Tracing Communicative Patterns: A comparative ethnography across platforms, media and contexts

In July 2017, we set out to conduct qualitative fieldwork in China, the US and Denmark following a shared approach that allowed for comparative studies of communication practices across culturally distinct contexts. Here, we present that particular methodological set-up, which facilitates the mapping of complicated, ubiquitously embedded and mundane communication as it occurs in various everyday contexts. In other words, the article focuses on the small, even deep data that enable the analysis of a person’s communication traces and the motives and incentives that trigger them. We pay particular attention to how communicative acts involving the internet interact and intersect with the total range of communicative practices of any individual. Applying an understanding of internet uses as communicative instances allows us to study connections and interdependencies between the myriad ways in which people can communicate and the ways in which people understand, differentiate, distinguish and choose, more or less consciously, between those possibilities. The method provides researchers with a way to place digital trace data in the context of a person’s total communicative repertoire by making it possible to discern how and why people move between digital and analogue communication. As such, the article offers a unique methodological design for comprehensively studying people’s everyday internet use by capturing all communicative acts moving across different platforms, media and contexts.

Researchers focusing on digital communication often define their research area by the level of online or offline integration (Garcia et al., 2009) or delimit their scope by focusing on specific platforms for (online) communication (such as social media: Miller et al., 2016), specific media (such as smartphones: Thorhauge & Lomborg, 2016) or specific contexts (such as the home: Bakardjieva, 2005). Although these limitation strategies define important areas for in-depth research, they also rule out the comprehensive study of connections and interdependencies between different platforms, media and contexts. Researchers studying social media cannot integrate this with knowledge about how and why people will move a conversation from a social media platform to a phone call. Researchers studying smartphones cannot integrate this with knowledge about how and why people use the same calendar software on their laptops as on their smartphones but in an entirely different manner. Researchers studying specific contexts like the home cannot integrate this with knowledge about how and why people use their tablets differently when they commute as opposed to when they are at home.

Taken together, the different strategies for limiting research scopes leave a gap in communication studies when it comes to analysing digital trace data in cross-platform (software), cross-media (digital/analogue hardware) and cross-context (physical/social situation) communication. To address this gap, we propose a qualitative methodological framework that, through the capture of small and deep data, enables mapping and analyses of the communicative practices that make up a person’s daily communication. Rather than looking at a single platform, digital media in particular, or online and offline contexts in isolation, we “follow the people” (Marcus, 1995) and focus on the proliferation of ways in which people can communicate. Essentially, by focusing on people, we are able to follow in their footsteps. This allows us not only to look at their digital communicative traces but also to understand why they were left there in the first place and to analyse them as they relate to other forms of communication in which a person engages. As such, digital trace data are not studied in and of themselves but are contextualized and understood as part of a communicative environment in which people switch seamlessly between online and offline communication practices. In effect, the research design addresses the interwoven nature of people’s online and offline presence (Simon & Ess, 2015) and targets intermediality (Jensen, 2008) on the basis of the individual respondents and in close relation to their actual communication purposes, networks and situations, departing from the question: how do we study the situated enactment of communicative practices empirically within a comparative framework?

The participants for the study were recruited via a combination of network sampling (Heckathorn & Cameron, 2017) and maximum variation sampling (Bryman, 2016). The approach combined recurrent interviews with 24-hour communication diaries in an iterative interview–diary–interview design. Interviews and diaries intersected by anchoring findings from one in the other. The diaries were time and context bound rather than reliant on memory, while the subsequent interviews contextualized the diary data points and allowed the respondents to infer meaning from, make sense of and co-interpret their communicative practices as they figured in their individual diaries. Unlike many other cross-national comparative studies that have focused on exploring differences, our research departed from the notion that, despite large cultural, political and economic differences between China, the US and Denmark (and indeed within them), communicative actions are driven by the same fundamental social mechanisms – such as a desire to maintain strong social ties (Granovetter, 1977). The shared methodological framework provided small and deep data that are comparable across contexts precisely because it focused on those fundamental incentives in a way that produced similar types of data. The comparisons, then, facilitated an analysis of the interactions between different cultural, political and economic contexts and the way in which those fundamental social mechanisms took shape.

The article is structured in three parts: first, a review of previous studies on the internet in everyday life; second, a discussion of different sampling strategies and qualitative data collection methods and an argument for the particular combination of network and maximum variation sampling with communication diaries and elicitation interviews; and, third, examples from our fieldwork that explain how the methodological design produced data that can comprehensively account for people’s individual understandings of their communication practices across platforms, media and contexts.

The “ethnographic turn” in media studies heralded a shift in focus from the specific content of media to the everyday contexts in which media are used. The social uses of television (Lull, 1980) and Reading the romance (Radway, 1984) were among the studies that introduced this focus on contexts of use, which later re-emerged in studies of internet use in everyday contexts (Bakardjieva, 2005). In the first generation of internet studies, however, a specific understanding of the internet as a placeless cyberspace (Slater, 2002) resulted in studies targeting online contexts such as MMOs (Turkle, 1995), MUDs (Markham, 1998) and discussion forums (Baym, 2000) rather than the empirical contexts of use.

By the end of the millennium, scholars had begun challenging the notion of the internet as a cyberspace that could be examined and understood as a context in and of itself. Through extensive fieldwork in Trinidad, Miller and Slater (2000) demonstrated that the internet rather consisted of numerous technologies, used by different people in quite different real-world contexts, which were both shaped by and shaping the internet. Trinidadians were using the internet in one of three contexts: through a stationary computer at home, at work or in an internet café. With the development of mobile media, uses of the internet have diffused into a myriad of other daily life contexts. With uses of the internet entering into nearly all aspects of everyday life, communication researchers are, as we outlined earlier, resorting to various strategies to delimit the scope of their research objects. These strategies can be divided into four categories, as summarized below.

Figure 1

First, scholars have delimited their studies of the internet by focusing on specific contexts. In one strand of research, the domestication tradition, scholars have been increasingly attentive to the home as a particular empirical context wherein the social norms and physical surroundings of specific households govern the appropriation of different media, including the internet (Silverstone, 2006; Silverstone et al., 1994). In line with this tradition, Bakardjieva (2005) challenged the dichotomy between virtual and real lives. Looking at people’s interactions in online communities, she developed a typology of five modalities of virtual togetherness, all of which are closely intertwined with people’s offline realities.

Second, the overall research scope can be delimited by focusing on a specific medium, like mobile and smartphone use in everyday life (Ling & Lai, 2016; Thorhauge & Lomborg, 2016) or laptop computers (Anderson et al., 2009). Third, other researchers have focused on specific platforms of communication that originated online, like MySpace (boyd, 2006), SecondLife, World of Warcraft (Poels et al., 2015) and Twitter or Facebook (Marwick & boyd, 2014; Miller et al., 2016).

Fourth, some scholars have taken individuals as the unit of analysis and followed their communication across platforms, media and contexts. Striving to obtain a holistic account of individuals’ communication, uses of the internet are understood only as they relate to everyday life and to the entire assemblage of communicative practices that amounts to the individuals’ communicative pattern. Applying this limitation strategy, Hasebrink and Domeyer (2012) mapped individuals’ media repertoires, Beneito-Montagut’s (2011) user-centred ethnography focused on interpersonal communication and Helles (2011) connected lifestyle types with different configurations of media and internet use.

It is within this fourth strand of research strategy that we place our study. In effect, we examine communicative patterns, with or without the internet, through mundane actions across any platform, medium and everyday context. This requires an intense and comprehensive relation to the individual participants, which in turn constrains the number of research subjects. Hence, we learn a lot (deeper data) about a few (smaller samples).

To sum up, this study is placed in the tradition of everyday-life internet studies and strives to access people’s communicative practices in a comprehensive manner while accounting for the increased complexity, mobility and fluidity in those practices. In the following part, we discuss the sampling strategies and qualitative data collection methods as they relate to the research design proposed in this article.

Although the function of any sampling strategy is to lend validity to the study at hand (Morse & Niehaus, 2016), researchers have often neglected to argue thoroughly, if at all, in publications for their choice of sampling strategy. Rather, sampling has been treated shallowly as a technical research state, without explaining the ways in which the qualities of the sampling strategy coincide with the aims of the study.

With the purpose of gaining in-depth knowledge of communicative practices and their relationships with culture, regulation, market and technological infrastructure across comparative research contexts, we rest our argument on Marcus’s (1995) study, which noted the pitfalls in attempting to explain macro-structures through a single-case micro-perspective. To avoid this, Marcus called for an ethnography that is multi-sited so that each site is differentiated enough from the others to be able to give a unique perspective on the same macro-structures (be they for instance cultural, political or economic). In Marcus’s framework, multi-sitedness is the result of following things, metaphors, stories, biographies, conflicts or, as is the case here, people across diverse sites. We propose to achieve this by sampling respondents through a combination of network sampling with maximum-variation sampling. This section outlines the values in the combination of these specific strategies.

Maximum variation sampling

One way of constructing the spatial components of a networked field site is by applying a maximum variation principle to the sampling of the network(s). This method helps to locate cases that are heterogeneous regarding selected characteristics that are thought to have important influences on the object of study (Bryman, 2016; Morse & Niehaus, 2016). As such, the term maximum variation refers to the variation between cases rather than to the variation of the full population of the field site(s). While being aware of the limitations that a small n-study puts on the ability to achieve maximum variation in a high number of variables, we sought to achieve the maximum achievable variance in seven variables when sampling our participants, namely 1) age, 2) gender, 3) place of residence and household, 4) ethnicity, 5) educational level, 6) religious beliefs and 7) relationship and parental status, to obtain the broadest possible representation of those factors. Applying maximum variation as a sampling strategy in a qualitative study can, firstly, provide deep descriptions of the individual cases and show the diversity between them and, secondly, highlight patterns that traverse variations and prove to be important precisely because they are unaffected by diversity (Patton, 2015). Figure 2 illustrates the different networks in our study.

Figure 2

Diary studies have been an established approach in social science research since the 1920s, with time use diaries in the 1960s and onwards being the most widely recognized form (Couldry et al., 2007). Critics have identified many limitations to time use diaries, particularly with regard to low response rates and participants’ ability to estimate reliably the time spent on mundane activities. However, the advantages of the method are evident in that diary data are collected close to the time of the registered event (ibid.) and that self-produced diaries generate data on the everyday context of mundane practices that would be difficult to obtain otherwise (Zimmerman & Wieder, 1977). Incorporating diaries into our design supports the research aim by providing a broad representation of participants’ communicative practices while minimizing our interference with their daily lives.

Research diaries can be devised on a continuum ranging from fully structured to entirely unstructured, which is also a distinction between “objective” and “subjective” approaches. Whereas time use diaries belong to the more structured and “objective” approach, “subjective” unstructured diary studies emphasize not confining the reflection of the participants and making the researcher as unobtrusive as possible in the data collection process. For instance, in a highly structured diary study, Hargittai and Karr prompted their participants recurrently at random times, asking them: “Please reply with your location, current activity, people you are with (number, your relationship to them, gender) and any media you are using now” (2009: 20). In contrast, Markham and Couldry, using unstructured weekly diary entries, explained that “The priority was to access people’s processes of reflecting on media use rather than cataloguing consumption habits” (2007: 688).

Couldry and colleagues (2007) described how research tends to favour either structured or unstructured approaches, with unstructured diaries being the most common component of ethnographic studies. The diary design that we argue for here, however, is innovative as it combines the two approaches into one: an objective approach with subjective measures. That is, while our diary method is quite structured in what to report and when, it is fully unstructured when it comes to how to report. Through this coupling, the type of data collected was manifold, although it reported on media use instances each time, and the diary was more organically incorporated into the existing communication practices of the participants. At the end of the introductory interview, the participants were given oral and written instructions on what to report in the communication diary, allowing questions to be answered in person. We asked them to register and report answers to the following questions whenever they used media (analogue or digital), throughout the day and temporally as close as possible to the actual uses, for one or two full days.

When was it and for how long (if this is not implied)?

These rather structured instructions enabled the data collection to stay close to the actual instance of media use as well as the potential trace data left in its wake, giving greater accuracy to the data reported. Therefore, we were able to infer communicative purposes of the media uses as they were happening and to ask clarifying questions then and there. They also allowed for the small, mundane acts of communication – such as the many times a person notices messages on the lock screen of a smartphone or the ritual text messages containing nothing but phatic conversations – to enter into the foreground. To provide further guidance regarding what to report, a definition of what we meant by media, which had been open to interpretation during the introductory interview, was given to the participants, so they all shared a similar idea of what media use entailed when reporting back to us.

For each report instance, the participants were then free to decide on the report platform and medium (e.g., social media, email, phone calls, tablets and paper records) and report type (e.g. texts, visuals, as well as audio and video messages). Additionally, they were introduced to automated data collection apps and software for different digital devices, and they could opt in and out of these. Hereby, the diary reporting was anchored in their existing media habits and contingent on their regular media uses, which minimized the effect of the reporting on the routines of the participants (Carter & Mankoff, 2005). Hence, the specific amalgamation of structured registration with unstructured reporting allowed the researcher to stay close but also out of the way during the data collection. Moreover, the ways in which the participants reported to us in terms of report types, platforms and media provided valuable insights into their general communicative choices and patterns.

We recognize, however, that the diary method is not able to catch every instance of communication, which was also not the purpose. The data collection depends on the participants’ ability to remember to register and report whenever they use a medium. Even though the diary was structured to enable the reporting of trivial and habitual media use, it was inevitable that some uses, to some of the participants, were so engrained in habit that they did not register the acts themselves. In other cases, the participants became too self-conscious and in effect changed their media habits. We argue, however, that the design of the diary registration and reporting process minimized our intrusion into participants’ routines while maximizing the participant feedback. In addition, the combination of diaries and interviews allowed us to ask about media uses that had passed unregistered as well as about face-to-face communications throughout the diary day. Hence, we employ the concept of communication diaries, since the specific fusion of diaries and interviews captures both mediated and non-mediated communication. The diaries, then, represent a different type of trace data from that found in mining uncontextualized and non-consented Facebook comments, retweets or YouTube likes. The situated enactment of the participants’ communicative practices amounts to different traces that they themselves decided to pass on to us (or not). As such, rather than being left out of the research process, the participants owned, reflected on and co-interpreted the traces of their communicative actions, thereby adding information on contexts of use, motives and purposes. In the following, we discuss the advantages and challenges of interchangeably augmenting diary data with interviews.

In the interview–diary–interview structure of our study, the purpose of the introductory interview was to become acquainted with the participants, their everyday life and their media use in general, whereas the second interview was an elicitation interview taking prompts from both the diary and the introductory interview.

The methodological combination of diaries and interviews in social science can be traced back to the 1970s, when Zimmerman and Wieder (1977) proposed a diary–interview method to document participants’ activities in a more feasible manner than through participation observation. Similarly, in recent media repertoire research, Hasebrink and Domeyer (2012) explored the diversity and differences between participants’ media repertoires by documenting their media use through diaries and in follow-up interviews, generating detailed and descriptive data of participants’ media use as well as facilitating further interpretations of those uses. In other words, the diary–interview approach enables researchers to employ diary data in two ways: first, as data themselves and, second, as prompts for the elicitation interview questions. The diaries in our study – the ways in which participants chose to report, the platforms and media that they used, their frequency of reporting, the people with whom they communicated and the ways in which they reached out to us – provide important data on the communicative practices of those respondents. The follow-up interviews allowed for participant reflections and a line of questioning that can be rooted in the actual communications of each participant.

In the introductory interview, we became acquainted with the participants (asking them about their daily routines, their close network, the things that are important to them and the media that they use). We asked for “tours” of their smartphones, tablets and computers, posing questions concerning notifications and sound settings, particular apps and the general organizing principles of their devices (app locations, folders and pinned apps).

Elicitation interviews have commonly been described as interviews in which another data source is used to elicit or spark a conversation within a mixed-method framework (Gundelach et al., 2014). Examples of different constellations of interviews and elicitation data include observational studies (Beneito-Montagut, 2011), screenshots and GPS data points (Thorhauge & Lomborg, 2016), open-ended reflexive diaries (Markham & Couldry, 2007) and log data (Anderson et al., 2009).

Although we had a preliminary interview guide for the elicitation interview, it was comprehensively rewritten to combine discussion topics with concrete prompts from the introductory interview as well as the diary. The tailored line of questioning made it easier for the participants to relate to each topic and encouraged them to fill in the gaps in the diaries and to co-interpret their data. As such, the diary data allowed us to ask more abstract questions, as the questions were tied to the participants’ own concrete acts of communication. It also brought attention to otherwise ubiquitous and indiscernible communication instances, allowing the participants to elaborate on the circumstances and contexts of those specific instances. Finally, by asking questions concerning the blank spaces and periods of non-media use in the diaries, we targeted the communication that was not present in the diary entries, namely face-to-face interaction. As part of a comprehensive study of the role of the internet in everyday communication, face-to-face interactions play a significant role in the multitude of ways in which people can communicate.

Having outlined the sampling strategies and empirical methods, the last part of the article elaborates on our design by providing data examples that show how the combination of communication diaries and interviews enabled us to follow people and analyse their communicative patterns.

We have argued that, unlike research that delimits its scope to specific platforms, media or contexts, our research design enables us to follow communication flows wherever they go: when participants suddenly switch between the affordances (Hutchby, 2001) of different platforms and media, when they communicate in several contexts at once or when they find themselves in private contexts in which researchers cannot usually follow them. We now present data examples from three participants, who are maximally different from each other, as they move across platforms, media and contexts. The examples show how we empirically approached the situated enactment of the participants’ communication patterns insofar as they could be traced in the diaries as well as in the subsequent conversations with them. Ena, in Denmark, a 46-year-old first-generation immigrant from Eastern Europe, works as a part-time nursery teacher and lives with her husband and children; John, from the US, is a 56-year-old painter of Catholic icons who lives alone in Chicago and works part time as a church janitor; and Chun, from China, is a 19-year-old college student who lives in Beijing and majors in Japanese studies.

This study was designed to enable comparisons of communicative patterns in each field and across national contexts. Within each field, the coupling of network and maximum variation sampling produced a number of inherently different networks of inherently different people whose communicative patterns can be comparatively analysed relative to their everyday lives. By applying the same combination of sampling strategies across the three fields, the diverse individual communicative patterns can also be comparatively analysed in relation to the cultural traditions, regulatory regimes, market structures and technological infrastructure pertaining to each field.

The article presented an answer to the methodological question of how to study empirically the situated enactment of communicative practices within a comparative framework. In this context, the question of how not to delimit the research area to specific platforms, media or contexts was also addressed with the introduction of the novel approach of following participants’ cross-platform, cross-media and cross-context practices through their self-reports as well as their subsequent elaborations on these data traces. Any methodological selection is simultaneously a deselection, to which the above-mentioned limitations are also a testament. We do not argue that this approach is a way to capture everything; rather, it is an attempt to study communication comprehensively through the collection of small and deep data. The research design makes it possible to capture the intermediality of the basis of an individual’s unique linking of all the communication tools and channels available to him or her and, in turn, a way to identify the role of digital communication tools as they interact and intersect with other forms of onlife communication. As such, the methods that we presented address how to study the internet in its increasingly ubiquitous presence in everyday life by mapping communication – online, offline or both at once – as it happens and wherever it happens. The approach is a novel iteration of mapping and analysing digital media use and digital trace data as part of a holistic view of individual communicative practices. It has been suggested time and again that online and offline experiences cannot be explained separately. We take this argument one step further and suggest that future research into digital communication needs to take the interwovenness seriously, not just between online and offline contexts but between communicative practices in general, as they are performed across platforms, media and contexts.