1.3 Methodological orientation

13 05 2009

***This is an outdated version. Please do not cite or reproduce in any way.***

LE attempts to connect micro-level phenomena to wider public discourses. This orientation towards illuminating socially relevant issues in situated processes of social action derives from the combination of linguistic and ethnographic modes of analysis, the effect of which is eloquently characterized as:

i. ‘tying ethnography down’: pushing ethnography towards the analysis of clearly delimitable processes, increasing the amount of reported data that is open to falsification, looking to impregnate local description with analytical frameworks drawn from outside […].

ii. ‘Opening linguistics up’: inviting reflexive sensitivity to the processes involved in the production of linguistic claims and to the potential importance of what gets left out […].

(Rampton et al. 2004: 4)

1.3.1 Tying ethnography down

Running with this characterization, this study ties ethnography down to the analysis of newsroom practices of inception, enactment and inscription, as outlined in section 1.1.2 above. The institutional, multimodal, collaborative and computer-mediated nature of these practices calls for a holistic approach to news production research. I try to achieve this goal by complementing traditional fieldwork methods – participant observation, interviews, document analysis – with computer-assisted writing process analysis. Writing process analysis

By adopting some of the methods of writing process analysis, I significantly increase the amount of data “open to falsification”. One of the signal strengths of keystroke logging is that these applications open up aspects of the writing process that were previously inaccessible. Until fairly recently, analyzing the writing process was very much a post hoc enterprise. At the risk of generalization, you could observe writers, interview them or have them reflect on their practices. Alternatively, you could compare source texts, style guides or draft versions with final versions. With the emergence of digital writing tools such as keystroke logging and screen recording software, it became possible to track and compare what journalists say they do to how they do what they say. The work of Daniel Perrin (Perrin 1999; Perrin 2001; Perrin 2003a; Perrin 2006) is groundbreaking in this respect.

While writing process research in the tradition of cognitive psychology (MacArthur, Graham & Fitzgerald 2006) is primarily concerned with cognitive theories of text production, it has been used to study writing in PR (Sleurs, Jacobs & Van Waes 2003), academic (Van Waes & Schellens 2003) and workplace settings (Perrin 2003b). Throughout this book, writing process analysis is used as a window on the journalistic (reproductive) writing strategies: how information from press releases and news agency copy is transformed into a print newspaper article.

For the purpose of my study, writing process analysis is significant in at least three ways. First and foremost, keystroke logging and screen video applications generate second-by-second records of news writing, allowing analysts to focus on micro-level events which can then be linked up with macro-level concerns. In Chapter 2.3 for instance, I illustrate how writing a lead speaks to the technological mediation of the news production process.

Second, data generated by keystroke logging tools decrease the reliance on ethnographic authority, i.e. making theoretical claims based on ‘having been there’ (Hammersley & Atkinson 1983). One of the criticisms of ethnography is that its findings have limited validity; ethnographies cannot be falsified (lived reality is unique; it can never be duplicated) nor do their findings lead to empirical generalizability or abstract knowledge (e.g. the laws of physics). As Fielding and Schreirer observe (2001: 46):

when analyses are challenged, qualitative researchers […] defend their interpretation not by adherence to systematic, established, externally-validated analytic procedures but by the (usually unassailable) fact that “they were there”. They did the fieldwork, they collected the data, therefore they have the “best sense” of what the data may mean.

In my experience, the integration of writing process analysis not only focused my analytical gaze, but also forced me to plan, test, revise and follow-up “systematic field strategies and […] analytic procedures [in order] to constrain self-indulgent idiosyncrasy” (Rampton et al. 2004: 2-3). In Chapter 1.4, I describe my research protocol and discuss methodological strengths and weaknesses and analytical affordances of computer-assisted writing process analysis.

Third, Inputlog, the keystroke logging application I used, generates datafiles for statistical, text, pause and mode analyses (Leijten & Van Waes 2006). The multidimensional nature of these data offer various analytical inroads, both qualitatively and quantitatively. While most of the chapters in Part II offer qualitative case study illustrations, Chapter 2.7 offers a quantitative analysis of reproductive newswriting, suggesting “what to see” (churnalism vs news sourcing) but also providing a sense of “where to look” in follow-up research. Describing local rationalities

A final way in which I have tried to tie ethnography down is by infusing descriptions of local rationalities with concepts drawn from, inter alia, discourse analysis, folklore studies, journalism studies, social theory and economics. The goal here is, in the words of Clifford Geertz (1983), to combine ‘experience-near’ and ‘experience-distant’ concepts. Experience-near concepts are those which journalists use to explain what they do, for example newsworthiness or objectivity. An experience-distant alternative would be one that scholars use to explain journalism, for example epistemological positioning (Bednarek 2006). In Chapter 2.5, I draw on a journalist’s introspective analysis of his own news production practices to compare my interpretation of his writing process.

1.3.2 Opening linguistics up

This study opens linguistics up by broadening its analytical scope. Specifically, the linguistic claims made here about news production as an institutional process of decontextualization and recontextualization should be read against a disciplinary tendency of skating over the production process in favor of the analysis of news products, i.e. the various textual artifacts of news journalism: interviews, editorials, feature articles, headlines and the like[1]. As argued in Section 1.1.1, I try to fill in a blind spot in news scholarship by examining the discursive processes that shape the news product. I am not alone in this endeavor. On the contrary, under the heading of NewsTalk&Text, a number of colleagues in the UK, Italy, Belgium, Germany and Switzerland are calling for a linguistics of news production which (i) recognizes the centrality of the production process to the analysis of news and (ii) claims that text-only analyses yield weak hypotheses. Galtung & Ruge’s ‘news values’ are a case in point (NewsTalk&Text 2009: 6):

The ineffectiveness of relying solely on journalistic texts or a corpus of news stories alone to make claims about the work of journalists is evident in critiques of Galtung & Ruge (1965). They analysed media reports about three international crises to devise a list of ‘news values’ which, they suggested, were employed by journalists in gauging an event’s newsworthiness and could be used as predictors of what would be more likely be reported as news. However, more recently Harcup & O’Neill (2001) have argued that some of Galtung & Ruge’s news values are in fact a product of the way in which events are written about – the way that journalists construct news – rather than the characteristics that an event needs to possess in order to be reported.

Put differently, I try to open up linguistic analysis of news discourse by extending the focus from what the textual characteristics and discursive functions of economic print journalism are to how journalists source, negotiate and write economic news. In doing so, I work with four methodological assumptions concerning case study logic (, data triangulation (, discourse representation ( and secondary analysis ( The ethnographic logic of case study methodology

As Small writes (2009: 26), the advantage of ethnographic work is that “a well executed single-case study can justifiably state that a particular process, phenomenon, mechanism, tendency, type, relationship, dynamic, or practice exists”. It should be clear at this point that the central aim of the case studies I present in Part II of this book is to provide empirical evidence for the complex intertextual trajectory that defines news production. Note that this aim implies “logical rather than statistical inference, for case rather than sample-based logic, for saturation rather than representation” (2009: 29). Instead of quantifying hypotheses, which I largely refrain from,  I offer qualitative empirical statements that revolve around the journalistic reliance on source texts, the contingency and creativity involved in news production and the added value of professional journalists’ mediation of economic news. The underlying logic of this qualitative approach is captured by Fielding & Schreier (2001: 49):

qualitative work can assist quantitative work in providing a theoretical framework, validating survey data, interpreting statistical relationships and deciphering puzzling responses, selecting survey items to construct indices, and offering case study illustrations. In some cases the theoretical structure itself is a product of field experience. The rationale of data triangulation

Having pointed to the limitations of text-only analyses of news production processes in Sections 1.1.1 and 1.3.2, I infuse analyses with different sets of data in an attempt to ‘know more’ about news production. Agreeing with Moran-Ellis et al. (2008) that triangulation is

an epistemological claim concerning what more can be known about a phenomenon when the findings from data generated by two or more methods are brought together

my research design employs multiple methods, some of which were conducted simultaneously, others sequentially. My multi-method approach was designed to generate audio, video, interview, text and keystroke logging data (see Chapter 1.4 for an overview of my ‘core’ data). Specifically, I try to compare findings from one data set with findings from another data set in attempt to offer a multidimensional account of natural histories of desktop news production. This is an inductive strategy motivated by two contrastive field observations: on the one hand, the messiness, chaos and contingency inherent in social action, and on the other, the identification of standardized newsroom practices, routines and values. An awareness of the limitations of discourse representation

The entextualization of social action into fieldnotes, transcriptions, vignettes and other forms of ‘data’ is a hallmark feature of qualitative social science (see esp. Bucholtz 2007b; Slembrouck 2007; Vigouroux 2007). The role of the ethnographer/analyst in this process is crucial:

The ethnographer ‘inscribes’ social discourse; he writes it down. In so doing, he turns it from a passing event, which exists only in its own moment of occurrence, into an account, which exists in its description and can be reconsulted.

Geertz (1973: 19) cited in Emerson, Fretz, Shaw (1995: 8)

It should come as no surprise then that, for instance, fieldnotes or transcripts of spoken (and written) text are seen as institutionalized abstractions, decontextualized approximations of the strips of reality that these artifacts attempt to represent. What is more, transcripts have social histories of their own and tend to display significant variation in terms of depth and breadth[2]. These limitations point to the irony of using “artefactualized simulacra of lived language in order to be able to sudy it” (Blommaert 2007: 828). In the analyses that follow, I illustrate that transcription practice can be enhanced by electronic, automated transcripts and other forms of visualization generated by keystroke loggers such as Inputlog (Leijten & Van Waes 2006). A reflexive openness towards secondary analysis

During data analysis, I have enjoyed the privilege of sharing my data with a number of colleagues. While I am aware of the traditional reluctance towards secondary analysis[3], I feel that sharing writing process and interview data, as well as audio data of newsroom talk has enhanced my analyses in two ways. First, it has forced me to reflect on my own position in the field and on my knowledge trajectory and second, it has sharpened my own analytical gaze. It was in fact by talking my research design and experience through – thus adding another layer of entextualization onto my own research experience – that I gradually came to grips with my data and how I wanted to mine them. Writing up research papers with colleagues and, on one occasion, an informant, furthered this maturation process.

[1] Let it be clear that I do not argue that media discourse analysis has examined news texts as static artifacts. On the contrary, studies of news discourse structure, function and effect (as in the work of Allan Bell) and those that gauge its ideological impact (to name only three: Fairclough, van Dijk, Fowler) both assume “an emergent, dynamic mechanism that results in the unique display of media discourse over time, culture, and context” (2006). For the sake of clarity then, my central argument posits that linguistic analysis of the contexts in which news discourse is produced (and consumed) remains underexplored .

[2] In a discussion article, Mary Bucholtz (2007b) distinguishes between four types of variation in transcription: variation in the global representation of talk, varitation in notation and format, orthographic variation and variation in translation. The author attributes these levels of variability to analytical foci and theoretical investments of transcribers (Bucholtz 2007a).

[3] It is commonly held that ethnographic data cannot easily be shared because of the intense personal commitment involved in fieldwork research. Jan Blommaert, for instance, has written about how he refused to collaborate with a researcher who had not been exposed to the field because “our colleague could not understand the full depth of what was going on in the transcripts” (Blommaert 2004: 94).




Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: