5. Investigating the Press Reform Debate

The term “paradigm repair” was used by Bennett et al. (1985) to describe “how journalistic self-criticism protects existing paradigms rather than confronts entrenched deficiencies and contradictions” (cited in Carlson 2015, p. 4). It refers to a situation in which metajournalistic discourse is used to protect press standards and values from scrutiny (Carlson 2012b, p. 267). The concept of paradigm repair was drawn from Kuhn’s (1962, cited in Reese 1990) work which linked creation with paradigmatic allegiances. Kuhn (1962, cited in Reese 1990, p. 392) stated that “paradigms provide examples rather than explicit rules” such that the paradigm is learnt “by engaging in the discipline”. As stated in Chap. 1, the notion of paradigm repair has been employed by previous scholars to examine journalistic metadiscourse on press “deviancy” in relation to objectivity (Reese 1990), fabrications (Hindman 2005; Carlson 2009), reporting errors (Cecil 2002), paparazzi (Bishop 1999; Berkowitz 2000), scapegoating (Berger 2008), media scandal (Carlson and Berkowitz 2014) and press standards (Thomas and Finneman 2014).

One news paradigm that is of importance to this book is the perception of the journalist as a crusader; one who uses the weapon of “the pen” to fight for justice for the less privileged; the voice of the voiceless and the provider of information to the public that enables them to hold the powerful to account—the watchdog of society. Franklin (1997, cited in Frost 2007) breaks this self-perception (and to some extent public perception) down into six norms which are journalism is a quest for truth, journalism is independent of government, newspapers are pluralistic organisations, journalists are independent of economic pressures, journalists are watchdogs and journalism creates a public sphere with the bottom line being that journalism is central to democracy. Closely linked to the “crusader image paradigm” is the “press freedom” or “press autonomy” paradigm. Press freedom is critical to journalism’s fulfilment of its crusader or watchdog role, particularly freedom from state interference. As stated earlier, from neoliberal perspectives, if journalists must call politicians to order and expose corruption, they must be independent of government (Deuze 2005). They are also allowed to use clandestine means to obtain information if that information is in the “public interest” (see Chap. 2 for more on public interest). This freedom enables the press to carry out investigative journalism as well as fulfil other watchdog and informative roles that will help enhance democracy (Waisbord 2013). The British press would frown at any effort or perceived effort to rob it of its autonomy and would likely use its power to control information to protect the freedom of the press and repair its image as a crusader.

This view of the journalist as a crusader is not without contestation. There are several instances where journalists have been accused of abusing their freedom (McQuail 2003, p. 81; Squire et al. 2005, p. 254). Scholars have contested the image of the journalist as a crusader based on these lines: journalists do not always tell the truth; their relationship with government is collusive; they are not independent of economic pressures because competitive markets cause them to throw ethics to the wind; they are not watchdogs but lapdogs; they are not pluralistic and do not serve as a democratic public sphere (Franklin 1997). Previous studies show that oftentimes when public outrage against press “deviancy” leads to the setting up of a press commission, steps taken by such commissions to check the abuse of press freedom have been interpreted in journalistic metadiscourse as a “threat to press freedom” and, by extension, a threat to democracy (Putnis 2000). Reese (1990), Cecil (2002), Carlson and Berkowitz (2014) and Thomas and Finneman (2014) examined how journalists’ metadiscourse followed certain patterns when they perceived threats to their paradigm. In line with Eason’s (1988, cited in Carlson 2015, p. 4) argument, these studies showed that the journalistic metadiscourses were defensive rather than self-critical.

In his study about how journalists in the USA responded to perceived threats to the objectivity paradigm, Reese (1990) examined three types of paradigm repair: “(a) disengaging and distancing the threatening values from the reporter’s work; (b) re-asserting the ability of journalistic routines to prevent threatening values from ‘distorting’ the news, and (c) marginalising the man and his message; making both appear ineffective” (Reese 1990, p. 390). Cecil (2002, p. 46) talks about a type of paradigm repair “in which the logic of journalism is reasserted in response to an outside challenge”. Referring to it as paradigm overhaul, Cecil (2002) found that while criticising challenges to the “objective news paradigm”, journalistic metadiscourse overhauled that paradigm, “reasserting objective news without altering or even questioning its underlying assumptions” (Cecil 2002, p. 47). He pointed out that “no changes were deemed necessary” (ibid.). The journalist(s) who were accused of deviant behaviour were described as a few “bad apples” and “the problem was “solved” by simply purging the reporters from the ranks” of good journalism (ibid.).

In this way, the journalistic metadiscourse sought to repair its paradigm by localising the bad behaviour and asserting the profession’s importance and achievements without considering making amendments to the paradigm. Alternatives to the paradigm were labelled deviant (ibid.). The process by which journalists strive to build walls to isolate deviant journalists from the “noble profession” have been described as boundary maintenance (ibid, p. 50). In addition to maintaining its boundaries through journalistic metadiscourse, the news organisation responds by either correcting or eliminating the bad apple(s) responsible for the mistake (Reese 1990, cited in Cecil 2002, p. 50). Cecil (2002, p. 50) emphasised that “the individual reporter, editor, and producers; not news organisations or the news paradigm itself, tend to receive the blame for breaks in the paradigm”.

While individualisation (also, ostracisation, localisation or bad apple) strategy has proved useful in the analysis of paradigm repair, limiting paradigmatic markers to individualisation of deviancy restricts the tools for analysis of paradigm repair. As Carlson and Berkowitz (2014) observed, paradigm repair interpretive strategies can also include extensions of deviancy from the individual to the general. The general can include a news organisation; a class of newspaper, for example, the tabloid; a media platform or even the whole media. Using as an example, the media coverage of the demise of two US regional newspapers, Rocky Mountain News and printed Seattle Post Intelligencer, Carlson (2012b, p. 267) demonstrated that paradigm repair can include situations where the perceived problem of an individual newspaper is interpreted by the press as a reflection of the challenges faced by all newspapers. Here, the perceived threat to the printed press paradigm was “repaired” by re-asserting the “importance and superiority” of the printed press over other forms, especially online news forms. Carlson referred to this extension from the individual to the general as “second—order paradigm repair” (ibid.).

Another case for expansion of paradigm repair to include generalisable interpretive strategies can be found in Carlson and Berkowitz (2014) where the deviancy (phone hacking) of an individual newspaper, the News of the World, was interpreted by both journalistic and non-journalistic actors as a manifestation of the deviancy of all printed press in the UK; and the Valerie Plame’s case where the misconduct of Miller was translated as the deviance of the press (Carlson 2012a, pp. 111–137). Carlson and Berkowitz (2014, p. 392) described this extension of deviancy from the individual to the general as “synecdoche deviancy”, a term he borrowed from the figure of speech; synecdoche, which means using part to represent the whole. He says, “with journalism, this may refer to efforts to extend a scandal from its immediate context to incorporate a broader set of news practices—such as using a scandal at one tabloid newspaper to extend to the whole of tabloid newspapers” (ibid.). I argue that both individualised (ostracisation, localisation) and generalisable (synecdoche) deviancy can emerge from the same incident and could occur at different stages of, for example, a scandal.

Synecdoche deviancy in journalistic metadiscourse is often triggered by scandals in which the press is the culprit. Examples of such scandals are the Valerie Plame’s case of 2003; the Pulitzer Prize–winning “news” story in the Washington Post, written by Janet Cooke about a non-existent eight-year-old heroin addict, published in 1980; the Time magazine cover “photo illustration” of O.J. Simpson at the time of his murder trial in 1994, with darkened skin tone; the “Rathergate”—a scandal in which the CBS anchor-man Dan Rather was accused of using fake documents in a report on George W. Bush dodging National Guard duty (Bettig and Hall 2012, p. 16); and more recently, the News of the World phone hacking scandal. Scandals can encourage public discussions on sensitive and controversial issues (Lull and Hinerman 1997, p. 1). According to Lull and Hinerman, contemporary media scandals provide “a clear and compelling entry point for criticisms of the media in society” (ibid.). In the case of the News of the World phone hacking scandal, the controversy generated important questions relating to how the media cover themselves and the implication of their manner of coverage for democracy.

In line with Carlson’s argument for the use of the notion of paradigm repair to study synecdoche deviancy in journalistic metadiscourse, Thomas and Finneman (2014) used paradigm repair to study the media coverage of the Leveson Inquiry, an offshoot of the phone hacking scandal (see Chap. 3 for more on the Leveson Inquiry). Their study sample consisted of “editorial comment in mainstream national daily and Sunday newspapers on the Leveson Inquiry from its inception to the conclusion of its hearing phase”. Using the media coverage of the hearing stages of the Leveson Inquiry as its study sample, Thomas and Finneman (2014, p. 172) summarised interpretive patterns used in previous analysis of metadiscourse into four main strategies:

Localisation, which is also known as ostracisation, individualisation and the bad apple interpretive strategy, is a type of deflection strategy. The press can use different strategies of deflection to exonerate themselves of blame by deflecting it to others. Deflection strategies are also referred to in this book as self-exoneration strategies. From their findings, Thomas and Finneman (2014, p. 172) concluded that the journalistic metadiscourse during the hearing stages of the Leveson Inquiry revealed “an institutional ideology that is quick to assert rights but largely resistant to notions of attendant responsibilities”. Though the authors’ categorisation is very relevant to the study of journalistic metadiscourse, their categories do not make room for durational modes of interpretation (ibid.).

According to Zelizer (1994, cited in Cecil 2002, p. 51), journalistic metadiscourse takes place on two planes: the local and the durational modes of interpretation. While the local mode of interpretation is concerned with the immediate occurrence and how it is made meaning of in journalism’s interpretive community, the durational mode of interpretation “offers a historical perspective, a discussion of past occurrences, which is often reinterpreted to fit into a more localized frame” (ibid.). Journalists discuss past stories such as Princess Diana’s death and paparazzi (1997), the topless princess photo (2012) and the Watergate scandal (1970), and talk about how this fits in with the present occurrence and what that means for journalism and democracy. Such discussions are often used to reinforce the boundaries of journalistic practice. This is what I refer to, in paradigm repair studies, as the strategy of historicisation (White 2004, p. 98; Mumford 2009, p. 72).

“Historicisation” is a term I borrowed from dramatic theory where German playwright Bertolt Brecht (1898–1956) used it to describe the following theatre scenarios:

Historicisation becomes a paradigm repair strategy when similar techniques as those outlined by Brecht are employed by the press in an attempt to repair its paradigm (ibid.). The concept of paradigm repair strategy assumes that the press would only protect its paradigm(s); but the press can also challenge journalistic paradigms (Carlson 2012a). Interestingly, the strategy of historicisation can also be used to challenge an existing paradigm. In such a case, it is not a paradigm repair but a paradigm challenging strategy. The strategy of historicisation can, therefore, be defined as an interpretive approach in which history is used to strengthen or explicate contemporary meaning-making in journalistic metadiscourse, in an attempt to protect or challenge a journalistic paradigm.

In journalistic metadiscourse, historicisation is mostly (though not exclusively) located in the opinion sections of newspapers where it is used to strengthen diverse arguments. Historicisation would most often express the newspaper’s position in an argument. For instance, newspapers that propagate the neoliberal ideology can use the strategy of historicisation to warn against state intervention in press regulation while those advancing social democratic perspectives can use the strategy to stress the need for state intervention. Historicisation can be used to either call for or oppose press reform. It can also be used as an element of drama; to assert the media’s importance; or to affirm that the press has gone too far too often. Historicisation is a broad paradigmatic approach that can embrace other paradigm strategies such as self-assertion and individualisation. Adding the strategy of historicisation to the four paradigm repair strategies—the strategies of threat to the paradigm (catastrophisation), self-assertion (affirmation), minimisation and individualisation (localisation; bad apples)—enabled me to investigate both historical and local modes of interpretation in the press coverage of the debate that arose from the NoTW phone hacking scandal and the Leveson Inquiry. Having discussed the terms and strategies that I engaged with in my analysis of the coverage of the debate that followed the NoTW phone hacking scandal, the following section explains the model of content and discourse analyses used for my investigation.

Content analysis is the systematic and objective analysis of texts such as can be found in newspaper articles, television clips, books, adverts and so on (Holsti 1969, p. 14, cited in Stemler 2001, p. 17). In content analysis, textual components (words, phrases, images, etc.) relevant to the findings of one’s research are counted, recorded and then calculated with the use of statistical methods (Riffe et al. 2005, p. 3). The understanding is that the results when analysed can provide answers to the question(s) posed by the study. Content analysis is suitable for analysing huge volumes of texts (Mosdell and Davies 2006, p. 98). This made it the most suitable research method for my investigation which examined more than 800 newspaper articles. In my research, content analysis was used to measure the use of sources, the hierarchy of importance accorded to different issues of concern, the range of alternative views, the dominant themes and the extent to which paradigm repair strategies were used in the coverage of the media policy debate.

Measuring sources enabled me to evaluate the sources used and ones most frequently quoted. This provides empirical data on how the media used sources of information in their coverage of the press reform debate. In a democracy, the normative expectation is that all stakeholders in a debate would have proportionate access to the public sphere. My decision to investigate sources in the media policy debate is in response to this expectation and the claim that ordinary citizens and sources critical of the press are allotted a weak position of access to the media (Galtung and Ruge 1965, cited in Harcup and O’Neill 2010, p. 270). Using content analysis, I was able to identify the dominant theme in each newspaper and the whole coverage; ascertain whether the press gave limited coverage to views that were critical of them; measure attributions of blame and alternative views; and gauge the degree of manifestation of different paradigm repair strategies in the journalistic metadiscourse. In sum, content analysis helped to provide statistical data on how the British press covered the media policy debate that arose from the News of the World phone hacking scandal and the Leveson Inquiry. A coding scheme was designed to enable me to input the data for my content analysis.

My study sample comprised all news articles on the press reform debate in six of the top ten British national newspapers (based on combined print and online readership figures for April 2011 to March 2012—Source: NRS PADD 2012): two newspapers from each category of the main newspaper classification in the UK. My decision to do only two from each category is for the purpose of manageability, considering the depth of analysis and available period of study. As mentioned in Chap. 1, I chose to look at national papers because of their nationwide reach. The national newspapers in Britain are categorised in terms of social class, although this classification does not always reflect reality (McNAir 2000, p. 14). The categories are the broadsheets, mid-market and the tabloids (Williams 2009, pp. 9–10). These categories have different target audiences and diverse manners of coverage. This was taken into consideration during the analysis of the coverage. However, these categories were bypassed where a phenomenon cuts across paper classifications.

The broadsheets, also known as quality newspapers, deliver hard or “serious-minded” news content. They are regarded as “the most information dense of the print media” (McNair 2000, p. 16). Their target audience consists of the upper- and middle-class people “with higher levels of income, educational attainment and social status” (Williams 2009, p. 9). They are often referred to as the elite titles (McNair 2000, p. 16). At the time of my research, the broadsheet newspapers included The Times, The Sunday Times, Daily Telegraph, Sunday Telegraph, Guardian, Observer, Independent, Independent on Sunday and Financial Times. The broadsheet newspapers examined are the Daily Telegraph and Guardian. The choice of these two quality newspapers was based on their wide reach and consequent potential to impact society. The Guardian had the biggest combined (print and digital data) monthly readership for broadsheet newspapers in the year to March 2012 with a total of 8.95 million readers, followed by Daily Telegraph with 8.82 million readers (NRS PADD 2012; Halliday 2012). Both papers were, thus, the most read British national quality newspapers within the period of the debate that followed the phone hacking scandal. The role Guardian newspaper played in exposing the extent of the scandal adds to its importance in the sample.

Next in line are the mid-market titles which at the time comprised Daily Express, Sunday Express, Daily Mail and Mail on Sunday. Their target audience is believed to be middle and upper working-class people. There is some confusion with this classification because papers in this group were former broadsheet newspapers (The Mail and Express newspapers) which changed from the broadsheet to tabloid print format (McNair 2000, p. 14). Some scholars use the term “red-top tabloids” to refer to the group commonly known as the tabloids, in order to differentiate them from the “broadsheets” in tabloid format. Representing the mid-markets in the study are Daily Mail and Daily Express. Their Sunday editions, as well as the Sunday publications of all other newspapers studied, were excluded from my research to reduce it to a manageable size. The Daily Mail also has great potential to impact society; it came next to the Sun as the paper with the second largest combined monthly readership across all classes of UK newspaper in the year to March 2012 with a total of 16.4 million readers (NRS PADD 2012; Halliday 2012). Though the Daily Express was far behind in readership figures (4.6 million readers), it serves as the only other mid-market newspaper, and examining it evens up the number of newspapers being studied to two newspapers per category of British national newspapers (ibid.). Providing an alternative newspaper per category of newspaper examined makes available opportunities to investigate whether a phenomenon was characteristic (or not) of a newspaper category.

The “popular” press, red-top tabloid or tabloid titles are at the end of the spectrum in terms of social status. They are known to have a high level of readership despite the fact that they publish less “serious” and more celebrity, sensational and entertainment-style news. At the time of this study, the tabloid newspapers were The Sun, Sun on Sunday, Daily Mirror, Sunday Mirror, Daily Star, Daily Star Sunday and People. My study examined The Sun and Daily Mirror. The Sun’s reach is significant being that it had the highest combined monthly readership across all categories of UK newspapers in the year to March 2012 with a total of 17.8 million readers (ibid.). The fact that The Sun newspaper belongs to the owner of the defunct News of the World strengthens its relevance to the study sample. It afforded me an opportunity to examine how one of Rupert Murdoch’s papers covered the misdemeanour of one of their own. The Daily Mirror’s 10.6 million monthly combined readership also made it a paper to be reckoned with; the figure makes it third in the ranking of overall UK national newspaper monthly combined readership in the year to March 2012 (ibid.). As previously stated, all six newspapers examined are among the top ten in terms of combined readership of national newspapers in the UK (ibid.). It is important to note that with the current trend of “tabloidisation” (the “dumbing down” or “going down market” of the more rational press to the sensational in order to attract a numerically larger audience), the lines are blurring among these three categories of newspaper (Williams 2009, pp. 9–10).

All news articles on the media reform debate in the six national daily newspapers were examined. The period of coverage was from 14 November 2011 (when the hearing began at the Leveson Inquiry) to 14 November 2013 (the aftermath of the Privy Council’s approval of a Royal Charter on press regulation). This two-year period falls within the time frame when media coverage of the press reform debate was at its peak in the UK (Macfarlane and Torpey 2012, n.p; Independent2013). Although editorials are where the newspaper’s opinions are often heard (Hindman 2003, p. 671), I decided against limiting my research to editorials because as Wahl-Jorgensen (2008, p. 67) pointed out, “in the British context … expression of judgements and opinions is frequently not limited to the op-ed and editorial pages, but increasingly pervades every section of the newspaper”. Therefore, limiting the study to editorials risks leaving out interpretations of the debate that featured in the news section of the newspapers. My data, thus, included both opinion and news articles that captured the media policy debate which followed the NoTW phone hacking scandal.

The news articles were obtained from Nexis UK, an electronic archive service with full text access to all UK national newspapers. Using the search terms “press regulation” or “press laws” or “public trust” or “media ownership” or “public interest” or “privacy” (anywhere in the text) and “Leveson” or “News of the World” or “phone hacking” (anywhere in the text), my search produced a total of 1485 news articles. A broad range of issues on the press reform debate including articles relating to the Independent Press Standards Organisation (IPSO), the Press Standards Board of Finance (PressBoF), Hacked Off, the Royal Charter and the Privy Council were represented in the result. After cleaning the sample by deleting repeats and unrelated stories, the sample was reduced to 870; 323 from Guardian, 199 from Daily Telegraph, 173 from Daily Mail, 28 from Daily Express, 96 from the Sun and 51 from Daily Mirror. The large reduction in the number of articles from 1485 to 870 was largely due to the high number of duplicate articles in Nexis UK, especially with articles from Guardian newspaper where the results, at the time of study, contained articles from both their online and print versions, despite excluding websites through the search preferences. My research examined written content only. Although a study on the visuals would also be interesting, that is outside the scope of this study. This study is quite broad and excluding visuals helped to make it a manageable project. Visual analysis of the coverage would be an interesting focus for future study.

Berelson (1952, p. 18, cited in Richardson 2007, p. 15) emphasised the characteristic of content analysis as an objective research procedure, free from the researcher’s interference. This “objectivity” requirement of content analysis also requires that the research be done in such a way that it can be replicated by anyone who chooses to do so (Krippendorff 2004, pp. 18–19). To take care of this requirement, a coding sheet was drawn up along with guidelines that helped to make the study replicable. The coding sheet was tested and re-tested by two trained postgraduate student coders. Thirty stories randomly selected from the study sample were tested until the overall percentage agreement reached an average of 95.9 per cent, with the lowest variable reaching 80 per cent agreement. The high level of percentage agreement across all variables helps to guarantee that this research can be replicated. The calculations were made using ReCal2 0.1 Alpha (dfreelon.​org). ReCal2 is an online reliability calculator for two coders which calculates intercoder reliability coefficients for nominal data and produces results for percentage agreements. The result of my intercoder reliability test was Krippendorff’s Alpha 0.822. A codebook that explains each variable was designed to enhance comprehension of the coding sheet. Though content analysis has huge benefits, among which are its cost effectiveness, unobtrusiveness and replicability (Berelson 1971, p. 18; Krippendorff 2012), the results it produces can be problematic because statistics from quantitative measurements can be interpreted out of context (Richardson 2007, pp. 15–18). To take care of such shortcomings, discourse analysis was used as a supplementary method to content analysis

Discourse analysis is multifarious and so are its definitions (Wodak and Meyer 2009, p. 24). van Dijk (1988, p. x; 1998 cited in Devereux 2007, p. 174) explains that discourse analysis allows the analyst to investigate the underlying meaning of words. Gee and Handford (2012, p. 1) advanced this understanding by defining discourse analysis as “the study of language in use”. Discourse analysis (DA) is “the study of the meanings we give language” and what that does within a particular context (ibid.). This is based on the “understanding that language does not passively reflect our experiences” but can be manipulated (consciously or unconsciously) by the powerful to advance their interests (Phillips and Jorgensen 2002, p. 63). “Powerful” within the context of this study includes those with access to the media’s public sphere and those upon whom the public has conferred the authority to both introduce and coordinate discourses within this public space (Phillips and Jorgensen 2002, pp. 2–3; Hall 1997). Some approaches to discourse analysis are largely linguistic, some are more focused on the construction of “themes or images” in texts while others are interested in linking language to social and cultural issues of contention with the aim of locating the resultant social consequences (Gee and Handford 2012, p. 1). The latter, which is known as critical discourse analysis, is the approach used for my research. Though discourse analysis was not the main method for the research, some of its principles were used to explicate my findings. Hence, the need to briefly discuss critical discourse analysis.

Critical discourse analysis, hereafter referred to as CDA, is diverse (Wodak and Meyer 2009, p. 5). Forms of CDA include the Foucauldian CDA and dispositive analysis (Jager and Maier 2009); the social cognitive approach (van Dijk 2009); the social psychological approach (Wetherell and Potter 1992); the discourse-historic approach of the Vienna School (Wodak and Meyer 2009) and Norman Fairclough’s approach to CDA (Fairclough 1992a, b, 1995a, b, 2005). I used principles from Norman Fairclough’s approach to CDA because unlike some models of CDA that focus primarily on linguistic analysis, Fairclough’s CDA makes room for the analysis of power relationships in communicative discourse in relation to wider social and cultural structures (Leifeld 2016, p. 39). Fairclough’s approach to CDA is beneficial for this study because an investigation into how the press covers debates about their policy and the implication of their manner of coverage for media accountability and the sustenance of democracy involves investigating the distribution of communicative power between the press and other stakeholders in the media policy debates. Fairclough expects that by using his approach to CDA, people can contribute to social change along the lines of more equal power relations in communicative discourse (Fairclough and Wodak 1997, p. 258). One normative expectation of CDA is to promote democracy by pointing out non-egalitarian discourses so that steps can be taken to make them democratic (Phillips and Jorgensen 2002). Though some aspects of van Dijk’s (1988, 1998, 2009) approach to CDA can be useful in this study (e.g. its investigation into communicative power and ideology analysis), the systematic and well-developed analytical construct of Norman Fairclough’s CDA and its emphasis on social issues as against van Dijk’s (1988, 1998, 2009) focus on the political gave it an edge in this study. Fairclough’s three-dimensional analytical model (discursive, linguistic and social practice) makes his CDA a good supplement to my content analysis.

As stated earlier, CDA was a supplementary method of analysis in my research; it was used to elucidate the results of my content analysis and for the analysis of how paradigm repair strategies were used in the coverage. News articles on the debate were scrutinised in search of words, linguistic devices and discursive patterns that could reveal underlying meanings. Fairclough’s (1995a, b, 2005) style of linking language-use to social practice was also employed in my analysis. For example, results from the coverage of the debate on media ownership were linked to the structure of media ownership in a democratic society. One unique feature in this style of mixed methodology is that there is no specific number of articles set aside as the sample for discourse analysis. Feltham-King and Macleod (2016, pp. 1–9) used a similar pattern of mixed methods when they used content analysis to supplement discourse analysis. The flexibility with which I could use principles drawn from CDA to analyse the data from content analysis at various points of my analysis is a feature that attracted me to this method. Supplementing content analysis with critical discourse analysis allowed me to interpret statistics based on the context within which they appeared, enabling in-depth analysis of the results of my investigation. It also enabled me to examine the construction and extent of usage of different paradigm repair strategies.

Following the lines of Fairclough’s three-dimensional analytical model (linguistic [grammar], discursive and social practice), I analysed linguistic devices such as “you-centeredness”, metaphors, hyperboles and adjectives to unearth their underlying meanings as suggested by Fairclough (1992b, pp. 158–194). This helped to affirm the use of some paradigm repair strategies in the coverage of the press reform debate. For example, my analysis of doom-laden adjectives in the media’s description of measures aimed at reforming their policy helped to affirm the use of “threat to the paradigm” strategy. On the discursive level, I identified the “us” and “them” pattern of discourse which attempted to portray the “us” (the press) as good and the “them” (campaigners for stringent press reform) as bad. I scrutinised discourses to see if arguments were based on the neoliberal or social democratic theory. I also investigated discursive patterns including silences in media texts and their implication. In terms of social practice, I analysed texts in my study sample based on the social, economic and political conditions under which they were produced. For instance, I linked the journalistic metadiscourse on media ownership to issue of media ownership concentration in the UK.

In this study, Fairclough’s three-dimensional model of analysis was not used chronologically but at various points of my analysis. Despite its richness, CDA does have some weaknesses. Scholars have argued that the fact that there is no specific way of selecting the study sample for CDA makes studies done with CDA subjective and prone to researchers’ interference (Blommaert 2005). This weakness is addressed in this study because CDA was used only to complement my content analysis. Consequently, all discourses analysed were drawn from the study sample for my content analysis which was chosen through a relatively objective method. Complementing content analysis with critical discourse analysis enabled me to provide comprehensive and comprehendible analysis of the coverage of the media policy debate.

This chapter presented the method used for my investigation into how the media represented the debate that arose from the News of the World phone hacking scandal and the Leveson Inquiry. Content analysis was supplemented by critical discourse analysis to provide comprehensive data on how the media cover debates about their policy. The chapter clarified the meanings of the terms “metacoverage”, “metajournalistic discourse” and “journalistic metadiscourse”. I called for a review of the definition that limits metacoverage to the increasing trend of journalists covering themselves during political campaigns (Esser et al. 2001). I argued that based on the composition of the word, it should refer to all forms of self-coverage by the media. Going by my definition of metacoverage, the media content analysed in my research falls within the category of metacoverage in the field of journalism, what Brin and Drolet (2009, p. 271) described as journalistic metadiscourse. I explained that journalistic metadiscourse is what Carlson (2015) referred to as metajournalistic discourse on journalistic platforms.

The concept of paradigm repair can be used to analyse journalistic metadiscourse on press reform. Journalists use paradigm repair strategies to protect press standards and values from scrutiny (Bennett et al. 1985, cited in Carlson 2015, p. 4). The paradigm repair strategies discussed in this chapter include “threat to the paradigm” (warnings of attacks on journalism), self-assertion (affirming the importance of journalism to democracy), minimisation (downplaying the significance of the phone hacking scandal and questioning the legitimacy of measures aimed at checking press irresponsibility), individualisation (localising the damage to the community to acts committed by a few journalists) and historicisation (using history or collective memory to repair (or challenge) journalistic paradigms).

I examined all news articles on the media policy debate in the Guardian, Daily Telegraph, Daily Express, The Sun, Daily Mirror and Daily Mail; from 14 November 2011 when the Leveson Inquiry was set up, to 14 November 2014—the aftermath of the sealing of the cross-party Royal Charter on press regulation. A total of 870 news articles were examined: 323 from Guardian, 199 from Daily Telegraph, 173 from Daily Mail, 28 from Daily Express, 96 from the Sun and 51 from Daily Mirror. A coding sheet, a codebook, Nexis UK database and the Statistical Package of the Social Sciences (SPSS) were used for my primary investigation. Results from this research are used to support my analysis of how the media cover debates about their policy.