Rethinking social media for qualitative research: The use of Facebook Activity Logs and Search History in interview settings

Introduction

Digital devices and data have offered new possibilities for social research and the generation of research data (Back & Puwar, 2012; Burrows & Savage, 2014; Marres, 2017; Marres & Gerlitz, 2016; Savage, 2013). In their ‘Manifesto for Live Methods’, Back and Puwar (2012) encouraged sociologists to question prevailing research conventions and to develop ‘empirical devices and probes that produce affects and reactions that re-invent relations to the social and environmental’ (p. 9). Their manifesto is an excellent place to spark sociologists’ methodological imagination but it does not provide practical strategies to develop tools for ‘live methods’ and respond to the challenges of such methods. Following Back and Puwar’s provocations, Marres and Weltevrede (2013) began to envision what new forms of live research might look like. In their methodological reflection on data scraping, they argued that it is important to distinguish between two forms of real-time research: namely studies dedicated to monitoring live content in a given moment (scraping data) and studies analysing the liveliness of issues over a given period of time (scraping the social) (p. 313). These two types of research have often been conflated and the latter overlooked. While Marres and Weltevrede’s discussion was concerned with online data extraction, the distinction between live and lively is helpful for social research interested in digital platforms and data both as research objects and as part of the research process. Marres and Weltevrede (2013) clearly understood the research value of the lively character of digital data and called for new forms of research that seek to render analytically productive ‘the dynamism or “shape-shifting” of online data, turning these into a resource and an object of digital social research’ (p. 327).

Building on Marres and Weltevrede (2013), there are other forms of liveliness that arise in research using data and digital platforms in interview settings. Indeed, in this context liveliness is not confined to data but expands to the interview situation itself during which the researcher as well as the participants encounter these data. In this way, the dynamism or liveliness of data mediates and affects the data but also the research situation. Liveliness also stems from the specific design and architectures of the digital platforms (and their features) which shape the data generated through them (e.g. data displayed as a chronological stream or data displayed on a newsfeed regulated by algorithms). Added to this, digital platforms allow users to interact with data (e.g. scrolling down and clicking on links) which in turn creates yet more liveliness during the situation of the interview. These two latter forms of liveliness distinguish the use of digital platforms and data from other types of research probes and offer both opportunities and challenges for social enquiry. To take advantage of these possibilities, social researchers must engage specifically with what liveliness means in the context of their research.

Recently, the notion of ‘assemblage’ has been mobilised to better understand the broad challenges of using digital devices and data in social research (see Marres & Gerlitz, 2016; Ruppert, Law, & Savage, 2013). As an analytical device, this notion allows to grasp the co-constitutive relationship between users, digital technologies, the data generated from their interactions with these technologies and the broader political, economic and discursive contexts in which these interactions take place. As researchers we need to attend to the specificities of the digital data assemblages that we generate in our research and to critically reflect on ‘their materialities, productivities and mediating capacities’ (Ruppert et al., 2013, p. 25) and on the ways in which they reconfigure our practice. In the last decade, researchers have considered and engaged with some of the challenges brought up by turning digital devices and data into sources and/or objects of social research (see for example Burrows & Savage, 2014; Marres, 2017; Marres & Gerlitz, 2016; Marres & Weltevrede, 2013; Rogers, 2013). However, the different forms of liveliness produced by the digital assemblage created by using digital platforms and data in qualitative research, and the ways in which using them impact on the research data and situations have been overlooked. Drawing on empirical data, this article addresses this gap by examining the specific forms of liveliness generated by the use of Facebook Activity Logs and Search History in interview settings as well as the broader challenges of using social media features as probes in qualitative research.

Digital technology and data in social research

The pervasiveness of digital devices and data in the everyday has contributed to the re-evaluation of the role of sociological research in the production of knowledge – in particular with regard to Big Data (Back, 2012; Burrows & Savage, 2014; Fuchs, 2017; Kitchin, 2014; Marres, 2017; van Dijck, 2014). In this context, the so-called computational turn (Berry, 2011) and the emergence of Big Data have participated in an increased reliance on quantitative methods and/or predictive data analytics programmes in social research and led to a new empiricist epistemology and new forms of positivism (Fuchs, 2017; Kitchin, 2014; Langlois & Elmer, 2013). As Ruppert et al. (2013) pointed out, digital devices tend to track and record activities and ‘doings’ – what they called ‘the doing subject’ – but not to track ‘the reflexive and self-eliciting subject’ (p. 35). In this way, digital devices are often primarily designed to extract and record data in ways that depart from the epistemological premises of qualitative and critical research traditions which examine not only people’s ‘doings’ but how they interpret, experience and make sense of their social worlds (Edwards & Holland, 2013). Combined these tendencies have overshadowed research and methodological discussions on the uses of social media platforms and data on a ‘small’ scale in qualitative research (Fuchs, 2017; Latzko-Toth, Bonneau, & Millette, 2016). Only a few studies have started to address this gap (see for example Beneito-Montagut, Begueria, & Cassián, 2017; Housley, Dicks, Henwood, & Smith, 2017; Latzko-Toth et al., 2016; Mare, 2017).

A large body of qualitative research has explored the complex ways in which people negotiate and engage with digital technologies and the data generated through them (Beer & Burrows, 2013; Couldry & Hepp, 2016). As part of this stream of research, scholars have examined how the designs, architectures and affordances of digital devices – social media platforms in particular – shape users’ engagement with and understandings of these technologies (Bivens & Haimson, 2016; Bucher & Helmond, 2018; Kaun & Stiernstedt, 2014; Lupton, 2015, 2017). Importantly, studies have also pointed to the specific political economy in which users engage with digital technologies and in particular the role of corporations in promoting notions of a ‘social’ Web and of platforms as ‘neutral’ technical infrastructures facilitating social connections while downplaying the ways in which personal data are turned into commodities and users subject to ubiquitous profiling and targeted advertising (Fuchs, 2014; Gillespie, 2010; van Dijck, 2013). Furthermore, scholars have called attention to the importance of understanding that the digital data generated by users, produced through the platforms (meta-data) and/or inferred by corporations are not static, nor isolated. As Lupton (2016) argues, digital data should be understood as ‘lively’, not only on the premise that they are information about human life, often with consequences for users, but because digital data have ‘a vitality and social life of their own’ (p. 114). In the context of social research, scholars have argued that digital devices and data are co-constitutive of the production of knowledge (see Kitchin, 2014; Marres & Gerlitz 2016; Ruppert et al., 2013). They are ‘both the material of social lives and form part of many of the apparatuses for knowing those lives’ (Ruppert et al. 2013, p. 24, original emphasis). While understandings of digital technologies, data and users’ engagement as mutually constitutive have become commonplace in media and communication studies, it is urgent that all researchers using digital devices and data for social enquiry reflect more critically on what it means for their research, not only in relation to the technological infrastructures and affordances of the digital devices they use but also to the liveliness of the data they generate through them in the context of research. A methodological approach that does not engage concretely with how digital devices and data are part of the production of knowledge runs the risk of fixing researchers’ attention to the opportunities generated by technology for how we do sociology while downplaying the complex ways in which digital devices and data also actively shape the doing of sociology.

This article argues that conceptualising digital data and devices as lively will allow researchers to reflect upon the ways in which social research has become mediated by human–technological encounters and knowledge shaped by and recirculated within the digital devices and data that they use and/or repurpose. Without an open discussion about these processes and the development of procedures to negotiate their impacts on sociological practice, researchers will not be able to take full advantage of the opportunities created by digital devices for social enquiry and more specifically qualitative research.

Using digital probes in interview settings

There is a long tradition of using digital and non-digital probes in qualitative research to elicit discussion and encourage participants to reflect upon their lives and past experiences. In this context, researchers have used a range of artefacts including photographs, videos, graphics, maps and diaries (Bagnoli, 2009; Banks, 2008; Copeland & Agosto, 2012). However, using digital platforms and the data generated through them as probes in interview settings differs in several ways from other forms of research probes. Digital platforms and their specific features (timeline, profile, activity logs, etc.) offer users the possibility to interact with data (e.g. scrolling down and clicking on links) in a responsive manner. This responsiveness is generated by the specific designs and architectures of the digital platforms used in the interview settings as well as by the algorithms of these platforms which are working in ways mostly unknown to the researcher (Kennedy et al, 2015). These differences offer new opportunities but also challenges which are specific to the dynamic and digital character of the probes. It is therefore important to critically engage and reflect upon the ways in which using digital platforms and data as probes in interview settings shape and reconfigure research practice.

To date, only a few qualitative studies have incorporated social media features as probes in interview settings (Duguay, 2014; Latzko-Toth et al., 2016; Robards & Lincoln, 2017). For example, in their work on the sustained use of Facebook among young people, Robards and Lincoln (2017) asked their participants to ‘scroll back’ through their Facebook timelines and to comment on the content they posted on Facebook over time to explore longitudinal life narratives. The timelines themselves in combination with the involvement of participants as ‘co-analysts’ of their own digital traces during interviews allowed Robards and Lincoln to ‘bring to life’ participants’ data as well as to capture the specific contexts in which information was disclosed on Facebook (2017, p. 720). In the same vein, Latzko-Toth et al. (2016) have developed a new type of interview – the trace interview – in which they visited and commented with participants the items that appeared on their Facebook Activity Logs to explore the ways in which young adults used the platform to obtain information and form their opinions about political issues. Their approach involved asking participants to reflect on their own digital traces in order to generate ‘a metadiscourse about them’ (p. 203). Latzko-Toth et al. argue that trace interviews – and the use of Facebook Activity Logs in interview settings – can be used as a strategy to thicken data with elements of context, description and meaning generated by ‘the concomitant dialogue between the interviewer and the participant about the [digital] traces’ (p. 206). However, their approach by conceptualising the digital traces mostly as an object of discussion tends to overlook the challenges brought up by the different forms of liveliness that using digital traces as probes generates in research. In other words, the ‘dialogue’ which takes place in this type of research is not only about the digital traces but with them, which in turn actively shapes the interview.

In order to engage with the challenges that arise when using social media features as probes in qualitative research, this article argues that we need to conceptualise the interviews and the data generated through the features as specific forms of ‘digital data assemblages’. Digital data assemblages are best understood as crystallised ‘configurations of discourses, practices, data, human users, and technologies’ (Lupton, 2015, p. 336); a form of ‘frozen data’. Each assemblage represents a specific moment in time before being rearranged in new ways as ‘users’ encounters with digital technologies occur and as different data sets come together, interact and are taken up for a range of purposes by various actors and agencies’ (p. 336). Following this logic, the use of social media features and data in interview settings produces a specific type of digital data assemblage that is constituted by the continuous and changing interactions between participants, the researcher, digital technologies and digital data as well as deeply connected to the moment and situation of the research interview. This digital data assemblage actively shapes the research data and situation by its liveness and the unique and specific moment in time in which it takes place – but also by its liveliness and the ways in which the assemblage comes together and interacts in that moment. Thus, liveliness is generated by the dynamic character of data collected over time, decontextualised from its original setting and repurposed in a synchronous way during interviews. In other words, liveliness arises from the interactions and tensions between the formatting of data (i.e. social media features), the entries listed on social media features (i.e. history of activities or searches), the meanings that participants and the researcher ascribe to these entries, participants’ recollection of them as well as how these entries reflect back on them in the context of the interview. Liveliness also stems from the uses of social media features which have been designed to select, order and present information in specific ways and as a result take part in shaping the pace and flow of discussions.

Drawing on empirical data, this article examines the different forms of liveliness embedded in the use of Facebook Activity Logs and Search Histories as digital probes in interview settings and the broader implications of using such probes for sociological research and practice.

Methodology

This article draws on data collected during my doctoral research, which explored the meanings that young adults ascribed to social media, their everyday engagement with the platforms and in particular their perceptions of peer monitoring and profile checking practices on these platforms (see Gangneux, 2018, 2019, for a detailed analysis). As part of the research, I conducted 32 semi-structured interviews with young adults aged 20–25 living in Glasgow (N = 19 women and 13 men) and coming from a relatively privileged background. All participants were regularly using social media, albeit the platforms they used, and their levels of engagement varied. All were active on Facebook and a large proportion of them also regularly used Instagram (n = 21) and Twitter (n = 20). To explore participants’ engagement with and understandings of social media platforms, I used two Facebook features as digital probes during interviews. Facebook Activity Logs (FAL) (Figure 1) were set up in the second half of the interviews to elicit discussion about participants’ engagement with Facebook using specific examples from their recent activities. Facebook Search Histories (FSH) (Figure 2) were used much later in the interviews to provoke specific discussion on searching and profile-checking practices. Both probes were devised to generate thick data regarding young people’s understandings of social media platforms and to encourage them to reflect upon their practices.

OPEN IN VIEWER

Figure 1.

Facebook Activity Log.

OPEN IN VIEWER

Figure 2.

Facebook Search History.

FAL were introduced by Facebook in 2011 alongside the Timeline Feature to let users review and manage what they shared on the platform. The FAL (Figure 1) is only accessible to users and displays a stream of user’s activities on the platform that is a dynamic and automated flow of entries presented as a single column and ordered chronologically (Manovich, 2012). Activities listed include content posted, content liked and commented on, profile updates, friend requests, new contacts, videos watched, events joined, etc.

The Search History (FSH) is part of the FAL (Figure 2) and is presented in a similar way. Searches, however, do not appear in the default mode of the FAL; users need to activate the ‘search filter’ to be able to see them.

The use of social media platforms in research raises a range of ethical issues, in particular regarding privacy and informed consent (see Townsend & Wallace, 2016). For this research, it was especially important to consider the possibility that participants might not know that their activities were systematically recorded by Facebook on their logs, and/or that they might not be fully aware of the types of information that their logs will display. To ensure participants’ informed consent and minimise intrusion, I used concrete examples to explain to participants the type of information that their FAL and FSH would contain before guiding them verbally on how to access them on their phones. Participants were asked to look at their FAL and FSH on their own first and were given a clear opportunity to withdraw their consent each time. Participants’ digital data were not recorded and verbal utterances (e.g. participants using names of people or places to describe their data) recorded on audio files were anonymised in the interview transcripts.

Using the FAL and FSH in interview settings also meant that I had to negotiate the specific power structures and layers of data circulation in which these features were embedded. These features have not purposively been designed for qualitative enquiry and knowledge creation. This meant that the probes relied on Facebook’s design of these features as well as on information selected and sorted by Facebook’s algorithms. The FAL presents only a partial and constructed record of the ‘activities’ users undertook on Facebook. For example, the FAL lists posts, comments, likes, pictures, friend requests but does not include activities such as scrolling down the platform or clicking on pages/profiles. In addition, the information displayed on the FAL varies in relation to users’ engagement with the platform. For instance, some of my participants used the search box instead of the chat bar or Messenger to send messages to friends. Because of these participants’ uses of the search box, a larger number of entries would appear on their FSH. By contrast, it is also possible to search and look at people’s profiles without them being recorded on the FSH. The FSH only includes the information typed in the search box and not practices such as clicking on profiles from the newsfeed or using the networked aspects of the platform to find people (e.g. friends of friends). In this way, the FAL and FSH are in effect limited and passive (as they do not include ‘movement’ activities such as scrolling down or clicking through other pages) recording of one’s activities and searches on the platform. Through these features, users are not given access to the full range of data amassed by Facebook but only to a fraction of it. However, this ‘invisible’ data are effectively reintegrated into users’ newsfeeds, shaping the content that they are exposed to and to some extent their engagement with the platforms.

From the onset of the project, I was aware of the shortcomings regarding the selection of the data displayed on FAL and FSH. These shortcomings, however, did not conflict with my research aims as I was interested in young people’s perceptions and understandings of their engagement with social media platforms and not in collecting information about their activities or searches per se. However, as I will discuss below, the use of FAL and FSH enacted different forms of liveliness, mediating in complex and unanticipated ways the data generated during the interviews, the proceedings of these interviews as well as potentially reinforcing specific understandings of data and technology.

Lively features: Negotiating dynamic data streams

The use of FAL and FSH as digital probes during interviews fostered a range of discussions and as a result generated ‘thick data’ regarding young people’s understandings of social media platforms. The probes facilitated the discussion of concrete examples and allowed participants to recontextualise their engagement with the platform within their everyday lives. Indeed, the design of the FAL and FSH as chronological data streams helped participants to contextualise the entries which appeared on their screens in relation to a date but also in relation to adjoining content. For instance, Luke checked his FAL during his interview to remind himself of his last activities on Facebook:

Luke made use of the chronological sequence of his log to remember and recontextualise his activities (i.e. yesterday, Saturday, Wednesday). The entries displayed on his log, then, became the starting point of an insightful discussion about the expectations and implicit codes of behaviours in relation to the tagging of friends on pictures and posting on someone else’s timeline. Similarly, Hugo’s (25) activity log sparked off a reflection about negotiating having parents on Facebook and the ‘politics’ of commenting. His log displayed a comment about gambling that he had made on a picture posted by a friend, only to delete it a few hours after. Hugo explained that his friend asked him to remove the comment to prevent his parents from knowing that he was gambling online. These two examples illustrate how using FAL in interview settings elicited discussion and generated thick data by encouraging participants to reflect upon their activities using their data. The FAL and FSH as dynamic chronological streams contributed, in this way, to what Kaun and Stiernstedt (2014) described as the ‘possibility of telling stories and performing memory’ (p. 1163).

However, the spatial and temporal organisation of the FAL and FSH as well as their dynamism also incited a move away from narration and instead encouraged an enumeration of successive entries with no beginnings or endings. The extract below from Emma’s interview illustrates this shift:

[Looking at her FAL] Yeah okay so that was just one of my classmates’ birthday today, so I just said, ‘happy birthday’. That is from the group … here we were discussing where to meet for the group project … Here one of my friends was saying that she is beginning her fourth year for her degree … ‘Page Name’, that is one of my friend[s] who is doing … well actually he is another person from my internship and I did friend him as well, but we were in different departments, so I don’t know him that well, but I just friended him. So, he just sent me a request to like his photography page, so I did that … right that is one of my friend and her boyfriend was saying that they are together and that he is really happy, something like that so I just liked it. (Emma, 22, my emphasis)

Here, Emma discusses very briefly each ‘entry’ before quickly moving on to the next, scrolling down almost visibly the dynamic and never-ending ‘stream’ of her activities. The log is impersonated by the multiple uses of the pronoun ‘that’ and of the adverb ‘here’. In the second half of the excerpt, Emma starts expanding on why she liked a friend’s page, providing interesting insights into implicit codes of behaviours on Facebook in the context of friendships. However, the liveliness of the log interrupted her thoughts by displaying more entries which she felt compelled to discuss as her use of ‘right’ to move on revealed. In this way, the dynamic stream of activities elicited discussion but also undermined the flow and pace of it by providing a dynamic and continuous list of entries to scroll down without allowing space and time to reflect on specific entries. Further, this dynamism also restricted the space and time that I had as researcher to follow up with participants on what they said. Indeed, participants not only tended to scroll down their data streams and move quickly from one entry to another but also looked at their phone screens, preventing me from making eye contact with them.

The FAL and FSH also triggered emotional responses from participants, evidencing the significance of emotions in users’ engagement with their data (Lupton, 2017). When first confronted with their logs, some participants were overwhelmed by or expressed surprise at the amount of data displayed on their logs. Some participants had clearly forgotten some of the activities that appeared on their screens, but this does not alone explain participants’ emotional responses. Participants were also receptive to how their information was presented. As shown in Figure 1 and Figure 2, FAL and FSH condensed all users’ activities or searches in one single column. The spaces between each entry of the log are not proportional of the ‘real’ time which has passed between them, illustrating one of the complex ways in which different forms of temporalities and ‘realtimeness’ are embedded within social media platforms (Kaun & Stiernstedt, 2014; Weltevrede, Helmond, & Gerlitz, 2014). The condensation of all activities in one place contributed to create an impression of their Facebook activities. Surprise was usually overcome by engaging more specifically with their logs, which indicates that the presentation of the log itself rather than necessarily its content was causing an emotional response.

These examples demonstrate the ways in which the liveliness embedded in the design of the features and more broadly of the platform (e.g. speed and the motion of scrolling down) shaped the content of the interview as well as the research situation itself, at times closing rather than encouraging discussion. The FAL and FSH organised not only participants’ discussions of their activities and searches but also their recollection of these activities, how they perceived them and also how it reflected on them. These processes were further exacerbated when participants’ FAL or FSH seemed to contradict their accounts of their engagement with the platform, as I will discuss in the next section.

Lively research situations: Negotiating conflicting accounts

As the previous section discussed, using social media features during interviews allowed and encouraged participants to retrospectively reflect upon the meanings of their past activities on Facebook while remediating these meanings in complex ways. The FAL and FSH also rendered visible and salient during interviews discrepancies between participants’ accounts of their practices and their practices as recorded in their logs. These discrepancies directly speak to the challenge of using devices designed to capture the ‘doing subject’ (Ruppert et al., 2013, p. 35) in interview settings with the risk of solely focusing on practices (i.e. the entries displayed on the logs) and what they seem to say in terms of users’ engagement with the platform at the expense of exploring users’ understandings and perceptions of these practices. This type of focus in the context of interviews can create unintended experiences of scrutiny, in turn pushing participants to account for, rather than giving their accounts of their practices. This section analyses such occurrences in the situation of the interview.

A common dissonance underlined by the probes during interviews was the seeming contradiction of participants describing not using Facebook ‘very much’ whereas their FAL displayed numerous recent entries. Such dissonances were mediated by participants’ subjective understandings of what constitutes an important usage of platforms as well as by the spatial layout of the logs (see discussion above). Occasionally, the FAL contradicted participants’ accounts in more apparent ways. Aaron (20), for example, who carefully emphasised the fact that he would not send friend requests during his interview was disconcerted by his log which showed that he had requested and friended someone the day prior to the interview. To cope with the apparent contradiction, Aaron distanced himself from his log by skimming through the entries appearing on his log as well as by describing the activities from the viewpoint of the AL using the third person:

[Looking at his FAL] Right … ‘Aaron sent James a friend request’, ‘Aaron liked a photo’ ‘Aaron liked a photo’, my brother Dean posted a link to my timeline. ‘Aaron and Carry are now friends’, ‘Aaron and Spencer are friends’ … yeah I’ve been meeting a lot of people since I am here. (Aaron, 20)

Similar contradictions between FAL and what participants had said triggered unease or awkwardness, in turn affecting the proceedings of the interviews. To account for these discrepancies and/or to save face, participants tended to explain the specifics of each entry that appeared on the screen and how the activities listed squared with their earlier accounts of their engagement with social media platforms. David, for example, provided a lengthy explanation of one of his Facebook posts after he realised that his log contradicted his account:

[Looking at his FAL] … I did actually do something on Facebook today. You know Furious 7? There was an advert to say … if there is an advert I’ll just write something negative on it but not like ‘why is there an advert here?’ but something stupid, so the advert came up and I was like because I have seen it and I hated it so I put ‘Furious 7 no point to seeing it, it is shit’ kind of thing, I don’t know just for a laugh. (David, 24, my emphasis)

By using ‘actually’, David acknowledged the discrepancy between what he previously said (i.e. not using Facebook that day) and what his log recorded. This type of forgetfulness demonstrates how Facebook has become deeply embedded in everyday life and more broadly reflects the functioning of social media platforms which relies on principles of immediacy, speed and forgetfulness (Kaun & Stiernstedt, 2014). Thus, dissonances between participants’ lived experiences of their engagement with Facebook and the digital traces of their activities on the platform were not surprising. However, by making apparent these dissonances the probes created a lively motion during the interviews which directly conflicted with participants’ accounts and made them feel that they had to account for their practices rather than giving their own accounts of their engagement with Facebook.

The use of FSH exposed similar discrepancies between participants’ accounts – and presentations – of their searching practices and the data displayed on their histories. For example, many participants asserted that they did not look up people on Facebook, mostly discussing these practices in impersonal and general ways. Using FSH raised discomfort as searching practices became suddenly visible and personal. For example, Molly (20) stressed during her interview that she would never look up people’s profiles on Facebook and that she felt very uncomfortable with these practices:

When my friends do it, like even when my friends are like ‘what is her name, I am just going to totally Facebook her’, I don’t like that. I feel, even though it is only information that the person put out to in the public view, I still find it very intrusive, I would never do that. (Molly, 20)

Later in the interview, she logged into her FSH, only to delete it straightaway explaining:

It looks really weird like if I have checked people out and I am like ‘no’, just clear that because I don’t want it, why would you ever want to remember this? … It is just kind of annoying like ‘ow no I’ve searched for these people, I look like a creep’. (Molly, 20)

Molly clearly did not want to be seen to use Facebook to look up people, especially after she had accentuated the fact that she would categorically not do it. When confronted with it and presented with an image of herself that contradicted her self-presentation, she deleted her FSH. In this case, the FSH undermined instead of elicited discussion.

Only Molly decided to delete her FSH and not use it during the interview. Other participants deployed a range of tactics to negotiate the seeming discrepancies between their accounts (i.e. not using Facebook to search people) and their FSH. For example, participants often detailed the context and specifics of each search to establish their appropriateness and legitimacy, distancing themselves from moral transcripts associated with ‘being nosey’ and the image of the ‘voyeur’ (see Gangneux. 2018, for a detailed analysis). Lucy (24), for instance, provided a lengthy justification of one of the entries that appeared on her FSH. She described the person that she searched for as ‘a friend of a friend’ who she had ‘briefly met’ during a night-out and explained how she only had a ‘quick look’ at the first page of the profile and how she would ‘not always do that’. Some participants clicked directly on the entries displayed on the log to show me the profiles of the people they searched for in order to demonstrate the legitimacy of these searches (e.g. close friend, searching someone before friending them).

These accounts demonstrate the capacity of FAL and FSH to reveal dissonances between participants’ practices and their presentation/lived experiences of their practices. However, by directly contradicting or being in conflict with participants’ accounts, the probes also created a new motion of scrutiny in the interviews in which participants felt that they had to account for their practices and that the probes were set up as a means to ‘check’ on what they had said. Eleanor (22) briefly glimpsed this possibility: while looking for her FAL on her phone she commented: ‘I just hope that what I said is actually accurate’. When asked to reflect on why she felt uncomfortable when seeing her FSH, Molly explained that she experienced the possible use of her FSH as a form of scrutiny in itself:

It is one of these things [looking up people on Facebook] that people do but it is different when you … it is different when you kind of look over … and see that someone searched for so many people. (Molly, 20, my emphasis)

As a researcher, I struggled to negotiate this unintended and unwelcome situation of scrutiny generated by the probes in which I was potentially perceived as both ‘checking’ on participants and on the veracity of their previous statements. As discrepancies became apparent, the interviews themselves became sites of negotiations of the liveliness of the digital data assemblage created by using the FAL and FSH within the interview settings. In these situations, the interviews often moved away from participants’ accounts and perceptions of their practices (which was the focus of my research) and instead seemed to encourage participants to account for their practices, generating awkwardness and unease. The perceptions and the responses to the probes in turn impacted on the discussion (or lack of thereof) and on the research situations themselves.

In addition, the ways in which some participants understood social media platforms and their functioning reinforced experiences of the interview as a form of scrutiny. Participants often perceived the digital traces recorded by FAL and FSH as objective, systematic and reliable recordings of one’s activities and searches on the platform. In Eleanor’s own words, they were seen as ‘accurate’, more so at times than their own accounts. Such understandings of technology have partly been constructed by social media corporations themselves which promote their platforms as neutral and technical structures while downplaying the role of the platform and the corporation in sorting, mediating and extracting content and data (Gillespie, 2010). By using the FAL and FSH in interview settings, I depended not only on how data were processed and sorted by the features of the platform but also on the pervasive representations of the platforms as technical structures and of data as objective. Thereby, the FAL and FSH performed an unintended but dynamic role in interviews by enacting objective recordings of ‘the doing subject’ (Ruppert et al., 2013, p. 35) and in doing so reinforcing experiences of scrutiny. Prior the start of the research, I was well aware that FAL and FSH were technological artefacts which presented and sorted a selected amount of information in specific ways (see methodology discussion). However, I did not fully foresee the ‘lively, productive and performative qualities’ (Ruppert el al., 2013, p. 22) of using these technological artefacts in interview settings, thereby producing specific and lively digital data assemblages which shaped the research data and the situation of the interview itself.

Conclusion

Digital platforms and data offer great opportunities for social research, not only in terms of Big Data and quantitative research but also for qualitative enquiry. For example, using social media data and features in qualitative research can be used as a strategy to ‘thicken data’ (Latzko-Toth et al., 2016). Given the ubiquity of social media platforms and their intertwining in everyday life, using features such as Facebook Activity Logs can inform research about social media but also more broadly examine social relations and practices which are enacted, experienced and captured through social media platforms such as for example friendships (Chambers, 2013), understandings of self (Lupton, 2016) or life narratives (Robards & Lincoln, 2017). In this article, I demonstrated that the use of the FAL and FSH in interview settings produced different forms of liveliness which allowed to produce thick data but also impacted on the data generated as well as on the research situation. While some of the issues that arose throughout my research are specific to the use of FAL and FSH, it is possible to extrapolate some of the broader political and theoretical implications of incorporating social media platforms and features in qualitative research.

By using FAL and FSH in my research, I was reliant not only on the design of these features (i.e. as dynamic datastreams) but also on the unexpected forms of liveliness generated by the ‘digital data assemblage’ produced by using these features in the context of a research interview. The main challenge lay in using platforms which were designed to (partially) record activities and ‘doings’ (i.e. the activities participants undertook on Facebook and captured by their logs) which in turn push participants to account for their practices as recorded in the context of the interview. Similar processes are likely to play out in qualitative research using other social media features in interview settings. Using digital platforms and data as probes in interview settings can also render visible discrepancies between participants’ accounts of practices and the practices as recorded and sorted by the platforms. Making such dissonances apparent can be useful to encourage participants to reflect upon their lived experiences of social media in relation to their uses as recorded by the platforms and explore the tensions between what people do and what they say they do. However, such probes can also create unintended situations of scrutiny, as this article has demonstrated. It is therefore critical that researchers carefully assess the ways in which such probes can be productive in thickening data and developing knowledge while taking into consideration the different forms of liveliness that the probes are likely to generate and their impacts on the research data and situation. To do so, it is important to consider the responsiveness of the features used in interview settings as well as the temporalities embedded in these features as they can be conflicting (e.g. time of the publication of data and its re-emergence in the situation of the research interview).

Finally, it is worth reflecting on how incorporating social media features in social research risks legitimising the use of personal data as a normal part of social enquiry and reinforcing an already pervasive ideology of dataism – the belief that all human behaviour can be collected through neutral technological channels, turned into quantitative data and objectively analysed (van Dijck, 2014, p. 198). These tacit understandings of data and their role in our contemporary society played a likely part in the forms of scrutiny experienced by some of my participants. While reproducing to some extent a neoliberal culture of measurement might be difficult to avoid, we urgently need to rethink the repertoires of sociology to take up the opportunities raised by digital data and platforms for qualitative research while avoiding the pitfalls of dataism. In this context, as social researchers we need to critically engage with the data and tools that we use, their architectures, designs and affordances but also with the broader ideological and economic systems of power relations in which they are embedded (Fuchs, 2014; Langlois & Elmer, 2013), and the ways in which they form an active part in mediating how we know the social (Ruppert et al., 2013). To respond to these challenges, we need to reflect collectively on the lively characteristics of using digital platforms and data in interview settings, and the complex ways the digital data assemblages we produce in doing so shape and feed back to the research data and the research process.