Media convergence and publicness: Towards a modular and iterative approach to online research ethics

1 Introduction

The aim of this article is to build a bridge between assumptions about publicness and research ethics in traditional (mass) media research (e. g. newspapers, TV and radio) and similar issues pertaining to research in so-called ‘new media’ environments. The reason for doing so is two-fold: first, discussions about online ethics are inevitably informed by perceptions of publicness previously articulated in relation to research on mass media; and, second, convergence of multiple modes and media in everyday communication makes ‘old-new’ dichotomies particularly difficult to keep separate. By moving beyond the ‘old-new’ dichotomy in relation to research ethics in media investigation, the article complements previous work that offers historicised understandings of textual and mediated discourse practices (e. g. Herring 2004: 3, Jones 2011: 336), genres (e. g. Heyd 2016: 89–90), and media ideologies (e. g. Thurlow 2013: 243; Spilioti 2016: 134). More specifically, understanding how publicness has been defined and applied to the process of researching mass media discourse paves the way for contextualising and critically revisiting ethical debates about collecting and analysing internet data, on the one hand, and presenting and publishing data extracts from new media contexts, on the other.

Media convergence, as a key property of new media environments (cf. ‘convergent media computer-mediated communication’ Herring 2013: 4), challenges assumptions about publicness. In particular, the study of websites, which increasingly include multimodal, multiauthored and multimedial content, questions any assumed clear-cut boundaries between public and personal data. With reference to language-focused research on multilingual digital writing in such contexts, I critically engage with the challenges to the process of ethical decision-making and the potential risks for internet users of a priori assumed public-private dichotomies. Drawing on modularity as a key organising principle of web design and discourse (Androutsopoulos 2010: 208; Pauwels 2012: 251), the article proposes a modular and iterative approach to research ethics that takes into account the complex and fluid configuration of web environments and attends to the conditions of multiple authorship and multiple publics that are increasingly typical of such contexts.

2 Mass media and the ‘public domain’: Research ethics

Any empirical research endeavour involves information management, typically referred to as data collection, analysis and presentation in academic discourse. In the discipline of applied linguistics, in particular, such information can be intimately linked with the persons who provide, produce and consume texts the researcher is interested in. It is this relationship between the researcher and the persons researched that is primarily facilitated and regulated by the key principles of research ethics, such as respect to human dignity, autonomy, and safety, as well as protection from harm (BAAL 2016; AoIR 2012; ESRC 2015). Acknowledging all stages of the research process, from initial design to dissemination (AoIR 2012: 3), we realise that ethical considerations do not only target the process of data collection and analysis (i. e. informed consent to access, record, store and analyse language/texts) but they also concern the process of data presentation and dissemination (i. e. consent or permission to quote and present language/texts).

With respect to the above ethical considerations, media discourse, available in various forms of published press, radio, television and film, has attracted the attention of language and communication scholars, as mass media are deemed to offer ample and convenient access to language data (see Yates 1996: 30, for a similar argument about new media). With media discourse being produced for purposes other than language research and arguably readily accessible, research using such sources seems to kill two birds with one stone: eliminating any influence on the data caused by the presence of the researcher–what Labov (1972) called the ‘observer’s paradox’–and quickly passing through ethical scrutiny from institutional review panels.

Understanding mass media research as low–or no–risk of harming the persons researched draws on a key premise that distinguishes personal data from material already available in the public domain. To quote the UK’s Economic and Social Research Council framework for research ethics,

While data collected and stored as a record at an individual level are considered personal data, material already in the public domain are not. For example, published biographies, newspaper accounts of an individual’s activities and published minutes of a meeting would not be considered ‘personal data’ requiring ethics review, nor would interviews broadcast on radio or television or online, nor diaries or letters in the public domain.

(ESRC 2015: 12)

Unpacking the notion of ‘already available in the public domain’ as used for press and broadcasting material, it appears that publicness, here, is understood in terms of three key assumptions. First, publicness is associated with a domain, i. e. a space that makes such material available and, thus, accessible to the wider public, including researchers (see also Giaxoglou 2016). Second, publicness is conceived in terms of the material’s purpose and targeted audience: material is produced to be consumed by the public, understood in the context of broadcasting as a mass and unidentified audience (cf. the ‘broadcast audience’ Marwick and Boyd 2011: 128–129). Third, publicness is associated with the roles and relevant identities of the authors of such media texts: for example, media professionals, politicians, and sports figures are usually invoked in their professional or public roles in mass media contexts.

However, there are cases where the three aforementioned assumptions may not justify media material as public. One is programmes that are referred to as ‘public or audience participation programmes’ (Livingstone and Lunt 1994; Thornborrow 2015). Radio phone-ins and reality TV shows (like Big Brother or the Jeremy Kyle Show) involve ordinary people (i. e. non-media or non-expert professionals) who often disclose personal and sensitive information about themselves. In such cases, researchers arguably deal with what is, essentially, mediated personal data, if the latter is defined as ‘information relating to an identifiable living individual’ (ESRC 2015: 23). Thus far, though, such material is also considered low risk, given that ‘the information contained in the personal data has been made public as a result of steps deliberately taken by the data subject’ (UK Data Protection Act 1998). It can be argued that people who willingly come forward, agree to be auditioned, enter into contractual agreements with media organisations and put themselves in front of open microphones and TV cameras have made decisions that involve assessing, as well as confronting, the publicness of such media interactions. Although one can never be sure about the intentions of such individuals or their understanding of the potential (harmful) impact of public sharing of their personal life on air, researchers into media discourse have long considered such interactions as public. Here, publicness also invokes assumptions about the speakers’ or authors’ agency in making such material accessible in the public space and to be consumed by the public.

Against this backdrop of current authoritative ethical guidelines, the media discourse researcher does not need to seek informed consent for collecting and analysing such material. Her ethical stance involves primarily issues related to the process of data presentation and dissemination, i. e. copyright and respect to intellectual property. In the case of press, TV and radio, ownership of published material lies primarily with the media organisation. As a result, appropriate attribution of material quoted or analysed, as well as respect for restrictions regarding the legal reproduction of media content, constitute the main ethical concerns of the media discourse analyst. To what extent, though, does increasing media convergence challenge current perceptions of publicness and, thus, call for revisiting ethical decision-making in language-focused media research? After defining media convergence, the following section questions assumptions of publicness as material that are (i) accessible in a domain, (ii) produced for a mass public, (iii) authored by public or professional figures and/or (iv) the result of deliberate agentive decisions of individuals.

3 Media convergence: Questioning ‘old-new’ dichotomies and ‘publicness’

Media convergence is a key phenomenon in the contemporary mediascape. From a technological perspective, convergence refers to the merging of previously distinct technologies within a single device or media platform. As a sociocultural phenomenon, Jenkins (2006: 2) defines convergence as ‘the flow of content across multiple media platforms, the cooperation between multiple media industries, and the migratory behaviour of media audiences who would go almost everywhere in search of the kinds of entertainment experiences they wanted’. In other words, convergence resists dichotomies between ‘old’ (e. g. press and broadcasting) and ‘new’ (e. g. email, online forums, Twitter, Facebook, etc) media, by foregrounding the increasing integration of multiple media platforms and the mobility of content across media environments. In the context of broadcasting media, for example, there are three key ways in which intersections between ‘old’ and ‘new’ media manifest. Firstly, public participation from TV and radio audiences is no longer limited to telephone calls or to the odd appearances of lay people in talk shows; instead, broadcast programmes increasingly rely on digital technologies for interaction with their audience and often embed emails, tweets, and other digital texts in the flow of the TV or radio programme (Thornborrow 2015). Secondly, the public engages with ‘old’ media snippets across digital platforms through commenting on, copying-and-pasting, liking, sharing, editing and remixing TV or radio material (e. g. Page 2012; Georgakopoulou 2014). Thirdly, and primarily discussed in this article, media industries work together to augment the inter-medial presence of news corporations, TV and radio programmes through dedicated websites, Twitter and Facebook profiles, used both for (re)broadcasting content and interacting with different audiences.

The challenges media convergence raises for researchers studying language use online have been repeatedly pointed out in the literature (e. g. Androutsopoulos 2011: 281; Herring 2013: 5; Bolander and Locher 2014: 16; Georgakopoulou and Spilioti 2016: 4). Among the key methodological and theoretical issues noted are: (i) the need for a methodological framework appropriate for analysis of ‘media coactivity’ (Herring 2013) or multitasking across and within media; (ii) attending to multimodality, as the range of available semiotic resources and their potential combinations have dramatically increased; and (iii) orienting to multiauthorship processes, as multiple participants are involved in the production of media content and any boundaries between media producers and consumers become increasingly blurred (cf. ‘prosumer’ Ritzer and Jurgenson 2010).

From a research ethics perspective, media convergence–together with the increasingly multimedial, multimodal, and multiauthored environments it affords–resists assumptions about media content as something available in a domain, i. e. given, static, located in a particular space, and, thus, accessible or amenable for collection and recording. Instead, digital content is mobile and moves–or, ‘flows’ in Jenkins’s (2006) words–across multiple sites for media engagement and activity. How can one discern whether this circulating and mobile content has been produced for a mass and unidentified public–if such a public exists in internet contexts? In a similar vein, how can one identify certain individuals as authors or discern the relevance of their professional or public roles to the communicative context, especially in circumstances of multiauthorship production where boundaries between content producers and consumers are no longer fixed or static? Last, but not least, how easy or clear is it for the ethical researcher to assess individual or deliberate agency in making personal content public, especially in environments where, as we will see, social media plug-ins may create conditions of automatic disclosure? These questions unsettle the key assumptions that arguably underpin established distinctions between personal data and material available in the public domain, complicating ethical decision-making, on the one hand, and questioning the scope of applying this distinction to digital contexts, on the other.

With reference to research on multilingual writing in digital environments, this article revisits the above questions in the context of websites that bring together content produced by multiple authors, for multiple publics and through combination of multiple modes and media. As will be shown, assumptions about the publicness of perceived public facing websites can be nuanced through attention to the ways in which individual areas, or modules, of a website are organised, combined, and discursively constructed by–and for–different social actors.

4 Websites as sources for research on multilingual writing

Unlike social network sites or personal websites as sources for language-focused research, official websites run by institutions and professional organizations arguably represent a low risk digital environment from the point of view of research ethics. With respect to criteria for publicness, material embedded in such websites can be considered to be ‘already in the public domain’ (ESRC 2015: 12), as the web domain is created and regulated by relevant professionals (i. e. web designers and organization represented), with content produced for public or mass consumption. Nevertheless, demonstrating shortcomings of the public-personal data distinction even in these less controversial areas can be revealing of the need to radically rethink and move beyond received dichotomies.

As mentioned in the previous section, a key factor in challenging current ideas about publicness and ethics is media convergence. In the context of broadcast organizations, convergence between the so-called ‘old and new’ media is manifest, among others, in websites dedicated to popular TV and radio programmes. Such websites typically have the following purposes: to provide the programme’s audience with further opportunities to (re)consume broadcast material, to enable them to engage with a wealth of other material relevant to the show’s topic (e. g. alternative news sources) and to interact with media professionals (e. g. show presenters) and other audience members. These purposes usually shape – and are fulfilled in – two distinct communicative spaces: (i) an edited web space that features content designed and/or selected for (re)broadcasting to TV/radio and web audiences (cf. Section 5.1) and (ii) an interactive space that includes messages posted by visitors to the website who often identify themselves as audience members of the particular programme (cf. Section 5.2). Content in both areas is highly dynamic and volatile, with material and messages updated (i. e. added or deleted) and reconfigured on a regular basis.

From the perspective of language-focused research, the conditions of mobility, multiauthorship, multimodality, and multimediality afforded in such environments often result in the mobilization and, at times, strategic and reflexive configuration of multiple linguistic resources, such as styles, registers, codes, etc (cf. Deumert 2014; Tagg 2015). For that reason, the developing field of the sociolinguistics of writing acknowledges the central role of research on digitally-mediated communication in reconceptualising writing (Lillis and McKinney 2013: 421–424) and pushing written multilingual discourse from the periphery to the core of sociolinguistic research on multilingualism (Sebba 2013: 98–99).

5 Opening Pandora’s box of ‘convergence’: Revisiting publicness in the research process

What counts as data in a multi/bilingual website, together with the ethical issues raised by such methodological decisions, is a reflexive process shaped and driven by one’s research question and wider epistemology. Recent research on multilingualism and writing (Sebba 2013; Lee 2016: 121; Tagg and Seargeant 2016: 341) has stressed the multimodal aspect of such phenomena, driven by a shift in sociolinguistic research towards writing as an (ideological) semiotic process (Lillis and McKinney 2013; Maybin 2013: 554). This multimodal approach involves taking into account the ‘visual contexts’ of multilingual discourse, ‘as a text surrounded by other texts, potentially with differing font sizes, colours and styles all potentially providing context for interpreting the content of the [multilingual] text’ (Sebba 2013: 102). This theoretical shift in the approach of the written word as a (multi)semiotic object has methodological consequences: from the decontextualized written word as unit for analysis, the focus shifts to the study of writing in multimodal ensembles, i. e. multimodal representations or interactions that are ‘seen as a material outcome or trace of the social context, available modes and modal affordances, the technology available and the agency of an individual’ (Jewitt 2013: 254–255). Following this approach, attention to users’ – and, at the same time, web designers’ – strategic use of visual means to organise content on a webpage helps the researcher to delineate multimodal analytical units within which multilingual writing can be explored in relation to the surrounding linguistic, graphemic, and visual resources. Revisiting ethics iteratively within and across such multimodal ensembles is paramount as the varying social and material constraints at play in each unit impact upon perceived boundaries between public and personal data.

5.1 Ethics and edited web content: Towards a modular approach

Edited content on official websites of media institutions and professional organisations is generally thought to belong to the producers and owners of the web domain. This approach to web domain as an undifferentiated whole is challenged by research on web discourse that foregrounds ‘modularity’ as a key organisational concept (e. g. Pauwels 2012: 251): content of individual webpages is organised into building blocks or modules that have been selected and combined, either advertently or inadvertently, into a coherent whole. As a result, in addition to the aforementioned methodological issues pertaining to the study of writing, the identification of modules that make up a website is useful for processes of ethical decision-making, as each building block may have different conditions of authorship and, thus, perceived ownership of creative content.

In order to identify the relevant modules on the multimodal space of a web page, visual elements can be combined with functional criteria. Visual elements (as suggested by Kress and van Leeuwen 2006: 214–218), particularly lines that frame and separate spatially contiguous texts, as well as colour, font and shape discontinuities that distinguish visually different areas of the web domain, have been particularly useful for identifying areas likely to operate as separate modules on edited web content on the Greekophrenia website. Drawing on visual resources, the identification of relevant building blocks also takes into account the extent to which the visually and spatially distinct modules fulfil various functions. With respect to websites dedicated to TV/radio programmes, such functions orient to (i) information structure or organization (menus, internal links, navigational tools), (ii) (re)broadcasting of media material (external links and other embedded textual, audio and visual material), and (iii) interactivity with site visitors (internal links and prompts to post comments).

The application of these criteria to websites reveals that different modular parts entail varying attention to copyright issues. For example, elements designed by the site developers, such as the ‘Greekophrenia on the net’ logo, site menus and internal links, represent creative content belonging to the site. Proper attribution and permission by the creators and owners of the website to reproduce such content in published material is necessary, similar to practices used for reproducing mass media material. In contrast, decisions about copyright permission become less straightforward when considering other areas that embed and (re)broadcast, for example, articles or news bites from journalists or bloggers who are not directly involved in the site design and who probably produced such texts for different audiences. The multiauthored composition of edited web content, together with the multimedial make up of the web area embedding the show’s radio or TV episodes, present cases where web discourse involves texts produced for different publics and exhibiting different conditions of media production (e. g. broadcast vs. internet). The challenges arising from convergence, here, concern the need to clarify copyright issues related to the reproduction of creative content for presentation and publication purposes. Contacting the site developers and owners and seeking permission for reproduction in academic research is the starting point.

Although it is advisable to contact web developers and owners of individual sites, digital affordances for copying, pasting, embedding and circulating content often obscure ownership and complicate copyright permission. In such cases, researchers can think of alternative ways for (re)presentation of web material. In fact, what seems like an obstacle at first sight can be a useful research tool. Publishing research on any type of discourse data involves a process of purposive presentation of data in order to illustrate and support the arguments made. Rather than approaching potential issues with reproduction of multimodal and multiauthored discourse as mere challenges, the time is ripe for developing appropriate methods for presenting such sources where copyright clearance may not be adequate or logistically possible. Similar to transcribing and presenting (multimodal) spoken discourse, arbitrariness of symbols and selectivity in discourse (re)presentation are inherent properties of this part of the research process. Research on multimodal and multilingual writing can draw on publishing practices of discourse analytic research on artwork, where strict copyright restrictions do not allow the reproduction of such material for publication purposes. Nevertheless, this does not hinder researchers from examining and analysing such material as evident in Jaworski’s (2014) study of multilingualism and heteroglossia in artwork by multimedia and performance artists. In such cases, visual material can be (and are) represented through the use of diagrams (Jaworski 2014: 154). The following section makes use of a diagrammatic representation of web discourse in order to discuss ethical and privacy issues raised by research on the interactive areas of websites dedicated to TV and radio shows (Figure 1).

Figure 1:

Modular representation of comments section (www.ellinofreneianet.gr 2012–2015).

6 Towards a modular and iterative approach to ethics in applied linguistics

Challenges related to media convergence arise from the interplay of multiple modes, authors and media material. Approaching a website as a single domain or space runs counter to the modular make up of such communicative environments which, in turn, resists any blanket ethics statement about the publicness of institutional websites. Instead, unpacking media convergences and any associated ethical challenges foregrounds the need for a more contextualised and process-based approach that addresses ethical issues at different junctures of a research project and opens a dialogue between regulatory-driven ethical guidelines and context-sensitive understandings of ethical dilemmas (AoIR 2012: 4–5).

In the case of dealing with intersections of ‘old’ and ‘new’ media (as in the example of radio/TV programmes’ websites), tracing trajectories of media practices enables the researcher to enrich and contextualise understandings of publicness. It moves beyond equating publicness with accessibility and adds further dimensions to understandings of the public domain, by taking into account discussion topics (cf. Bolander and Locher 2014: 17; Buchanan 2011: 94) and professional roles of media producers (cf. similar distinction between amateur and commercial web environments in Androutsopoulos 2008: 9). At the same time, a process-based approach involves revising ethical decisions during different research stages. The research questions and methodological premises shape the process of ethical decision-making. For instance, methodological decisions about the study of multilingual writing online, ranging from the written word as individual, decontextualized units to writing as ideological (multi)semiotic process, have an impact on how the ethical researcher deals with informed consent and copyright issues (cf. Page et al. 2014: 60).

In terms of consent for using web material for research purposes, a modular perspective to ethical decision-making contributes to the development of a process-based approach to research ethics (AoIR 2012). As shown in the article, websites consist of several, functionally and spatially distinct, areas where conditions of multiauthorship and multimediality are (re)configured in various ways. Different configurations involve varying copyright restrictions and understandings of publicness across and within the pages of a website. Attention to the design, multiple audiences, discourse activities, and interactivity patterns in each individual area is important both for methodological (i. e. digital writing as hybrid, multisemiotic and multivocal) and ethical reasons, as digital content is generated through various means. The article also foregrounds the need for careful ethical consideration of metadata that might get inadvertently collected by the internet researcher. Metadata, such as users’ profile picture and other personal information, are often automatically embedded in websites that use social media plugins. In this case, human agency and publicness as framed in ethical guidelines (i. e. as steps deliberately taken by human subjects) need to be revisited in order to account for technologically mediated agency that involves instances of automatic disclosure of personal information.

Together with a modular perspective, the process of ethical decision-making online is iterative. Understandings of publicness have been revisited across the different stages of the research and across the different modules that make up the main web pages of the site. Concerns about publicness extend to the current moment of writing the article and, thus, the stage of publishing and disseminating research, particularly because visitors’ comments are no longer available on the site. What sort of ethical decisions should one make when certain material no longer exists in the public domain but only in the private record of the researcher? It seems that the process of academic research is also caught up in the ethical conundrums of online scalability and spreadability (Boyd 2011: 48). It is a truism that the speed and scale at which online content is circulated are unprecedented but academic research and publishing have always played their own role in spreading and scaling up (certain) types of discourse. My response to this ethical conundrum was to represent the different areas and scripts in a diagram, rather than including screenshots or verbatim messages from the comments section. Although this would be a questionable practice for publishing results and findings from a language-focused study, such an instrumental approach to research presentation (i. e. moving beyond the publishing of verbatim quotes if justified by the aims of the study) can be useful in digital contexts where public-private discussions are muddying the waters of research ethics.

The process of transforming recorded material in line with ethical commitments is well rooted in research practices of transcribing and analysing spoken discourse (Ochs 1979). Compared to collecting and transcribing speech, written discourse (e. g. written documents, newspaper articles) lends itself conveniently to academic research. In other words, it deceives researchers into thinking that writing is already available for scrutiny, i. e. given material or, to use the Latin term, data (‘given’) ready to be studied and analysed (Markham 2013). The element of deceit lies in the fact that this reasoning collapses together understandings about the materiality of the sources (persistence of writing on a page vs. volatile speech) and epistemological assumptions about texts and people as objects and ‘data subjects’, respectively (UK Data Protection Act 1998). Perceiving writing as a static object produced by data subjects shapes our stance towards research on texts and people (cf. Cameron et al. 1993: 87). Research on digital discourse carries over these assumptions about written discourse data. Perhaps we need to stop thinking about discourse data in terms of materiality and move beyond objectifying texts and people; then we can embrace richness and complexity as inherent properties of doing research (in any setting, face-to-face or mediated). This can take us a step closer to solutions that orient to research as process and shed light on, rather than obscure, the nuances of communicative encounters, online and offline.