate and comparable results. Third, only the first 50 comments of each article were coded, whereas future studies might choose to code all comments. Fourth, most data was not normally
distributed, which implicates that results must be interpreted with caution. Also, the news articles deriving from NU.nl covered an important Dutch election when I gathered my data. This could influence the interpretation of my results, and causes an imbalance, as HLN.be did not cover any recent elections. Lastly, I have not devoted much time into analyzing a possible relationship between incivility by journalists in news articles, and incivility in user comments, whereas other studies might opt to do so in order to examine if and how users are specifically influenced by journalists’ language.
In sum, the main findings of this study suggest that the term ‘civility crisis’ might be overstating the problem. There is incivility, but incivility often is meaningful. Future research should therefore put great effort into examining if incivility is a problem for our modern societies, from an academic or societal standpoint. Scholars could conduct more content analyses, and should opt to look more into the quality and nature of uncivil discussions, by examining engagement, and meaningfulness in regard to incivility. If research ascertains that
incivility is most of the time not merely meaningless, that people still discuss with each other politically despite incivility, and establishes that incivility might not be that destructive to our modern societies, this might implicate that Van Stokkum’s (2010) concern only applies to societies’ perspectives on public morality, and therefore, incivility should not be a serious concern to the academic community.
References
Albertazzi, D., & McDonnell, D. (2008). Twenty-First Century Populism: The Spectre of Western European Democracy. Basingstoke, United Kingdom: Palgrave Macmillan. Alonzo, M., & Aiken, M. (2004). Flaming in electronic communication. Decision Support
Systems, 36(3), 205–213. http://doi.org/10.1016/S0167-9236(02)00190-2
Anderson, A. A., Brossard, D., Scheufele, D. A., Xenos, M. A., & Ladwig, P. (2014). The “Nasty Effect:” Online Incivility and Risk Perceptions of Emerging Technologies. Journal of Computer-Mediated Communication, 19(3), 373–387.
http://doi.org/10.1111/jcc4.12009
Ansolabehere, S., Iyengar, S., Simon, A., & Valentino, N. (1994). Does Attack Advertising Demobilize the Electorate? The American Political Science Review, 88(4), 829–838. doi:10.2307/2082710
Arceneaux, K., Johnson, M., & Murphy, C. (2012). Polarized Political Communication, Oppositional Media Hostility, and Selective Exposure. The Journal of Politics, 74(1), 174–186. http://doi.org/10.1017/S002238161100123X
Blumler, J. G. (1979). The Role of Theory in Uses and Gratifications Studies. Communication Research, 6(1), 9–36. http://doi.org/10.1177/009365027900600102
Bos, L., Van Der Brug, W., & De Vreese, C. (2010). Media coverage of right-wing populist leaders. Communications, 35(2), 141–163. doi: 10.1515/comm.2010.008
Brants, K., & Van Praag, P. (2006). Signs of Media Logic Half a Century of Political Communication in the Netherlands. Javnost - The Public, 13(1), 25–40. http://doi.org/10.1080/13183222.2006.11008905
Brooks, D. J., & Geer, J. G. (2007). Beyond Negativity: The Effects of Incivility on the Electorate. American Journal of Political Science, 51(1), 1–16.
Calhoun, C. (1992). Introduction: Habermas and the Public Sphere. In C. Calhoun (Ed.), Habermas and the Public Sphere, (pp.1–47). Cambridge, MA: MIT Press. Cappella, J. N., & Jamieson, K. H. (1997). Spiral of Cynicism : The Press and the Public
Good. New York, NY: Oxford University Press.
Coe, K., Kenski, K., & Rains, S. A. (2014). Online and Uncivil? Patterns and Determinants of Incivility in Newspaper Website Comments. Journal of Communication, 64(4), 658– 679. http://doi.org/10.1111/jcom.12104
Conway, J. C., & Rubin, A. M. (1991). Psychological Predictors of Television Viewing Motivation. Communication Research, 18(4), 443–463.
http://doi.org/10.1177/009365091018004001
Daily Mail Reporter. (2012, September 18). UKIP leader loses appeal against £2,400 fine for saying EU chief had 'charisma of a damp rag and appearance of a bank clerk'. The Daily Mail Online. Retrieved from
http://www.dailymail.co.uk/news/article- 2205131/Nigel-Farage-Politician-fined-4-000-saying-EU-president-charisma-damp-rag.html
Dalisay, F. S. (2012). The Spiral of Silence and Conflict Avoidance: Examining Antecedents of Opinion Expression Concerning the U.S. Military Buildup in the Pacific Island of Guam. Communication Quarterly, 60(4), 481–503.
http://doi.org/10.1080/01463373.2012.704567
De Lange, S. L. & Akkerman, T. (2012). Populist parties in Belgium: a case of hegemonic liberal democracy? In C. Mudde & C.R. Kaltwasser (Eds.), Populism in Europe and the Americas: Threat or Corrective for Democracy? (pp. 27-45), New York, NY: Cambridge University Press.
De Persgroep. (2015). Het Laatste Nieuws. Retrieved from http://www.persgroep.be/nl/onze-media/het-laatste-nieuws
