EXAMINING ONLINE PUBLIC DISCOURSE IN CONTEXT: A MIXED METHOD APPROACH TAMARA WITSCHGE Abstract The Internet is often praised for its ability to provide a space to enable every person to present her or his view, thus (potentially) allowing for more inclusion and participation in the public discussion. This potential has led many scholars to examine online discussions and see what these can contribute to democracy and the public sphere. These investigations, however, often focus on a single aspect of online discourse: the actual content. It is important to realise that discourse is not constructed in a vacuum: in addition to the text, there is the environment in which the text is produced and consumed as well as the wider social practice to which it belongs. Every instance of language use, including that of online political communication, is a communicative event that consists of three dimensions: the text; the "discursive practice which involves the production and consumption of texts"; and the social practice (Jorgensen & Philips 2002, 68). Though important, the discursive and social practices are often neglected in studies of online political communication. The potential of the Internet for opening up public discourse cannot be fully evaluated if the context in which it is produced is ignored and if issues of power involved in this context are not addressed. In this article I introduce an integrated approach that looks at all three of the aforementioned dimensions of online public discourse from a critical discourse analytical perspective. The proposed mixed method approach allows for an examination and evaluation of the discourse in context, thus broadening the scope of explorations. Tamara Witschge is a research associate in the Department of Media and Communications, Goldsmiths, University of London; e-mail: t.witschge@gold.ac.uk. rN ■ LO rv. si r\T oo o o rN m £ £ 3 a 0 £ * 1 * (R 0 C K LO Introduction The development of the Internet1 as a communication medium for the multitudes has rekindled interest in democratic debate. The Internet's features seem ideal for it to enable the type of communication that should take place in the public sphere.2 The features of the Internet have led a number of scholars to examine the extent to which the Internet enables democratic discussion. The characteristics that are at the heart of these studies include the Internet's unbounded space for interaction and the anonymity of this interaction. They relate to the perceived openness of online space, both in terms of the number and types of people participating (the quantity of interactions) and in terms of the openness people experience (the quality of interactions). The Internet is celebrated for offering the possibility of many-to-many communication (Coleman & Gotze 2001, 17), for bridging time and place (Eriksen & Weigard 2003; Street 1997, 195), and for the transmission of large amounts of information (O'Hara 2002). It is generally seen as "contributing to new ways of knowing, new strategies for gathering, storing, retrieving, and utilising information" (Dahlgren 2004, xv). "Because of its horizontal, open, and user-friendly nature, the Internet allows for easy access to, and thus greater participation, in the public sphere" (Brants 2005, 144). Together with the low (social and economic) costs of publishing, this has created great optimism regarding the Internet's potential. There has been a lot of empirical research into these matters3 with varied and inconclusive result (see e.g., Witschge 2004). Some studies show the Internet to open up spaces for discussion, with participants seeking different viewpoints (e.g. Stromer-Galley 2002, 2003), introducing new participants to the discussion (e.g. Schneider 1997), and participants being generally more supportive of diverse and tolerant points of view than non-users (Robinson, Neustadtl, & Kestnbaum 2002). Counter to these positive findings, one finds that abusive postings, monopolisation of attention, control of agenda, and style of communication make it that some participants are heard more often than others (Dahlberg 2001, 15). A number of issues have to be addressed in relation to these findings. First, whether or not the Internet contains spaces that form or resemble something like £ an ideal public sphere depends on the way people use it. The Internet "itself does not bring about democratisation or openness, but its diffusion does create new openings to struggle for democracy" (Warschauer 2003, 183). In addition, different technologies and different user contexts may produce different experiences of Internet communication: the Internet does not exist (Thomas & Wyatt 1999, 694). Second, studies use various and differing notions of democratic discussion (if it is defined at all) and employ different methods to examine the Internet's potential for democratic discussion. These different definitions and methods to some ^ extent explain the differing findings. There seems to me, however, to be a more fundamental issue at stake here: A considerable part of the studies analyse only the content of the online discourse to determine its potential for democracy.4 Studies that only analyse the content of the online discourse ignore essential information regarding (1) how this discourse comes into being (which determines to a large extent the boundaries of the potential of online discussions), and (2) how this discourse relates to the larger societal discourse. These two dimensions, together with the text, form the online discourse. Examining them together will provide insight 0 £ into the role of online discussion platforms in the public sphere beyond which is possible in a content-only study. In this article, I propose an integrated methodological approach that looks at all ^ three of the aforementioned dimensions of discourse: critical discourse analysis.. In the following section, I will introduce the definition of public discussion that is used in the study. I shall then describe the process of critical discourse analysis. The fourth section introduces the study that will serve to illustrate the proposed approach: the debate on immigration in the Netherlands on Dutch web forums. Once this has been done I will use this study to show the added value of a mixed method approach for examining online discourse as a whole. Differences in Public Debate Public discussion is understood here as "public communication about topics and actors related to either some particular policy domain or to the broader interest and values that are engaged" (Ferree, Gamson, Gerhards & Rucht 2002, 9). It constitutes an important part of democracy and citizenship, especially in polarised societies. Public discussion takes place within the public sphere - a virtual space constituted by all public communication on political issues. Within the public sphere different discourses exist. Difference in discourse reflects a difference in perspectives, experiences and ways of speaking. In this article, public discussion is viewed as the political method to democratically respond to these differences between discourses, as is advocated in theories of public sphere or deliberative democracy. A number of concerns have been expressed regarding traditional, rational, accounts of deliberation and public sphere (most notably Habermas' 1989 account of the 18th century bourgeois public sphere), specifically where it concerns the inclusion of difference in the public sphere. In this paper I will put forward an alternative account of deliberation that takes into account these criticisms and that is focused on the interaction between different discourses and is open to difference. When one reviews the literature on alternative accounts of deliberation it becomes apparent that the main concern is with openness: openness of the debate to different participants, types of discourses and positions. The concept of openness is similar to the criterion of inclusion and equality in rational deliberative democracy theories, but is different in the sense that it does not merely seek the inclusion of all those affected, but also of different types of discourses and different forms of communication (besides or beyond what the majority would argue to be rational). It is also different in what is meant by equality. Here, equality means that everyone has the right to raise issues, open up debates, provide information, and question others. This does not involve the bracketing of one's identity or interests as is the case with traditional deliberation, but rather sees the discourse to be informed by these identities and interests. Moreover, the aim of public discussion (which ideally entails an open and equal exchange) is not the elimination or suppression of difference but rather a constant negotiation of it (Chouliaraki & Fairclough 1999, 135). The differences will, however, possibly be reduced. Public discussion allows, through fostering understanding and appreciation of the other's convictions, concerns, and needs, for different groups to transcend the awkwardness, fear, and hostility that might exist, and let people "appreciate the plausibility of seeing the world from a different perspective" (Valadez 2001, 34). Ultimately, if communication takes place in the open and inclusive way set out above, "this dialogue enables people to navigate and interact across cultural and racial boundaries" (Streich 2002, 138). Because of the importance of public discussion in pluralistic societies, Chou-liaraki and Fairclough argue that research is needed that focuses on dialogue "with the objective of arriving at detailed accounts of practices of dialogue in late modern societies which can discern the obstacles to, practices of and potentials for non-repressive dialogue across difference" (Chouliaraki & Fairclough 1999, 136-137). With this aim in mind, I will argue that critical discourse analysis is the best methodological approach that can be used in such an examination into the obstacles and potential for open discussion. Critical Discourse Analysis The perspective on discourse taken here in order to examine the openness of online public discussion is that of critical discourse analysis [CDA].5 CDA considers language to be social practice, views the context of language use as crucial to the analysis of it and takes particular interest in the relation between language and power (Wodak 2001, 1-2). An important element of CDA for this article is that it takes discourse not merely to refer to an isolated text, but includes in it the context in which a text is produced and consumed, as well as the wider societal practice in which the text exists: "Discourse and any specific instance of discursive practice, is seen as simultaneously (i) a language text, spoken or written, (ii) discourse practice (text production and text interpretation), (iii) sociocultural practice" (Fairclough 1995, 97). These three dimensions of discourse are depicted in Figure 1. Figure 1: Three-dimensional Model for Analysing Discourse (J0rgensen & Philips 2002, 68)6 U S i 0 £ * i * IR 0 C « text production TEXT text consumption DISCURSIVE PRACTICE SOCIAL PRACTICE The main claim of this article is that context matters, specifically the context of the production and consumption of texts. Not only genres such as "news," but also the online political discourse is a product of "specific professional practices and techniques" which are "based in particular social relations, and particular relations of power" (Fairclough in: Richardson 2007, 40). However, this is often disregarded. Easy access to the online discussion and increased possibilities to obtain a speaker's role in the online discussion, have led some scholars to neglect the power relations that determine who gets to say what. When analysing the extent to which online discussions provide an open and equal exchange, and when trying to discern the obstacles and potential for reaching C^ such exchange, one needs to look at the context in which a text is produced and ^ consumed as well as at the text itself. In the case of the production of online texts in web forum discussions, the norms and moderation of the web forums need to be taken into account, as well as the intentions of the participants in the discussions. Both the web forum's maintenance and its users determine to a large extent which contributions are allowed and by whom. Online, as offline, there are norms that guide (and restrict) communication. In terms of text consumption, even though the Internet seems to lessen status barriers, the question is: How do people perceive others and their contributions, and how they consume these contributions? Here, as well, the question of power comes into play: What role do, for instance, marginal voices play when they are produced online, but not heard? The answer to these questions can be found in participants' views, their reasons for discussion online and their evaluations of the discussions. In other words, what interpretation(s) do participants give to the context in which the text is produced and what power relations help shape this discourse (Fairclough 2001b, 134-138)? In addition, it is important to take into account the social practice in which texts are produced and consumed, which may be considered as the wider environment of such discourse. "In essence, CDA involves an analysis of how discourse (language in use) relates to and is implicated in the (re)production of social relations - particularly unequal, iniquitous and/or discriminatory power relations" (Richardson, 2007: 42). Thus, it is necessary to examine the implications of a text for the social context and examine the social (power) relations reflected, altered and reinforced in the discourse. How is the discourse positioned in relation to power struggles in society (Fairclough 2001b, 138)? Online texts are part of broader social practices and reflect, negotiate or resist, and feed back into existing social power relations like any other discourse and thus this aspect of the discourse needs to be examined. How does this theoretical framework translate into a method of analysis of online discourse? An important remark here is that CDA is as much theory as method (Fairclough 2001a, 121) and that there are no clear-cut universal "recipes" for analysis. CDA is not to be used as a method (or theory for that matter) in isolation, but rather it advocates using other methods and theories within the framework of CDA. The methodological approach proposed here is thus not meant to be employed as a universal standard. Rather, the idea is to provide an example of a study conducted using the CDA framework to indicate the benefits of analysing discourse in this broader way. Each study into political discourse will have to set up the method in such a way that it does justice to the particularities of the object of study. Before I describe the way in which this particular study was conducted let me introduce the context of the study from which this methodological approach is drawn. Online Public Discourse on a Contested Issue The research on which this article is based (Witschge 2007) focuses on the public debate on immigration in the Netherlands. The study was conducted in times of tension and anxieties that intensified after such incidents as 9/11; a shift in Dutch politics initiated by the late Pim Fortuyn;7 the Madrid bombings in 2003; the murder of filmmaker Theo van Gogh by a Muslim fundamentalist in 2004; and the London bombings of 2005. In this time, public debate about minorities has focused more and more on issues of social cohesion. The driving question has become whether long-time residents and recent immigrants can live together in a peaceful manner (Gisberts & Dagevos 2005, 66). The differences between groups in society are taken to be so substantial and fundamental that public debate seems at an impasse. The focus in the discussion on differences makes it difficult for a constructive exchange of ideas to take place on how to deal with the problems that face contemporary society. Different groups seem reluctant to give each other occasion to speak of their respective experiences and opinions. The question in this situation, where difference and tensions seem to dominate public debate, is how these aspects should be dealt with in such a way as to benefit, and not threaten democracy. In order to try and understand dialogue on a topic of immigration in the Netherlands, and to make sense of how people online voice, negotiate, and reduce difference, the following research question was formulated: To what extent is the public discussion on Dutch web forums on the issue of immigration an open and equal exchange and what are the obstacles and potential for obtaining such dialogue? Web forums are a specific form of Internet communication, even though they might employ different software formats. A web forum can be defined as "an online public discussion area where users exchange ideas and information" (Mann & Stewart 2000, 219). There are a number of shared characteristics for this type of discussion: they are public; participants can remain anonymous; the discussions are organised by themes and topics; the discussions are facilitated or moderated; and participants do not have to be online at the same time.8 On these sites, interaction is not only possible but also generates the main reason for existence of the web space. This is not to say that dialogue between discourses is impossible or less likely to occur on non-interactive spaces. Web forums are chosen, however, because they y are specifically intended for discussion. People can raise issues, publish ideas, and present arguments. Web forums are designed to foster responses to posts, even if authors of posts do not always (intend to) get one. The study focused on the following forums: Fok, Maghrebonline, Maroc, Nieuwre-chts, Politiekdebat, Terdiscussie and Weerwoord. This set of forums consists of various different types: political web forums (both right-wing and left-wing), web forums aimed at immigrants and general web forums. The forums are popular in terms of the number of participants, discussions, and posts. In addition, they produce a ^ large amount of discussion about immigration. Examining Discourse in Context There are three dimensions to online discourse that require examination: the discursive practice online (the context of both production and consumption of a text), the text itself, and its relation to social practice. These three dimensions are strongly interrelated and it is therefore not always feasible to study them in isolation. Four case studies were conducted in order to analyse how open the online discourse is along these three dimensions. The first two studies deal with the online QQ practice, each study focussing on one particular aspect of it. The first study focuses C > on the context of production; it seeks to identify the structural openness of web forums by examining the online rules and regulations. The second study concerns users perceptions of the online discussions and as such deals with both how texts are produced and how they are consumed online. The third and fourth studies look at the openness of the text itself, and analyses its implications for social practice, and to a certain extent also incorporates the discursive practice (as mentioned above it is not always possible or desirable to separate the dimensions in analysis). In the spirit of CDA, three different methods are used in these studies in order to analyse the discussion. These methods are discourse analysis, feature analysis, and a questionnaire. In what follows, I explain how these different methods are employed in the various studies to answer the central question regarding the obstacles and potential of an open and equal exchange. Structural Openness Web forums are often considered to be the "genuine" public sphere: They allow many-to-many interactions, and relative to other media constituting part of the public sphere there are few restrictions in accessing this space. In television, radio, and newspapers, there are technological barriers and journalistic gatekeepers limiting public access, making access impossible for a large section of the public. In contrast to the general idea, however, web forums are controlled and restricted spaces as well, not only because "providers, internet browsers and search engines pre-structure access to information" (Koopmans 2004), but also because the web forums themselves can actively exclude participants and discourses. Rather than assume that web forums provide platforms that are open for all, we need to examine if there are obstacles to open and equal exchange in the institutional environment. To uncover the institutional boundaries we first have to address questions relating to the regulation of the forum: 1. What are the rules on the web forums? 2. How and by whom are these forum rules maintained? 3. How do participants of the forums perceive this regulation? There are different aspects to the organisation of web forums. First there are the rules as published on the forums (netiquette). To analyse how the netiquette affects the relative openness of forums, a discourse analysis is applied that identifies who and what is allowed and disallowed on a forum. Second, web forums have moderators who enforce the netiquette. The next step then is to analyse the different types of moderation. The types of moderation are identified in order to assess power held by forum moderators. Every web forum has its own rules regarding what is allowed in discussion (often called netiquette). I use the term "rules" here for the directives that are explicitly laid down in the netiquette. "Norms" refer to the implicit guidelines held by moderators and participants. A discourse analysis of the netiquette and examination of the moderators upholding the rules shows that online interactions in political discussions are very much bound by the rules and regulations of a forum. The analysis makes clear that there are limits to how open discussions are, even though web forums generally aim at providing an open environment. The netiquette is not intended to rigidly determine how participants should behave. It is focused on what is not allowed, and how participants should not behave. An examination of the netiquette and its maintenance, shows there are two types of platforms: (1) forums focusing on a broad inclusion and extensive freedom of expression, forming general platforms aimed at providing an open space for all; and (2) forums focusing on protection of a specific group of users, aiming to provide openness for a specific group. These two types of platforms do focus on openness but exclude certain forms of communication in trying to establish it, thus limiting the content of political discussions. Moderators have different means available to them for applying a web forum's rules ranging from participating in discussions to changing and removing texts and banning participants. These measures provide them with control over the production of texts, and as such moderators have the potential to obstruct an open and equal exchange. Thus it is important to look at how the rules are applied, by whom, and how transparent these issues are on the forum. In the forums examined in this study, the transparency on these matters leaves much to be desired. The rules are often poorly defined (if defined at all) and the way the moderators enforce them is unclear. The web forums are lacking in information pertaining to the appointment and presentation of moderators. On almost all forums it is unclear who the moderators are and how they were appointed. Only if one is completely immersed in a particular forum does it become clear how moderation takes place and by whom. Furthermore, there are few possibilities for users to appeal these matters. Users at times feel they are treated unfairly and are unjustly excluded from the discussion. The actions of the moderators may thus create a type of atmosphere in which some people feel more comfortable to voice their opinions than others. Through this, the possibility of different discourses interacting in a certain space may be limited. The way web forums are moderated (which is part of the context of online discourse) determines to a considerable extent their contribution to public discussion. By excluding texts and/or authors, moderators are impinging on the openness of online discussions. This observation would not have surfaced from looking solely y at web forums on a textual level (web forum postings). The context of the text thus needs to be included in the analysis to properly establish the democratic nature and potential of online discourse. Critical discourse analysis, with its attention for power in language, allows an analysis of the role of the moderator in debate, and shows more generally the obstacles to obtaining open debate. Participants' Views on Discussing Online The environment in which people produce and consume text is important for understanding the way in which information and opinion is exchanged. One part of this context has been addressed in the previous section by analysing responses to forum moderation. This, however, provides information on only part of the participants' perception of forums. How do the different participants of web forums T71AH7 f TA7/-\l-> t~/~\f*1 1 fY"l C f I—I A^AT t~i f~\ fl-lAT7 rAAr/^Ai "t )■ A 4" 1"» l~\ Ar\T71 VAr>TV>An-(- 11"» T»7ni/in f nAtr C > view those web forums? How do they perceive the environment in which they read and contribute postings? There are certain sets of norms that people take into account or adhere to when writing letters to the editor of newspaper, participating in a face-to-face political debate or phoning in to the radio. For all of the different ways of participating in QQ the public debate, there are different settings of interaction. What is the setting of interaction for online web forum discussions? How do people view this platform of debate? CO Users' views can be examined by in-depth interviews and by large-scale surveys. In the study I opted for the latter of these data collection procedures, since I was interested in the variety of perceptions on online political discussions. Open questions regarding the motives for discussing online were included, however, to allow for users to share ideas outside those pre-specified by the researcher. The research was conducted using an online survey instrument. Respondents were recruited through an online request to participate in the survey distributed on the web forums (233 respondents of different web forums took part in the survey). Even though the characteristics of the population of web forum users do not allow for generalisation to the national population (or, for that matter, to Dutch Internet users), the method allows for coming to considerable understanding concerning participants of large Dutch web forums (for issues of sampling in online studies see: Hewson, Yule, Laurent, & Vogel 2003; Witschge 2007). The way I propose to examine text production and consumption should provide information on the general perception of online discussions (though it does not provide information on the interpretations of specific texts). One of the main aims of the questionnaire was to determine how participants view online discussions and why they discuss online. Their views provide insight into the potential of web forums to provide a platform for an open and equal exchange in the public sphere. The results of the survey show that the users generally view web forums as open platforms for discussion in terms of the opinions expressed on the forums. The most prominent reason provided by the participants for discussing immigration on web forums is their wish to exchange ideas and access different opinions. The respondents attach great value to encountering a diversity of opinions as well as to the possibility to express their own opinions online. Participants consider web forums to be spaces where this can be found. The results of the survey show that respondents consider the forum they use to be open, and they regard it to be more open to differences than traditional media. This particular perception may make for a different type of discourse online than offline as it is determining the setting if the interaction; if people feel more open to utter their opinion, they may also do so more readily. Their perceptions are thus of importance when considering the democratic potential of the Internet. The question that then arises is what people actually do with the diversity of opinions online: are they open towards alternative positions? Survey responses suggest that this is not the case and that participants only rarely change their opinion as a result of the discussion. The analysis gives insight into the role online discussions can play in the democratic debate as well as into how people view this particular platform. This particular study shows that the online discussion is the main means of discussing politics for the respondents. It also shows that participants seek diversity, and experience online discussions as open. However, it has also become clear that participants generally do not change their opinions. This type of information on the environment in which online political discourse comes into being demonstrates both the potential and limitations of this discourse. As such the CDA approach provides an extra viewpoint on the interactions online. Representation and Inclusion The studies discussed here were conducted to establish whether online discussions are open, and in particular open to difference. In the previous section it became clear that web forum participants do seek diversity. The question is whether this diversity is also present in the text and lead to possible changes in social practice. One could examine this solely by looking at the text of online discussions, but since the openness of these discussions is often evaluated in comparison to the openness of more traditional media (as the participants of the discussion themselves do as well), I propose to adhere to that approach and compare the openness of the online debate to that of the debate presented in the newspapers. The Internet is often hypothesised to allow for more inclusive discussion, because everyone can voice her or his opinion. The supposition is that there are no or fewer gatekeepers, thus allowing for more and diverse people and positions to be represented online than in offline media. But to what extent are different actors and viewpoints included in online discussions? And how does this compare to the representation in newspapers? These questions can be examined by mapping the different types of actors in the offline and online debate and subsequently comparing the extent to which different positions are included in these debates. The specific issue that was analysed is that of "honour revenge."9 The discussions on the topic of honour killings were analysed over the period of one month. Particular attention was paid to which actors were represented. For the selected newspapers10 actors were seen as represented, either by being an author of an opinion piece, column or letter-to-the-editor, or by being quoted in the newspapers (for a similar method of coding see: Ferree et al. 2002). The actors were coded for relevant identity markers, such as gender, ethnicity (of "immigrant descent" or "native Dutch") and whether they are members of the political elite. In the online discussions all the participants were seen as actors in the debate, and where possible, they were coded for relevant indicators (gender and origin). In addition, their "status" within the forum was coded by including the date of y registration and the average number of postings. This type of analysis provides ^ information on who produces the text, that is, who is included in the debate. £ The offline and online discussion platforms contained, respectively, 30 and 139 contributions on the issue of honour killings. In the newspapers, a variety of actors in terms of gender and ethnicity could be discerned, but almost no citizens or other non-governmental actors were represented. However, aside from the dominance of members of the political elite, the newspapers were more inclusive to women and those of immigrant descent than the online debate. In the online debate there was little evidence that those of immigrant descent were represented at all. Many references are made regarding religious or ethnic minorities, but these minorities were not themselves represented. Native Dutch participants seemed to dominate even the discussion on a Moroccan-Dutch forum that was part of the study. That the newspapers seem to represent a more diverse public seems to contradict initial expectations concerning to the openness of online debates. But to what extent does this representation influence the content of the debates; does more diversity in terms of participants mean more diversity in terms of positions? Turning now to the content of the debate, I found that this was not the case. Online, ^ even though there is little diversity in terms of participants, more information is 0 C provided, more positions are considered, and alternative solutions to the problem are discussed. What can we conclude regarding the openness of the online discus- LH sion? Is it more or less open than the newspaper debate? Looking at diversity from different angles, the analysis of the producers of the text and that of the content of the debate (the text itself) present different results. Adding the context of the text allows for a richer (and more nuanced) understanding of the text itself and its implications for social practice. Going deeper into the social practice of the discourse, the case of honour killings is used to provide an in-depth analysis of the ways in which the online discourse excludes certain positions and groups. Here, two elements are taken into account. First, the discourse can be exclusive in that it suggests that certain participants and/or viewpoints are inferior to others or because it altogether ignores certain participants or their contributions. Second, exclusion and inequality in a debate can be established by distinguishing between "us" and "them," thus creating a division between those who are seen as belonging to this society/public/group, and those who are not. A discourse analysis allows us to examine the mechanisms of exclusions that are present in the debate. The practice of discursive exclusion in the online debate shows that the "other" is not considered to be a Dutch citizen. Moreover, the power to determine what is and what is not Dutch is not readily extended to the "other." Denying someone citizenship and thus the legitimacy to participate in the debate, does not only pertain to actual perpetrators of honour killings but seems to include immigrants and Muslims in general. This practice seems to foreclose the possibility of open and equal exchange. Connecting the implications of the discourse to the position of groups in society (the social practice) is an important and valuable part of the approach proposed in this article. A mere analysis of text in isolation would not have allowed such conclusions. Looking at representations as well as exclusions through text thus allows for a more thorough analysis and a richer picture. In/exclusion of Alternative Voices The study of representation presented above provides information on the production of the text, its content and the context in terms of the social practice. It shows that no alternative voices are (re)presented. The question of how open online discussions are begs an analysis of what happens when an alternative voice is presented. How do different voices interact online when alternative voices are present(ed) in the debate and to what extent is this interaction open and inclusive? This section is devoted to introducing the method that will allow us to answer that question. The case at hand is that of a specific alternative voice online: a blogger called Ertan. Some years ago, Ertan voiced his understanding of a young boy of Turkish descent who shot and killed his secondary school teacher. Ertan presented his contested message on three different online forums. The 139 reactions to Ertan's message were examined using a discourse analytical approach in order to identify different strategies that people adopt to deal with an unconventional voice. The analysis shows that even though an alternative or radical voice is expressed online, it is not successful in opening up a dialogue. Instead, the participants were unanimous in trying to find ways to exclude it. They do so, not by addressing the content of the message, but rather by trying to "eliminate" it. The other participants acknowledge neither the participant nor the content of the post. This means that even though the discussion platform initially allows for the voice to be included, the participants are not receptive. Thus, the alternative voice only has a formal, but not a meaningful access to the debate. As a result, no understanding of the other comes about; the debate and its participants lack openness. The technology may allow for the alternative voice to reside in the public domain, but what happens with this voice depends on the other "inhabitants" of this space. It is apparent from the contextual analysis that, even though the text itself suggests inclusion (the voice is represented), there is no open and equal exchange in the debate. This again points to the need to look at the context of the text and its social implications, rather than at text in isolation. I have focused until now on describing and explaining the practice of current online debate and examine the obstacles to open and equal debate. However, if we go back to the challenge set by Chouliaraki & Fairclough (1999, 136-137) introduced above, we also have to look for the potential that exists for such debate to come about. To see whether and if so how, dialogue is possible online and how people may seek to bridge existing differences, an additional discussion was analysed. Here, strong and alternative voices were present as well but a few participants were actually open, and tried to understand the other's position. Discussants determined the tone of the debate (and softened it) by acknowledging the presence of other discussants through greeting, particularly those with whom they disagreed, through testimonials, narratives and the sharing of personal experiences. The debate featured inclusion of difference and through the dialogue some level of understanding was established. This last study shows that the context in which the online discourse is presented is fundamental in understanding its role in democratic debate. It shows that even though the alternative voice may be heard, interaction might not occur but rather lead to (further) inequality, not only in debate, but also in wider society. The cony textual analysis also shows which factors may help to establish dialogue and as such may help to enable understanding and reduce inequality in debate as well as in society. Here, again it is established that CDA allows for a rich analysis of text in its context and pays proper attention to issues of power both in text and society. i 0 C > Conclusion In this article I have proposed to employ critical discourse analysis to analyse online political discussion in context. I have argued that we need to look beyond the textual content of online discourse when examining the potential of online communication for democracy. The discursive and social practice that provide the context of online political communication are often neglected in studies into online political communication. However, one cannot fully evaluate the potential of the frw i i ni iKli/i 14- i rmrn'^r' tKfi /irt-n^-nvl- I T4rV\mV\ Internet for opening up public discourse if one ignores the context in which such discourse is produced. This study has introduced an integrated approach that looks at all three dimensions of discourse: the discursive practice (the production and consumption of text); the text; and the social practice. The mixed method approach proposed in this article allows for an examination and evaluation of the discourse QQ in its context, thus enlarging the scope of the conclusions. The benefits of looking beyond the text have been illustrated by drawing from four empirical studies regarding online public debate about immigration in the Netherlands. First, the study of the online rules and regulations (by means of a discourse and feature analysis) showed how online political communication is bound by the netiquette of the discussion forum. Even though the Internet is often assumed to be boundless, online discourse is highly controlled. This affects the actual text produced online and curtails the role of web forums for democratic debate. The discursive practice was further examined through an online survey examining participants' views on online political discussions and trying to establish how open the participants perceived these discussions. The analysis showed that participants did view online discussions as open. Their main motive for discussing online was to exchange views, though they rarely changed their opinion as result of the discussions. Both the focus on openness and the lack of actual openness further demarcate the potential of online discussions for democratic debate and as such provide insights that complement insights to be gained from analysing the online texts. The third and fourth studies presented here concern the text and its implications for social practice (and, to a certain extent, also the discursive practice). The studies examined who participates in the online debate (in comparison to the debate presented in the newspaper), what the content of the discourse is and what the implications of the online political discourse are for the wider social context. These combined analyses provide a richer picture of the role of online debate in democracy than would be possible through an isolated analysis of the text. The analyses show that more diversity in voices does not necessarily translate into more diversity in discourse. Also, by focusing on exclusion as well as inclusion, it shows the boundaries of the discourse and pointed to issues of power and inequality. Finally, by focusing on how alternative voices are received, I have been able to demonstrate the difference between mere inclusion and proper interaction, and factors were described that enable dialogue. All of these findings, seeming contradictions, and nuances provide a richer and more situated picture of the Internet's potential for democracy than that resulting from an analysis of text in isolation. The discursive and social practice need to be incorporated into an overall research design if investigators desire a thorough understanding of online political communication. The CDA approach has allowed for this rich and contextually embedded picture to be painted. CDA is as much a theoretical as a methodological approach. It does not provide universal methodological tools for analysis of discourse. The methodological approach I proposed here should, in other words, be taken as an illustration of how issues of power and openness can be examined in online discussions rather than a fixed and case-independent set of procedures. It serves to illustrate the importance of examining the context of online discourse. A number of these contextual matters can be analysed in different ways than proposed here and some aspects of the analysis may benefit from additional methodological tools. Given the scope of this article, I necessarily highlighted some aspects and only briefly touch upon others. The main argument, nevertheless, was that the context of online discourse should be included in the examination, and the main aim was to show how crucial information is gained by a contextual analysis often neglected in a content-only study of online political discourse. In this manner, I hope to have contributed to understanding of modern-day practices of public debate as well indicated how research can discern obstacles to, and potentials for, open and equal exchange in online political discussion. Acknowledgements The research reported on in this paper was part of the author's PhD project funded by the Dutch Science Foundation (NWO). The author wishes to thank Nicholas Jankowski, Maarten Cleeren and the two anonymous reviewers of Javnost - The Public for their insights. Notes: 1. Here, I refer to all Internet technologies that are considered to enable democratic discourse. These include the web (and all its different technologies for discussion), e-mail, USENET and newsgroups. For an overview, see for instance Barnes (2002). 2. There are many other types of political uses of the Internet, such as online campaigning, online voting, citizen information online, and e-consultation. I do not address these here, but rather focus on the literature that is related to public sphere theory and deliberation. 3. See for instance (from 2000 onwards): Ó Baoill 2000; Dahlberg 2000; Gastil 2000; Jankowski & Van Selm 2000; Wilhelm 2000; Coleman & G0tze 2001; Gimmler 2001; Muhlberger & Shane 2001; Sunstein 2001; Tanner 2001; Brants 2002; Hagemann 2002; Papacharissi 2002; Price & Cappella 2002; Savigni 2002; Stromer-Galley 2002; Tsaliki 2002; Albrecht 2003; Jenkins & Thorburn 2003; Liina Jensen 2003; Janssen & Kies 2004; Kiss 2003; Papacharissi 2004; Trénel 2004; Dahlgren 2005; Wiklund 2005. 4. This is not to say that there are no studies at all that look at multiple aspects of public discussion. However, the way in which I propose to examine public discussion entails a more holistic approach. 5. For overviews and different approaches within the field of discourse analysis see: Chouliaraki & Fairclough,1999; Van Dijk 1997a, 1997b; Fairclough 1995; Gill 2000; Howarth 2000; J0rgensen & Philips 2002; Potter 2003; Torfing 1999; Wetherell, Taylor and Yates, 2001. 6. This three-dimensional model stems from Fairclough (1995). In subsequent work this three-dimensional model is replaced by an alternate, but slightly different conceptualisation of discourse; see J0rgensen & Philips 2002: 70-71. 7. A Dutch politician who openly expressed his contempt of Islamic culture. 8. There are a number of ethical considerations that should be taken into account when using data from these forums. The main question is whether it is ethical to use data available (publicly) online without informing the participants about such use. I have argued elsewhere (Witschge 2007) that it is ethical as I am examining only those discussions that can be viewed as part of the public sphere, drawing a parallel between web forum contributions and letters to the editor that are published in newspapers. Continuing this parallel, I quote the original text and their authors, whether or not nicknames are used. 9. The Dutch term of 'Eerwraak' not only includes so-called 'honour killings' but also refers to other forms of (physical) violence in order to 'restore' the honour of a family. 10. The main national paid newspapers were included in the sample (Het Algemeen Dagblad, NRC Handelsblad, De Telegraaf, Trouw, and De Volkskrant), as well as one free paper (Metro), and one regional paper with national distribution (Het Parool). References: Albrecht, Steffen. 2003. Whose Voice Is Heard in the Virtual Public Sphere? A Study of Participation and Representation in Online Deliberation. Paper presented at the Research Symposium "Information, communication, society," University of Oxford, September 17-20. U S i 0 £ * i * IR 0 C « Barnes, Susan. 2002. Computer-Mediated Communication: Human-to-Human Communication Across the Internet. Boston, MA: Allyn & Bacon. ^ Brants, Kees. 2002. Politics is E-verywhere. Communication, 27, 2, 171-188. 00 Brants, Kees. 2005. Guest Editor's Introduction: The Internet and the Public Sphere. Political Communication 22, 2, 143-146. Chouliaraki, Lilie and Norman Fairclough. 1999. Discourse in Late Modernity: Rethinking Critical Discourse Analysis. Edinburgh: Edinburgh University Press. Coleman, Stephen and John Gotze. 2001. Bowling Together: Online Public Engagement in Policy Deliberation. Dahlberg, Lincoln. 2000. The Internet and the Public Sphere: A Critical Analysis of the Possibility of Online Discourse Enhancing Deliberative Democracy. Unpublished Doctoral Dissertation, Massey University, Wellington. Dahlberg, Lincoln. 2001. Computer-mediated Communication and the Public Sphere: A Critical Analysis. Journal of Computer-Mediated Communication 7, 1. Dahlgren, Peter. 2004. Foreword. In W. Van de Donk, D. Rucht and B. D. Loader (eds.), Cyberprotest: New Media, Citizens and Social Movements, xi-xvi. London: Taylor & Francis. Dahlgren, Peter. 2005. The Internet, Public Spheres, and Political Communication: Dispersion and Deliberation. Political Communication 22, 2, 147-162. Dijk, Teun A. van, ed. 1997a. Discourse as Social Interaction (Discourse Studies: A Multidisciplinary Introduction, Volume II). London: Sage. Dijk, Teun A. van, ed. 1997b. Discourse as Structure and Process (Discourse Studies: A Multidisciplinary Introduction, Volume I). London: Sage. Eriksen, Erik O. and Jarle Weigard. 2003. Understanding Habermas: Communicative Action and Deliberative Democracy. Continuum: London. Fairclough, Norman. 1995. Critical Discourse Analysis: The Critical Study of Language. London: Longman. Fairclough, Norman. 2001a. Critical Discourse Analysis as a Method in Social Scientific Research. In R. Wodak and M. Meyer (eds.), Methods of Critical Discourse Analysis, 121-138. London: Sage Publications. Fairclough, Norman. 2001b. Language and Power (2nd ed.). Harlow, UK: Pearson Education Ltd. Ferree, Myra Marx, William A. Gamson, Jürgen Gerhards, and Dieter Rucht. 2002. Shaping Abortion Discourse: Democracy and the Public Sphere in Germany and the United States. Cambridge: Cambridge University Press. Gastil, John. 2000. Is Face-to-Face Citizen Deliberation a Luxury or a Necessity? Political Communication 17, 4, 357-361. Gijsberts, Mérove and Jaco Dagevos. 2005. Uit elkaars buurt: De invloed van etnische concentratie op integratie en beeldvorming. Den Haag: Sociaal en Cultureel Planbureau. Gill, Rosalind. 2000. Discourse Analysis. In M. W. Bauer and G. Gaskell (eds.), Qualitative Researching with Text, Image and Sound: A Practical Handbook, 172-190. London: Sage. Gimmler, Antje. 2001. Deliberative Democracy, the Public Sphere and the Internet. Philosophy and Social Criticism 27, 4, 21-39. Habermas, Jürgen. 1989. The Structural Transformation of the Public Sphere: An Inquiry into a Category of Bourgeois Society. Cambridge: Polity Press. Hagemann, Carlo. 2001. Who Discusses What with Whom: Participation and Issues at Two Political Discussion Lists on the Internet. Paper presented at the Euricom-Colloquium "E-Networks & Democratic Life," Piran, September 15-20. Hewson, Claire, Peter Yule, Dianna Laurent, and Carl Vogel. 2003. Internet Research Methods: A Practical Guide for the Social and Behaviourial Sciences. London: Sage. Howarth, David. 2000. Discourse. Buckingham: Open University Press. Jankowski, Nicholas W. and Martine Van Selm. 2000. The Promise and Practice of Public Debate in Cyberspace. In K. Hacker and J. A. G. M. Van Dijk (eds.), Digital Democracy; Issues of Theory and Practice, 149-165. London: Sage. Janssen, Davy and Raphael Kies. 2004. Online Forums and Deliberative Democracy: Hypotheses, Variables and Methodologies. Paper presented at the Empirical approaches to deliberative politics, European University Institute, Florence, May 22-23. Jenkins, Henry and David Thorburn, eds. 2003. Democracy and New Media. Cambridge, MA: MIT Press. Jorgensen, Marianne W. and Louise Philips. 2002. Discourse Analysis as Theory and Method. London: Sage Publications. Kiss, Balazs. 2003. Politics and the Forum: Are Topics and Fora Deliberative? Paper presented at the ECPR, Edinburgh, March. Koopmans, Ruud. 2004. Movements and Media: Selection Processes and Evolutionary Dynamics in the Public Sphere. Theory and Society 33, 3, 367-391. Liina Jensen, Jakob. 2003. Public Spheres on the Internet: Anarchic or Government-sponsored - A Comparison. Scandinavian Political Studies 26, 4, 349-374. Mann, Chris and Fiona Stewart. 2000. Internet Communication and Qualitative Research: A Handbook for Researching Online. London: Sage. Muhlberger, Peter and Peter Shane. 2001. Prospects for Electronic Democracy: A Survey Analysis, Version 1.0. Pittsburgh, Pennsylvania: H.J. Heinz III School of Public Policy and Management. Ó Baoill, Andrew. 2000. Slashdot and the Public Sphere. First Monday 5, 9. O'Hara, Kieron. 2002. The Internet: A Tool for Democratic Pluralism? Science as Culture 11, 2, 287-298. Papacharissi, Zizi. 2002. The Virtual Sphere: The Internet as a Public Sphere. New Media & Society 4, 1, 9-27. Papacharissi, Zizi. 2004. Democracy Online: Civility, Politeness, and the Democratic Potential of Online Political Discussion Groups. New Media & Society 6, 2, 259-283. Potter, Jonathan. 2003. Discourse Analysis. In M. Hardy and A. Bryman (eds.), Handbook of Data Analysis. London: Sage. Price, Vincent and Cappella, Joseph N. 2002. Online Deliberation and its Inluence: The Electronic Dialogue Project in Campaign 2000. IT & Society 1, 1, 303-329. Richardson, John E. 2007. Analysing Newspapers: An Approach from Critical Discourse Analysis. Basingstoke: Palgrave Macmillan. Robinson, John P., Alan Neustadtl, and Meyer Kestnbaum. 2002. The Online "Diversity Divide": Public Opinion Differences among Internet Users and Nonusers. IT & Society 1, 1, 284-302. Savigni, Heather. 2002. Public Opinion, Political Communication and the Internet. Politics 22, 1, 1-8. Schneider, Steven M. 1997. Expanding the Public Sphere through Computer-Mediated Communication: Political Discussion about Abortion in a Usenet Newsgroup. Unpublished PhD dissertation, MIT, y Cambridge, MA. Street, John. 1997. Remote Control? Politics, Technology and "Electronic Democracy." European Journal of Communication 12, 1, 27-42. Stromer-Galley, Jennifer. 2002. New Voices in the Political Sphere: A Comparative Analysis of Interpersonal and Online Political Talk. Javnost - The Public 9, 2, 23-42. Stromer-Galley, Jennifer. 2003. Diversity of Political Conversation on the Internet: Users' Perspectives. Journal of Computer-Mediated Communication 8, 3. Sunstein, Cass R. 2001. Republic.com. Princeton, NJ: Princeton University Press. Tanner, Eliza. 2001. Chilean Conversations: Internet Forum Participants Debate Augusto Pinochet's Detention. Journal of Communication, 51, 2, 383-403. Thomas, Graham and Sally Wyatt. 1999. Shaping Cyberspace: Interpreting and Transforming the Internet. Research Policy 28, 681-698. Torfing, Jacob. 1999. New Theories of Discourse: Laclau, Mouffe and Žižek. Oxford: Blackwell Publishers. Trénel, Matthias. 2004. Measuring the Quality of Online Deliberation. Unpublished manuscript. Tsaliki, Liza. 2002. Online Forums and the Enlargement of Public Space: Research Findings From a European Project. Javnost - The Public 9, 2, 95-112. «Warschauer, Mark. 2003. Technology and Social Inclusion: Rethinking the Digital Divide. Cambridge, ____MA: MIT Press. Wetherell, Margaret, Stephanie Taylor, and Simeon J. Yates, eds. 2001. Discourse as Data: A Guide for £ Ì 0 £ * i * (R il o ^ Analysis. London: Sage. Wiklund, Hans. 2005. A Habermasian Analysis of the Deliberative Democratic Potential of ICT- enabled Services in Swedish Municipalities. New Media & Society 7, 5, 701-723. _| Wilhelm, Anthony G. 2000. Democracy in the Digital Age: Challenges to Political Life in Cyberspace. New ^ York: Routledge. Witschge, Tamara. 2004. Online Deliberation: Possibilities of the Internet for Deliberative Democracy. In P. Shane (ed.), Democracy Online: The Prospects for Political Renewal through the Internet, 109122. New York, NY: Routledge. Witschge, Tamara. 2007. (In)difference Online: The Openness of Public Discussion on Immigration. Unpublished Doctoral dissertation, Universiteit van Amsterdam. Wodak, Ruth. 2001. What CDA is About: A Summary of Its History, Important Concepts and Its Developments. In R. Wodak and M. Meyer (eds.), Methods of Critical Discourse Analysis, 1-13. London: Sage Publications.