Assessing the climate of public opinion in the user comments era: A new epistemology

Abstract

This article examines the role of user comments in evaluating the climate of public opinion. It aims to evaluate the relevance of quasi-statistical assessment of public opinion – which was tailored to traditional media – to the digital era. The article, based on 21 interviews with Israeli users of news websites, argues that comments-browsing on the Internet gives a new meaning to the notion of a quasi-statistical assessment of public opinion. The aggregation of different comments, each of which contains an implicit cue for the climate of public opinion, transforms them together into a direct cue. The effect of the merging of journalistic contents with user-generated contents side-by-side on the same website is also evaluated through the perspectives of persuasive press inference and exemplification theories.

Keywords

Epistemology Israel online news public opinion user comments

In the past decade, user comments have become an immanent component of news websites. Compared to other forms of participatory journalism, user comments are unique as they enjoy the amplification generally reserved for mass communication contents (Lee and Jang, 2010; Reich, 2011). The importance of user comments on news websites is driven by the role of these sites in the public sphere: although news sites, often linked to traditional print newspapers, face economic challenges, they still carry a journalistic position of authority (Singer et al., 2011).

One factor to take into account following the emergence of user comments is a change in the epistemological assumptions underlying the assessment of public opinion. In the traditional media era, scholars posited that the public evaluated public opinion based their observations both of journalistic outlets and of their immediate social circles. Within this process, the assessment of public opinion through journalistic outlets was based on a ‘secondary’ observation: the first observation is that of the journalist, who ostensibly observes ‘social reality’ and reports about it, whereas the secondary observers are those exposed to the journalistic work. These readers or viewers gather direct journalistic cues, such as public poll results, and implicit cues, such as the individual views of the journalists, all of which give them a sense of the socio-political trends in controversial issues on the public agenda.

In this article, I will examine whether, in the digital era, the secondary observation of the public opinion climate has been replaced, at least partly, by the seeming direct and unmediated observation of public opinion as reflected in user comments. The appearance of user comments following website news articles reflects a new intersection between authoritative news expression and public expression, two components that are often described as complementary sources for assessing the public opinion climate. While integrating letters-to-the editor in traditional newspapers was intended to reflect the same type of intersection – between the authoritative journalistic voice and the voice of the public – space restriction of the print culture limited the number of letters printed. Furthermore, letters were screened and selected by gatekeepers: in the form of editors and journalists. Online comments sections, on the other hand, are not restricted in the number of comments they can publish. Moreover, comments are assumed to be free of the editorial process of gatekeepers (with some exceptions, for example, the removal of inappropriate comments). Thus, browsing through comments can be seen as giving a new meaning to the notion of quasi-statistical assessment of public opinion. The user comments on news articles can be seen as a nonrepresentative quasi-poll through which people can express their views about issues. On sites that offer the option to ‘like’ or ‘unlike’ a comment – an act that takes much less effort than writing a comment and that bypasses possible pre-moderation and censorship of comments – the sense of quasi-poll becomes even stronger. In addition, in this article, I also consider the importance of comments in the assessment of public opinion climate through the lens of exemplification theory, examining the use of concrete, often vivid, exemplars in comments that catch the users’ attention (Zillmann, 2002).

Scholarship holds that participatory journalism has altered the epistemology of traditional journalism. We see this, for example, in the epistemology of weblogs, which, as Matheson (2004: 460) argues, differ from traditional news media by making fewer claims to holding knowledge. Weblogs’ authority emerges from their connectedness to other sites and knowledge sources. In this article, however, I am not interested in the epistemology of news artifacts but in how users frame their knowledge about the public opinion climate through their engagement with news websites. As the epistemology here relates to the users’ experience, my conceptualization of the possible shift in epistemology will be based on interviews with users of news sites. More precisely, I will base my claims on 21 in-depth semi-structured interviews with Israeli news site users.

I include a note of caution here with regard to the interviewees. All the people interviewed in this study reported being engaged, on various levels, in the user comment sphere. All of them had at least some experience in sending user comments in the past. But because the attitudes of this group might differ from those of lurkers, they may not represent this latter, larger population. Focusing on the former group, however, allows me to crystallize and sharpen theoretical phenomena on the issue of public opinion assessment in the user comments era that would be more difficult to grasp through interviews with wider groups of news sites users in general. Of course, many lurkers do not read user comments. But as interviews with users who have experience posting comments revealed, (at least) some users only read user comments and not the journalistic articles themselves. Moreover, this group allows us to assess how users perceive the possible effect their posts have on others. Nonetheless, while evaluating the interviewees’ attitudes about the role of public opinion climate, we need to bear in mind this is not a representative group.

We also need to take into account the specific commenting culture in user comments in Israel. User comments play an important role in Israel’s journalistic and public spheres – the culture in which the interviewees in this study are embedded. User comments were integrated in Israeli news sites relatively early in history, in the first half of 2000 (Nagar, 2011), and are well acknowledged in Israeli media. They are often cited in traditional mass media and seen as influencing editorial decision-making (Nagar, 2011) as well as indicating the public’s interest in specific news issues (Melamed, 2006). Although comment-posting often requires pre-registration, where users must provide personal details, this is not mandatory on Israeli news sites. In order to reduce the number of offensive comments, these sites are often moderated and comments are screened before their publication (Reich, 2011). While most users prefer to comment anonymously, the pre-moderation process in popular news sites might result in a high rate of comment rejection.

In the next section, I will refer to theories that have strong relevance in assessing the public opinion climate in traditional media and assess their relevance in the user comments era.

The quasi-statistical assessment of media contents

The media in general, and news organizations in particular, are seen as social institutions ‘that produce and communicate statements about reality on a regular basis’ (Ekström, 2002: 274). As McLeod et al. (1995: 73) argue, one role played by mass communications is as ‘epistemological devices’: mass communication influences how and what we know and connects the individual and the collective effort to know. Thus, mass communication provides ‘the raw material’ for individuals to understand the opinion climate around them (Hoffman, 2012: 464). In this spirit, news organizations often implement empiricist-oriented journalistic epistemology (Ekström, 2002) to provide the public with ‘information’, which is the defining mark of journalism as a profession (Mindich, 1998: 115).

Systematic observation of media and social environment

The image of people systematically scanning the media in order to assess public opinion is not a new one. In 1958, Davison described individuals engaging in ‘personal sampling’ of the public sphere, among them media outlets, in order to generalize about the opinion of certain groups or society in general on specific issues:

They [people] note views reported in the mass media and may even question taxi drivers or casual acquaintances they meet on the train as to the prevailing opinion on a given issue in one part of the country or another. (Davison, 1958: 99)

In a similar spirit, Noelle-Neumann (1974: 44) in her seminal theory of the spiral of silence assumes that people continually engage in assessing the public opinion climate with regard to controversial matters. She assumes this is done in a quasi-statistical manner, ‘by observing [their] social environment, by assessing the distribution of opinions for and against [their] ideas, but above all by evaluating the strength (commitment), the urgency, and the chances of success of certain proposals and viewpoints’. In trying to understand the tendency of the public regarding controversial issues, Noelle-Neumann imagined people assessing the atmosphere of their social surroundings, and at the same time scanning mainly consensual media outlets, categorizing reports according to their views on these matters. This image is based on the notion of secondary observation of the public sphere, wherein journalists function as public envoys or representatives (Muthukumaraswamy, 2010). The public cannot be everywhere all the time and therefore cannot be exposed in a direct way to all social occurrences; journalists are thus supposed to observe – to reflect the ‘social reality’ for the public.

While Davison and Noelle-Neumann assume that people view the public media as reflecting public opinion, later approaches treat this process as a third-person effect, where people assume the media affect others. Thus, Gunther’s (1998) persuasive press inference theory assumes, similar to Noelle-Neumann, that people are engaged in a systematic quasi-statistical assessment of the press. Yet, according to this theory, people do this because they believe the slant of the media influences other people and in this aggregated way affects public opinion. According to this line of thought, the ‘media may not be very influential in telling us what to think, but they do have the ability to influence our perceptions of what others think’ (Tsfati, 2003: 66).

These assumptions about the engagement of individuals in assessing public opinion are of particular importance in analyzing the role of user comments in this process. As both Davison and Noelle-Neumann suggest, this assessment is constructed by the media as well as by observing others’ opinions in surrounding social circles. In the user comment era, however, both the authoritative journalistic outlets and the opinions of ‘ordinary people’ (the taxi drivers, as in Davison’s example) appear on the same web page framework. This becomes more interesting if we take into account Gunther’s (1998) persuasive press inference theory. If people assume that the media strongly affect others, and therefore public opinion, how would they perceive the conjunction between media content and the comments of ‘others’? Thus, if a news article on a certain issue is followed by contradicting user comments, we can well wonder how this will affect the readers’ quasi-statistical assessment of the public opinion climate.

The new hybrid framework, where journalistic content is presented alongside public comments, raises questions about the motivation for observing the climate of opinion, in our case in a news website. One possible motive, stemming from Davison and Noelle-Neumann’s theories, is the fear of isolation. Other motivations that have been suggested for comments-reading include the desire to validate self-opinion against the opinions of others – assessing the real ‘pulse of community’ (Diakopoulos and Naaman, 2011). The readers’ use and gratification integrate with writers’ motivations: the latter might use the comment section to test the public reaction to their own views and express their experiences or emotions (Diakopoulos and Naaman, 2011).

User comments as a quasi-survey and possible base-rate fallacy

The quasi-statistical assessment of the social environment, in general, and media content, in particular, refers to the aggregation of evidence by individuals as part of the cognitive process of assessing public opinion. We see similar logic behind the scholarly differentiation between direct indications of the climate of public opinion (i.e. opinion polls) and indirect indications (i.e. news coverage), described as surrogates for more direct references (Mutz and Soss, 1997: 432). Studies have found that direct representations of public opinion have more influence on individual’s perception of the opinion climate than implicit cues, such as individual opinions presented by journalists. As Zerback et al. (2015) argue, ‘People simply judge survey data to be a more adequate indicator of public opinion than article slant’ (p. 433). Yet, while indirect cues – individual opinions – were found to have a weaker direct effect on the assessment of opinion climate, Zerback et al. (2015) found they strongly affected personal opinions. Furthermore, the role of such indirect cues in assessing public opinion increased in situations that lacked certainty or information, such as surveys. In another study, Hayes et al. (2011) tested the linkage between fear of isolation and the attention given to public opinion polls in six countries. They found that fear of isolation ‘does serve to stimulate the quasi-statistical organ to tune into the signal of public opinion transmitted through the mass media in the form of public opinion poll results’ (Hayes et al., 2011: 456).

The idea of direct representation of public opinion through numbers and polls is relevant to this discussion about the comment sphere and the assessment of public opinion, as the next sections reveal that user comments are often seen as direct, authentic indicators of the public sphere. In this way, user comments (which can, in countries such as Israel, include hundreds of comments per article) seem to lay out the public before the users in an unmediated, quasi-statistical manner. This might undermine users’ dependence on journalistic work in assessing the climate of public opinion. The number of the comments supporting or opposing certain issues can become a direct cue – an instant (unrepresentative) quasi-survey – indicating the public’s attitudes and trends. Many websites have integrated the technical capability of expressing sentiments supporting or objecting to the news article or user comment. The number of likes or unlikes a comment receives might strengthen readers’ perceptions that the comment arena provides direct quantitative cues of the popularity of a certain view.

It should be noted that while empirical studies do not support the finding that the number of likes/unlikes affects readers’ evaluation of public opinion climate (as some interviews in this study indicated), some findings prove that the content of comments influences readers’ attitudes (Lee and Jang, 2010; Winter et al., 2015). This raises the possibility of base-rate fallacy, that is, when a specific individual case following the presentation of abstract information (such as statistics or polls) has a greater influence on recipients’ perception of public opinion. This influence was found to be more profound than abstract information alone. It is in this regard that exemplification theory emerges (Zillmann, 1999, 2002). Exemplification theory assumes that concrete examples place fewer cognitive demands on people and therefore have a stronger influence than more abstract discussion (Zillmann, 2002: 25). As Zillmann (1999) argues,

Taken together, the effects of such case presentations have been interpreted as showing that recipients give disproportional attention to concrete, often vividly displayed events, especially to those that engage the recipients’ emotions, and that this attention preference comes at the expense of attention to more abstract, comparatively pallidly presented information. (p. 70)

As Winter et al. (2015) argue, ‘single statements by peers can be regarded as vivid exemplars (which also contain more potentially persuasive content to think about)’ (p. 435). Experiment results support the effect of others’ user comments on how people assess public opinion and their social environment (Lee and Jang, 2010: 840; Peter et al., 2014). This was linked to exemplification theory, seeing user comments as providing cognitive aid in judging public opinion (Zerback and Fawzi, 2017: 12).

Following the discussion above, several theoretical issues arise related to the public opinion assessment in the user comments era:

The notions of ‘personal sampling’ or quasi-statistical assessment of the public climate engage with assessment of both the media outlets and the social environment. In the user comments era, these two components merge. How does this affect the assessment of the public climate?

According to the persuasive press inference theory, people assume that the media have a persuasive effect on others. How does the fact that ‘others’ views’ are presented through user comments, in conjunction with the journalistic article, affect this belief regarding the media effect.

Following the logic of quasi-statistical observation, in the public sphere direct cues of public opinion, such as polls, influence individuals’ perceptions of public opinion climate. Is it possible that a high number of comments, along with the capability to like or unlike them, can be seen by users as direct cues of public opinion?

Can we find evidence in users’ experiences of the effects that exemplars presented in the comments have on the assessment of public opinion climate (exemplification effect)?

I will refer to these issues in my discussion of the interviews, which will follow the methodological section.

Methodology

The study is based on 21 in-depth semi-structured interviews, conducted between April and November 2015. Interviewees were Israeli news site users who identified themselves as engaged, to varying degrees, in reading and sending comments on news outlets. They were recruited through the posting of user comments on main Israeli news sites (such as Ynet and Haaretz) with a request to interview them for the study, as a response to comments they had posted. Users who responded were then asked to take part in a face-to face, Skype, or telephone interview (conducted in Hebrew) about their experiences in the user comments sphere. Interviewees also included communications students who, in responding to a Facebook advertisement calling for volunteers, identified themselves as engaged in commenting online on news sites. While it was relatively easy to recruit male interviewees, women proved a greater challenge, perhaps indicating a deeper phenomenon of female underrepresentation in the user comments sections (Pierson, 2015). Women mainly ignored the messages posted following their comments. Some agreed to take part in the study but later changed their minds. As a result, and following a wide survey conducted on the subject as part of a larger study, seven female interviewees, who identified themselves in a digital survey as active in the user comments sphere, were recruited (two were recruited following requests to interview users, posted in user comments sections, as in the regular recruitment procedure noted above). These seven women interviewees were offered gift certificates as an incentive for taking part in interviews. Following these efforts, interviewees included 12 males and 9 females.

Of the interviewees aged 25–67 years (M = 46.9 years), 13 of them reported having an academic degree, 4 reported partial academic education, and 4 reported completing high school (12 years of education). Six of the interviews were conducted face-to-face and eight through Skype. Seven interviews, all of them with women who refused to do face-to-face or Skype interviews, were conducted on the telephone. Of the 21 interviews, 20 were recorded and transcribed (1 interviewee refused my request to record the interview). Questions asked in the interviews dealt with various issues, following relevant scholarly literature regarding user comments and other theories of social behavior. Interviewees were asked about their comments posting and reading habits, their motivations, their attitudes toward posting a comment, and those of their significant others. They were asked about varied matters relating to public opinion and user comments: for example, the triangular relationship between news articles, comments, and user opinions; consensus versus conflict situations; the assessment of public opinion (spiral of silence effect); and the effect of the number of comments in general and quasi-statistical cues in particular. The questions referred to the interviewees and to other people in order to assess possible impact of the third-person effect. Questions were asked generally about the user comment spheres and also in a more concrete way about particular issues such as state–religion relations in Israel and the 2015 election – both issues that fit well into the general framework of highly commented-on items, as pointed out in the scholarly literature (Boczkowski and Mitchelstein, 2013; Tenenboim and Cohen, 2015).

Assessing the public opinion climate: Journalistic article versus user comments

As we have seen in the theoretical discussion, the assessment of public opinion is assumed to involve observation both in media outlets and in the social milieu. In the user comments era, the authoritative journalistic voice and the expression of ‘ordinary people’ merge on the same web page. This raises questions about the cognitive process of personal sampling and quasi-statistical assessment of public opinion.

While there was no unanimous agreement on the issue, several interviewees emphasized that when it comes to assessing public opinion, user comments are a much more useful tool than journalistic articles. The main reason, repeated over and over in the interviews, was that the journalist writing an article is only one person, whereas the comments are written by many people. Users identify each comment, as long it bears a different name or even a nickname, as the voice of a single person (Interviewee 8, 21 May 2015). Thus, comments are seen as ‘written by the masses’ and, as such, as reliable testimony of the public opinion climate (Interviewee 10, 5 June 2015). As one interviewee put it, ‘The writer [journalist] is one. And then 600 commenters follow. This can make a difference’ (Interviewee 18, 29 October 2015).

The preference here of the output of the many – even if their reliability is unknown, particularly given the veil of anonymity – echoes the idea of the wisdom of the crowds (Surowiecki, 2004). According to this concept, a large group of people who are willing to collaborate in the creation of a product will increase its quality. In our case, the comments of an expert journalist, even when he or she works according to ethical and organizational constraints, are seen as less reliable than the accumulated comments of many unknown users. The wisdom of the crowds is often cited in discussions of Wikipedia’s virtue as a source of information (Fallis, 2008; Kittur et al., 2007). Despite epistemological doubts about Wikipedia, often related to the reliability or expertise of the contributors (Fallis, 2008), it is still considered a reliable source due to the wisdom of the group. In the case of assessing the opinion climate, then, the role of the group is significant, as a variety of people in the crowds and the aggregation of ideas are seen to constitute and reflect the climate. In this way, it is not only a matter of assessing something that is detached from the commenters; the comments themselves constitute the thing that they assess:

You can see which way the wind blows there [in the comments section]. Who supports and who opposes. Who supports this way and who the other. Each person comes with his own baggage […] and this is the place where he lifts up the crane of his truck and unloads his stuff. (Interviewee 9, 28 May 2015)

The merging of the article and the comments seems to reflect, at least according to some of the interviewees, a weakening of the persuasive press inference when it comes to the authority of the journalistic outlets. One reason for this is the assumption that a great number of people do not read the articles. Comments are much shorter than articles and sometimes more amusing to read (Interviewee 8, 10 April 2015). Another, more popular reason cited by interviewees is that journalists are biased and lack objectivity, a problem that the comment sections can balance and correct. As one person argued, ‘At the end of the day, journalists simply bring their own agenda’ (Interviewee 3, 6 May 2015). In some interviews, journalists were described as having leftist political orientation. Interviewees argued that journalists did not reflect ‘reality’ and went against the people’s will and interest. This hostile perception of journalists reflects the general tendency to blame the media for society’s problems (McQuail, 2005: 53; Tsfati and Huino, 2014). Comments are described as attacking journalists’ views (Interviewee 15, 23 October 2015) and as a balancing mechanism to correct, at least partly, leftist-biased journalism:

[T]his wave [of comments] seems much more influential than the view of one person [the journalist], which is so slandered […] I always have my doubts if this is a real wave or not. But my impression is that it is more influential even than the article. (Interviewee 5, 5 December 2015)

Compared to journalistic outlets, user comments are seen as much more authentic and independent testimonies that are anchored in reality (Interviewee 13, 26 June 2015).

In interpreting interviewees’ attitudes, I found that some of them assume comments have a persuasive biased effect on others. Yet, the interviewees also emphasized their own resistance to such effects. Thus, one person interviewed, who complained about censorship in the comment section, emphasized that other users cannot really know the trends of public opinion from observing the comments. Yet, ‘they think they know. Unless, as I told you, they can do the manipulation, see how many likes and unlikes there are, to try to post a comment and see if they are [pre-]screened or not’ (Interviewee 6, 5 December 2015). Similarly, another interviewee argues that the comment sphere reflects the ideological agenda of the website, and also emphasizes that most of the commenters are not aware of this fact: ‘[Y]ou need to be more sophisticated, to read between the lines […] I try to be very critical. I am too critical’ (Interviewee 12, 14 June 2015).

Number of comments as a direct cue for public opinion climate

Based on the aggregation of a high number of comments common on Israeli news sites, some interviewees noted their use in quasi-statistical, or even real statistical, assessment of climate of public opinion. As one of the people interviewed noted,

You can use it [comments section] as a sort of … I will not say a study, but a sort of statistical model, some kind of research, or a market survey […] regarding peoples’ opinion. (Interviewee 7, 19 May 2015)

He explained that while it was customary in the past to conduct telephone surveys to study the public’s opinion on certain matters, today you can write a news story about any policy or political initiative and assess the public’s attitude through the comments on the article. He suggested that politicians plant such articles in news sites to gauge the public climate and plan political strategy. Other people interviewed noted the potential of using comments to assess public opinion on specific issues (Interviewee 15, 23 October 2015).

We can draw parallels here between the socio-political function of peoples’ participation in opinion polls and the social functions of participation in user comments sections. Surveys whose results are presented in traditional media are seen both as a channel for people to express their views and as a tool for the public to calculate political choices, especially at election time (Portilla, 2016: 36). In a similar way, user comments give people the opportunity to speak up. In fact, professional journalists see the comments as opinions, distinct from the journalistic content (Heinonen, 2011: 41). Yet, the process of commenting is just as important as the process of publishing opinion polls results, not only because it allows people to express an opinion but also because the accumulation of comments indicates the direction of public opinion. While each individual comment, similar to a journalistic article, may hold implicit cues of the public opinion climate, the aggregation of a great number of comments transforms them, at least in the eyes of some of the interviewees, into a direct cue for public opinion.

These statistical indications seem even stronger if accompanied by the capability to ‘like’ or ‘unlike’ comments. As some interviewees noted, this feature is an important statistical indicator of the opinion climate. Despite the lack of empirical evidence that the number of likes influences readers’ evaluation of a news story (Lee and Jang, 2010; Winter et al., 2015), some of the interviewees saw them as direct cues, bypassing institutional and editorial manipulations. While news sites can censor comments through pre-mediation in order to ensure they fit the site’s political orientation, the like/unlike numbers cannot be controlled:

The like and unlike give a picture. You can immediately see, you can see the numbers. Sometimes it goes to the extreme. You see 2 unlikes and 47 likes. And then you scroll all the way and you see, you can even make an adjusted calculation, to take these numbers, all of them, and to sum up. (Interviewee 10, 5 June 2015)

Interviewees described the comments as a tool to ‘understand the vibe’ of the public. As one of them put it, a user comment is ‘an experimental balloon to see where my views are [socially positioned] and to what extent they differ from those of others’ (Interviewee 3, 6 May 2015). One of them uses this assessment of the vibe as a tool that helps them decide whether to join other commenters. Thus, users look at the comments section to see whether any comments resemble their own views (Interviewee 20, 30 October 2015). They also use the comments to evaluate the popularity of their views:

The world completely changes […] you are exposed, your views, you know exactly where you are and where you are positioned. You are not only in your private sphere, among your friends and work colleagues, anymore. You are opened to the whole world, not to the whole of it, let’s say you are opened to the half of the country that can comment in this way or the other. (Interviewee 9, 28 May 2015)

In the eyes of some users, then, the comment sphere functions as a quasi-survey and therefore provides direct cues regarding public opinion. Yet, following the same rationale – of the quantitative characteristics of the comment sphere – some of the people interviewed noted its bias and unreliability as a measurement.

The bias of user comments

While acknowledging the theoretical potential of user comments as a tool to assess the public opinion climate, some of the interviewees emphasized the limitations of the ability to de facto assess public opinion through comments. Their reasoning echoes concerns raised about surveying populations: that is, the problems of sampling and deliberate bias of the comments prevent a reliable assessment of social trends. Thus, while users believe that some obstacles prevent the assessment of the opinion climate through comments, their reasons are anchored in quasi-statistical considerations. One of the interviewees even compared the failure of surveys to predict the 2015 election in Israel to the bias in user comments. According to his view, certain social groups were deliberately excluded both from the population of survey respondents and from the user comments sphere (Interviewee 10, 5 June 2015).

One of the interviewees argued that the medium for comments itself – that is, computer-generated typewritten text – creates an inherent bias in commenters: as he explained, not all Israelis have access to computers; moreover, some of the users have difficulties expressing themselves in writing. This distorts any assessment of the climate of public opinion based on user comments, because these excluded people, who might hold homogeneous views on certain issues, are not represented (Interviewee 9, 29 May 2015). The fact that user comments are a technology allowing personal expression is also seen as a source of bias as they are perceived as promoting negative discourse that challenges the news article’s opinion regardless of its view (Another interviewee argued that user commenting, whether supporting or opposing a certain viewpoint, promotes extremist discourse.) (Interviewee 8, 21 May 2015; Interviewee 3, 6 May 2015). The perception that user comments in Israel have a conflictual orientation is not unfounded. As Melamed (2006) shows, regardless of the subject matter of the article the readers are commenting on – politics, sports, entertainment, and so on – user comments tend toward conflict (p. 138).

Other reasons cited for the nonrepresentative nature of the user comments relate to intended manipulation of the comment sphere, for example, when people are paid to post comments to promote a specific view (Interviewee 18, 29 October 2015). Some of the people interviewed also claimed that news sites themselves, by actively censoring comments through moderation, make the political picture reflected in the comments sections unreliable. These claims were based mainly on interviewees’ sense that their comments were systematically ignored or censored for political reasons even though they did not contain inappropriate language or attitudes. According to Reich (2011: 106), the reject rate for comments in Israel is high compared to other counties; however, journalists he interviewed noted that at least 40 percent of the user comments are published.

The exemplification function of comments

The doubts expressed in empirical studies that likes/unlikes are used as a quasi-statistical direct cue for trends in the public opinion, on one hand, and the proof of the effect of the comments’ contents, on the other hand, raise the possibility of the exemplification role of comments (Lee and Jang, 2010; Winter et al., 2015). Generally, interviewees refer to comments as reflecting views. This might be linked to the fact that highly commented upon items in Israel are those dealing with political-controversial issues (Tenenboim and Cohen, 2015). Yet, in line with the results of previous studies (Diakopoulos and Naaman, 2011), the people interviewed repeatedly mentioned that one of the pleasure of being involved in the comment section was sharing experiences and learning from others. This strengthens the perception that comments using examples help users better understand the matters discussed in the article. Understanding lengthy articles was mentioned as a demanding cognitive task. Articles were described by some users as long and boring (Interviewee 9, 28 May 2015), whereas the shorter comments helped them to understand ‘what is really going on in the article’ (Interviewee 15, 23 October 2015). Thus, in line with exemplification theory, comments – with their brevity, variety, and simplicity – are seen as cognitive aids to help deal with the issues discussed in the article.

Having experience with the issue under discussion is also considered as a motivating factor for writing a comment, encouraging the receipt of supportive comments from other users. Thus, one interviewee, a single mother (Interviewee 17, 30 October 2015), mentioned that she read an article, felt strong empathy, and shared her own story as an example. In the same vein, a former teacher noted that she usually posts comments when she feels she can contribute from her experience:

Last year when there was the protest to reduce the number of pupils in classrooms, as a former teacher I explained my standpoint, how difficult it is to mange 42 kids in class. Following that there were comments of other teachers, saying, ‘Sara this and that comment is correct; I am also a teacher’. (Interviewee 16, 25 October 2015)

She noted that when she does comment, relaying on her experience, she often receives supportive comments from other users telling her she helped them to think about the issue in a new way, giving her the feeling that her comments are being read and influence others.

The authenticity of examples is associated with whether the comments are identified or not. Thus, one interviewee mentioned posting comments about her abusive relationship and her attempt to rehabilitate herself after her divorce, noting that when she shared this experience, in order to strengthen other abused women, she did it through identified comments (Interviewee 14, 29 July 2015). The use of her name and picture was part of the general message to the women she was commenting that they were not alone and not the first to suffer. However, she also noted that she prefers to comment in such an identified way in closed Facebook groups and not on news sites.

Other people interviewed also mentioned using examples as a marketing or commercial tool. One person, who works as a teacher, noted that she works with her students and her own children on comment literacy:

We read the article together and also look at the comments. We try to trace … a bit of literacy of these things. And then we see, after reading an article about a [weight loss] drug that has no effect, a comment stating: ‘Yes, I lost 84 kg because I used it’. So you understand […] it is important to see what is true and right. (Interviewee 14, 29 July 2015)

Another person referred to the commercial power of examples, as well as to the influence of the anonymous versus identified nature of such comments. Referring to comments on a car test report that was published on a news site, he described comments of personal experiences with the specific cars tested. He noted that it was hard to know whether these comments were about real experiences using the car or were written only for commercial reasons. If a specific commenter is identified, a reader can look for the name on Facebook to check whether the poster works at the car agency, for example (Interviewee 2, 29 April 2015).

In concluding this section, it can be said that some of the people interviewed showed they were aware of the effect using examples in comments can have on how other people perceive the public sphere. Issues related to sharing and evaluating the role of personal examples seem to be more dominant among women than men, although this may be coincidental and requires further study.

Discussion

Interviewees in this study often reported they scanned user comments in attempting to assess the public opinion climate. As some of them argued, the comments section provides them with an opportunity to learn about public opinion in a direct and authentic way. Some of them revealed awareness about the possible manipulation in this arena and the nonrepresentative nature of the comments. Yet, in doing so, they often spoke using statistical terms, as if the comment sphere were a survey that was manipulated or nonrepresentative. (Indeed, one format user-generated contents has integrated into news websites is ‘polls’, where users are invited to answer multiple questions posted by the website (Hermida and Thurman, 2008: 354).) This reflects their expectations of this sphere, which they assume the comment sections do not meet. Other interviewees described the comments as representing authentic voices: those of ‘ordinary people’. The authority of these comments, or their social strength, lies in their large numbers, which provide a wide spectrum of public opinion. As such, some of the interviewees, in line with previous research, saw user comments as a barometer of public opinion (Lee and Jang, 2010: 843).

Interviewees expressed the belief that the user comments will balance out the biased perspectives of the journalists. Some of them believed, in what can be seen as a variation on Gunther’s (1998) theory of the persuasive comments inference, that the slant of comments has a persuasive effect on other users. Given this effect, there is a need to systematically scan and assess the comments.

The findings in this study provide evidence that we can apply the idea of quasi-statistical assessment of the public opinion to participatory journalism – that is, user comments. The fact that these user-generated contents merge with journalistic articles might erode the role of the latter in modern media environments. While some interviewees expressed the view that the news stories have an advantage in their effects compared to user comments due to their typographic priority (interviewee 19, 29 October 2015), the variety of comments and their high numbers were seen to constitute a direct cue for public opinion. User comments also help readers comprehend the issues under discussion, as the articles are often seen as long and boring. The brevity of the comments, along with quasi-statistical mechanisms such as like/unlike buttons embedded in some of the websites, eased the quasi-statistical assessment of the public sphere.

Another approach to the comments sphere is to see it as a cognitive means to ease user assessment of public opinion climate, viewing the comments as concrete, often vivid and amusing, exemplars. In line with the exemplification theory, some of the interviewees noted the persuasive role of example in comments, reflecting their belief in the affective nature of examples in shaping others’ views on the issues being discussed and, following that, in shaping how others assess the social environment. When users posting the examples identified themselves, this was seen as strengthening the comment’s authenticity and therefore its effect.

Examining the findings of this study through a broader theoretical lens reveals a possible change in how people ‘know’ opinion climate through their engagement with news outlets. The integration of user comments on news sites seems to reflect, at least partly, a shift in the epistemology of assessing the public opinion climate: from a secondary observation, in which people scan journalistic reports and observations of social reality, to direct observation of the mass’ comments, which reflect the opinion climate on news topics. The idea that the systematic assessment of the public opinion climate involves observing not only media outlets but also people in their social circles is not new. But the merging of the possibility of assessing journalistic outlets and the opinions of peer readers at the same time and within the same framework is new and might lead to the erosion of the importance traditionally given to the journalistic articles in this assessment process.

A few limitations must be taken into account in assessing the results of this study. First, I focus on the attitudes of users who are engaged in the user comment sphere. It may be that the trends detected among this group are bolder than those we would find in society at large (in fact, this group was chosen because of this assumption). It is quite clear that interviews with users who tend to ignore user comments would have led to different conclusions. Second, user comments are part of a specific national journalistic culture.

Future studies can be undertaken in several directions. First, comparative cross-national research is needed to assess whether this study’s conclusions are unique to Israel’s cultural sphere or represent general global trends. Second, there is a need for a comprehensive comparative study among people who identify themselves as engaged in the comments sphere and the larger society. Among other things, such a study should examine whether the tendency to ignore news articles and read the comments only, which was evident among some of the interviewees in this study, is evident across a broader population. There is also need for such comparison in considering possible differences between groups in their attitude toward user comments as a tool aiding assessment of public climate. Third, while several studies have examined the role of comments as exemplars that affect public opinion perceptions, a more nuanced approach to comments’ content is needed. For example, scholars could turn their attention to the different effects that various genres of comments – for example, exemplars versus general political comments – have on public opinion perceptions.

Footnotes

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Israel Science Foundation [Grant Number 898/14].

Author biography

Oren Soffer is Associate Professor at the Department of Sociology, Political science and Communication at The Open University of Israel. His research interests include the study of new media, while using historical analysis to better understand and theorize new textual and social phenomena. He also studies the socio-political history of Hebrew mass communication.

References

Boczkowski

Mitchelstein

(2013) The News Gap: When the Information Preferences of the Media and the Public Diverge. Cambridge, MA: MIT Press.

Davison

(1958) The public opinion process. Public Opinion Quarterly 22(2): 91–106.

Diakopoulos

Naaman

(2011) Towards quality discourse in online news comments. In: Proceedings of the ACM 2011 conference on computer supported cooperative work, Hangzhou, China, 19–23 March.

Ekström

(2002) Epistemologies of TV journalism: A theoretical framework. Journalism 3(3): 259–282.

Fallis

(2008) Toward an epistemology of Wikipedia. Journal of the American Society for Information Science and Technology 59(10): 1662–1674.

Gunther

(1998) The persuasive press inference effects of mass media on perceived public opinion. Communication Research 25(5): 486–504.

Hayes

Matthes

Eveland

(2011) Stimulating the quasistatistical organ: Fear of social isolation motivates the quest for knowledge of the opinion climate. Communication Research 40(4): 439–462.

Hermida

Thurman

(2008) A clash of cultures: The integration of user-generated content within professional journalistic frameworks at British newspaper websites. Journalism Practice 2(3): 343–356.

Hoffman

(2012) When the world outside gets inside your head: The effect of media context on perceptions of public opinion. Communication Research 4(4): 463–485.

10.

Kittur

Chi

Pendleton

et al . (2007) Power of the few vs wisdom of the crowd: Wikipedia and the rise of the bourgeoisie. In: Proceedings of the 25th annual ACM conference on human factors in computing systems (CHI 2007), San Jose, CA, 28 April–3 May.

11.

Lee

Jang

(2010) What do others’ reactions to news on Internet portal sites tell us? Effects of presentation format and readers’ need for cognition on reality perception. Communication Research 37(6): 825–846.

12.

McLeod

Pan

Rucinski

(1995) Levels of analysis in public opinion research. In: Glasser

Salmon

(eds) Public Opinion and the Communication of Consent. New York: Guilford Press, pp. 55–85.

13.

McQuail

(2005) McQuail’s Mass Communication Theory, 5th edn. London: SAGE.

14.

Matheson

(2004) Weblogs and the epistemology of the news: Some trends in online journalism. New Media & Society 6(4): 443–468.

15.

Melamed

(2006) The talkback in Israel: A reflecting and magnifying mirror or a balancing and rectifying pendulum of the journalistic discourse. MA Thesis, The Hebrew University of Jerusalem, Jerusalem.

16.

Mindich

DTZ

(1998) Just the Facts: How ‘Objectivity’ Came to Define American Journalism. New York and London: New York University Press.

17.

Muthukumaraswamy

(2010) When the media meet crowds of wisdom. Journalism Practice 4(1): 48–65.

18.

Mutz

Soss

(1997) Reading public opinion: The influence of news coverage on perceptions of public sentiment. Public Opinion Quarterly 61: 431–451.

19.

Nagar

(2011) The loud public: The case of user comments in online news media. PhD Thesis, The State University of New York, Albany, NY.

20.

Noelle-Neumann

(1974) The spiral of silence: Theory of public opinion. Journal of Communication 24: 43–51.

21.

Peter

Rossmann

Keyling

(2014) Exemplification 2.0. Journal of Media Psychology 26(1): 19–28.

22.

Pierson

(2015) Outnumbered but well-spoken: Female commenters in the New York Times. In: SIGCHI conference proceedings. Available at: http://cs.stanford.edu/people/emmap1/cscw_paper.pdf

23.

Portilla

(2016) The inclusion of methodological information in poll-based news: How do Spanish newspapers conform to professional recommendations and legal requirements? Journalism 17(1): 35–48.

24.

Reich

(2011) User comments: The transformation of participatory space. In Singer

Hermida

Domingo

et al . (eds) Participatory Journalism: Guarding Open Gates at Online Newspapers. Oxford: Wiley-Blackwell, pp. 96–118.

25.

Singer

Hermida

Domingo

et al . (2011) Introduction: Sharing the road. In Singer

Hermida

Domingo

et al . (eds) Participatory Journalism: Guarding Open Gates at Online Newspapers. Oxford: Wiley-Blackwell, pp. 1–9.

26.

Surowiecki

(2004) The Wisdom Crowds. New York: Doubleday.

27.

Tenenboim

Cohen

(2015) What prompts users to click and comment: A longitudinal study of online news. Journalism 16(2): 198–217.

28.

Tsfati

(2003) Media skepticism and climate of opinion perception. International Journal of Public Opinion Research 15(1): 65–82.

29.

Tsfati

Huino

(2014) Research findings weaken perceptions of media bias. Newspaper Research Journal 35(1): 82–95.

30.

Winter

Brückner

Krämer

(2015) They came, they liked, they commented: Social influence on Facebook news channels. Cyberpsychology, Behavior, and Social Networking 18(8): 431–436.

31.

Zerback

Fawzi

(2017) Can online exemplars trigger a spiral of silence? Examining the effects of exemplar opinions on perceptions of public opinion and speaking out. New Media & Society 19(7): 1034–1051.

32.

Zerback

Koch

Krämer

(2015) Thinking of others effects of implicit and explicit media cues on climate of opinion perceptions. Journalism & Mass Communication Quarterly 92(2): 421–443.

33.

Zillmann

(1999) Exemplification theory: Judging the whole by some of its parts. Media Psychology 1(1): 69–94.

34.

Zillmann

(2002) Exemplification theory of media influence. Media Effects: Advances in Theory and Research 2: 19–41.