Introduction—Advances in Visual Methods

The ideas in this chapter have their heritage in the tradition of ‘visual organisation studies’, a field we have been involved in developing over the past fifteen years. In particular we have been working with participant-led photography as a way of gathering data in numerous participatory field studies located in a variety of organisational contexts, including hairdressing (Shortt 2010, 2015; Shortt and Warren 2012), office work (Warren 2002, 2008, 2014; Shortt 2018), hospitals, university buildings (Shortt 2019) and accountancy (Warren and Parker 2009; Parker and Warren 2017). We understand the term participant-led photographic field studies to describe a methodology where research participants generate image-based data that connect with an empirical investigation of some aspect of their lives (e.g. see Vince and Warren 2012; Shortt and Warren 2019). These images might be part of a participants ‘pre-existing’ personal collection or have been made expressly for the needs of the study, but a central feature of the approach is that the photographer should be the participant. As well as producing images, the participant is also asked to attribute meaning and/or an explanation of their image to the researcher either during a research interview, or by supplying a captioning sentence.

This approach to generating data is rooted in visual sociology and social anthropology (Bateson and Mead 1942; Harper 1998; Collier and Collier 1986; Knowles and Sweetman 2004), and is where the use of photography in social research has brought a different perspective to the texture of people’s lives, homes, communities and working practices. The past twenty years have seen significant contributions in this field and a growth in the debates in favour of visual representation rather than a reliance on purely textual accounts (Banks 2001; Pink 2001), as well as arguments for the value of more participatory methods. Placing the camera in the hands of the participants, the ‘researched’, allows for ‘native image making’ (Wagner 1979) and raises the voices of those that traditionally may not get heard and addresses the power balance between the researcher and the researched (Warren 2005). These advances in visual studies have also been propelled by the visual literacy present in our contemporary culture (Knowles and Sweetman 2004) where both visual representations of everyday and organisational life have become prevalent, as well as how we practically connect with, capture and share visual data through devices and online platforms.

Indeed, over this time, we have seen a fundamental shift in the way qualitative researchers have approached the use of visual methods, with technological advances being a major driving force for innovation. At the beginning of the twenty-first century, visual methods in business and management research were really just emerging as technology with cultural shifts enabling their spread and development. Over the last decade or so we have witnessed a growing interest among more mainstream management communities in harnessing the power of the visual to better understand organisational contexts (e.g. Meyer et al. 2013). In 2007 the ESRC funded International Network for Visual Studies in Organization (inVisio) was founded, at the same time as the European Institute for Advanced Studies in Management series of conferences on ‘Imagining Business’ (2008, 2011), followed by the 2010 Standing Conference on Organizational Symbolism themed ‘Vision’. Management journals have published special issues on the visual including, the Accounting, Auditing and Accountability Journal (2009), Qualitative Research in Organizations and Management (2012) and Culture and Organization (2012), and two edited collections have been commissioned by the international publisher Routledge; Quattrone et al.’s (2011) Imagining Organziations, and Bell et al.’s (2014) handbook, The Routledge Companion to Visual Organization. Recognising this groundswell of interest, the ESRC further funded inVisio through a researcher development initiative (2010–2012) tasked with building capacity in visual methodologies among business and management researchers (www.moodle.in-visio.org).

Certainly, the popularity of visual methodologies is likely to grow in the future, particularly since the visual culture we live and work within has seen such exponential growth over the same period with the digital revolution, most notably the emergence and rapid rise of social media. 71% of North Americans have a social media profile and the worldwide number of social media users is expected to reach 3.02 billion by 2021, a threefold rise in just 10 years (Statista 2019a). The global number of Instagram users is now 1 billion (Statista 2019b). These statistics demonstrate how widespread these new communication tools are in the lives of global populations. Importantly for this chapter, we are also seeing a shift towards circulating and sharing images as well as text-based content (such as status updates and ‘tweets’)—3.8 trillion photographs are estimated to have been taken from the invention of the camera until 2011, yet 1 trillion were taken in 2015 alone (Kane and Pear 2016). This image-explosion is in large part due to advances in Internet bandwidth, mobile data networks, smartphone storage and cloud computing. However, as we have argued at length elsewhere, (e.g. see contributors to Bell et al. 2014; Warren 2009, 2018), the visual has been a powerful communicative medium taking up a variety of technological forms throughout its history (Manghani et al. 2006) and so these developments should be seen in that light.

Our specific interest here is the role visual social media now plays in everyday life offering a wealth of opportunities to researchers who wish to explore the activities, behaviours, views and experiences of their participants. Specifically, our aim in this chapter is to contribute to a ‘new wave’ of visual studies in a pragmatic and useful way, as well as a step-by-step guide on ‘how to…’ set up, design and manage a research project that incorporates such platforms. We aim to help researchers make well-informed decisions when considering how visual social media might be used in organisational field studies. This is important for the development of visual organisation studies as a methodological field,—not least to assist with gaining ethical approval from institutional committees and review boards.

Why Use Social Media/Instagram in Organisational Field Studies?

The past five years have seen a growing body of research examining social media and its opportunities, challenges and risks for researchers. There are now a considerable number of studies of social media in relation to a range of organisation and management research questions. From questions of ‘free labour’ in the production of marketing content for brands on Facebook (Beverengen et al. 2015), to analyses of individuals’ visual identity in Facebook (Uimonen 2013), through to the use of ‘selfies’ in constructing consumer identities (Iqani and Schroeder 2016; Kedzior et al. 2016), including how brands are destabilised by user-generated content (Rokka and Canniford 2016). Of the top 100 brands, 90% have an Instagram account, and a reported 60% of users have found new products through Instagram (Brandwatch 2019) which shows the potential Instagram may have for this kind of study. If we widen the scope of social media to include company websites, there is a broad range of research taking into account how organisations portray, construct and seek to control their images in a digital age. Interesting examples include BP’s ‘greenwashing’ through images on its website (Kassinis and Panayiotou 2017), the marketing strategy for MBAs delivered by leading business schools (Elliot and Robinson 2011), and analyses of anti-capitalist ‘viral’ videos (Bell and McArthur 2014).

However, there is a very little methodological commentary on the potential of user-generated visual social media content in organisational research. This is perhaps surprising given the everyday character of social media image-sharing and the extent to which users routinely use the app to communicate the goings-on of their day to one another. 60% of Instagram users log in daily, putting it second only to Facebook in terms of user engagement (Brandwatch 2019), and in 2016, users posted some 3.5 billion photos every day (ibid.). Methodological advice is now beginning to appear in relation to using social media in social research (e.g. Sloan and Quan-Haase 2017). Instructional texts on how to conduct ethnographic research using the internet are also now quite mature, with some entering their 2nd editions, e.g. Kozinets (2009) landmark text on ‘netnography’ which is now titled Netnography: Redefined, (Kozinets 2015) such is the speed at which social media and Internet technologies are changing. But save for Laestadius’s (2017) chapter on ‘Instagram’ and Hand’s (2017) chapter on ‘Researching Social Media Images’ in the Sage Handbook of Social Media Research (Sloan and Quan-Haase 2017) there is scant advice for a researcher who wishes to exploit these technologies for their field-study research, and understandably, few examples of researchers doing so. One exception to this is Sergi and Bonneau’s (2017) work-in-progress, mining the hashtags that people give to their Instagram images of work. Although at an early stage of analysis, their findings show how assembling image-sets based on hashtags such as #officelife, #bankinglife, #workingonasunday, etc., reveal intimate, mundane and often backstage dimensions to everyday working lives that are not usually accessible to anyone but the participant. And in addition, McKeowan and Miller’s recent article, #tableforone: Exploring Representations of Dining Out Alone on Instagram (2019 forthcoming), which also uses the visual data created by the image-set based on the hashtag #tableforone to explore individual’s experiences of dining out alone and what more this can tell us about individual leisure practices in public spaces. Although such studies are vital in advancing our understanding of using Instagram as a resource in visual research, they nonetheless only focus on visual data already produced by participants, and not on images explicitly produced as part of a research project.

So, in line with our description of participant-led field studies above, we believe there are two main ways in which we see Instagram as a useful data generation tool for participant-led field studies. Firstly, the images on participants’ existing Instagram accounts can be seen a personal archive that can be analysed in order to generate data about their lifeworlds and habits. The two studies we mention above (Sergi and Bonneau 2017; McKeowan and Miller 2019) are most closely associated with this approach, their strength being that the participant’s time commitment is minimised (as the images already exist). Indeed, it may not be necessary for the participant to physically meet the researcher at all, because providing the Instagram feed is set to ‘public’ or access is granted to the researcher, the images will be accessible on any internet-enabled device capable of displaying graphics. Furthermore, since the images were generated from a ‘natural’ setting this might appeal to researchers striving to minimise their influence on the data collection. There is also an inherently historical and/or longitudinal aspect to Instagram feeds which could prove very useful to researchers interested in research questions with a temporal character, for example socialisation into a new profession, experiences of pregnancy or serious illness at work, or management of home-work identities. Particularly relevant to the last of these examples is the additional affordance of Instagram in that you can only upload photographs on a mobile device, meaning that images are emplaced (Laestadius 2017). Gomez Cruz (2016: 337) suggests this offers exciting possibilities for ethnographers to capture participants ‘trajectories’ for example by providing an emotional and sensorial way of virtually retracing steps. During research interviews, these images could act as powerful aide memoires for past happenings, given the iconic nature of photographs and their intimate fluting with memory and emotion (see Edwards and Hart 2004). Finally, for certain populations, the fact that Instagram may already be embedded in their visual/sociocultural practices offers the opportunity to probe those practices in ways that are already familiar to the participant, for example, asking why did you take that photo, like that, or use that filter and so on.

The second way we see potential in Instagram for participant-led field studies forms the rest of this chapter, through our worked example below. In short, this positions Instagram as a convenient repository and sharing tool for photographs made by participants expressly for the needs of the study, as a kind of diarying or storying technique.

The Case of #myUWEBBSview: A Participant-Led, Visually Led Study of a Business School

In order to better explain the key things to do and consider when using Instagram as tool to generate and share data for the needs of a project, we present a case study drawn from our own research, called ‘My UWE BSS View’—a visually led post-occupancy evaluation of a new Business School building at a UK University. At the time of writing this chapter, the project is ongoing.

The Business School: A Visual Approach to Post-occupancy Evaluation

In April 2017 the Faculty of Business and Law moved into a new £55 million building on the Frenchay Campus of the University of the West of England (UWE), in Bristol in the UK. Bristol Business School, as the building is named, is now home to over 300 staff and 6000 students, and the building represents part of the wider University’s ‘Masterplan’ project (UWE 2019) that aims to see the university develop one, consolidated campus and to create buildings and facilities that provide the latest facilities and learning environments for all users.

The architects of Bristol Business School, Stride Treglown, and the construction company, ISG Construction, approached Harriet and other members of the research team to suggest a collaborative research project examining the post-occupancy experiences of users in the new building could provide a unique and timely opportunity to investigate everyday life in this new space. Specifically, our industry funders were keen to design a post-occupancy evaluation (POE) that extends beyond current approaches to POE that are predominantly quantitative in nature and often provide only statistical information that focuses on the mechanics and practical functioning of a new building (see, for example, Capita 2017; HEFCE 2006; Williams 2001). In addition, Stride Treglown and ISG, who frequently specialise in designing and building Higher Education spaces, wanted this research to provide them with in-depth, rich data with which they may translate into further learnings for future buildings.

With this brief in mind, we decided to experiment with using Instagram to gather data about the building users ‘on the spot’, everyday encounters with their surroundings. Our rationale for doing so was that the majority of our potential research participants would be in possession of a smartphone with camera, enabling instant upload of images (and captions) throughout the course of their normal day; the Instagram app is free to download and use and is widely known; it has the capacity to include hashtag identifiersFootnote 1 that would make identification of the project data straightforward; and it includes functionality to caption images at the time of upload, something of importance to the method as we discuss further below. What we had not anticipated was the resistance to using Instagram that we would encounter on account of its deeply embedded social function as an identity-work device. But we will return to those issues later in the chapter.

Research Design

Using social media, promotional postcardsFootnote 2 (see Figs. 1 and 2), a project website (see Fig. 3) and other modes of communication, we recruited participants by asking them to take pictures of their spatial experiences in the building that addressed two simple questions that directed them to their sensory and ‘emplaced’ (Pink 2009) experiences:

  1. 1.

    How do you feel about the building?

  2. 2.

    How are you using the building?

Fig. 1
figure 1

One of the promotional postcards used for the research project

Full size image
Fig. 2
figure 2

Participant information on the back of our promotional postcards

Full size image
Fig. 3
figure 3

A snapshot from the ethics pages of our project website

Full size image

We also asked participants to add a short caption explaining their reasons for taking the photograph, then post their images to Instagram using a dedicated project hashtag: #myuwebbsview. The importance of captions in participant-generated visual research will be discussed in more depth below, but briefly here, it is needed to enable appropriate analysis, ensuring the image can be coded according to the meaning it has for the participant (see also Shortt and Warren 2019: 543).

There was a great deal of discussion among the project team and the industry funders around what would be a suitable hashtag, given it has to be simple and memorable but equally not something so common that it could be used by another group, thereby contaminating our feed with pictures from other online discussions, debates and groups (Laestidius 2017: 576), and given #myuwe and #myuwebristol were hashtags already in use through the UWE Marketing team and followed and used by students, it made sense that our own project hashtag included this. In addition, the abbreviation ‘BBS’ for ‘Bristol Business School’ was part of a common language internally and externally.

Overwhelmingly, one of the most complex parts of this project was seeking ethical approval, particularly given the use of Instagram and the public nature of the platform, as well as the potential for such a large sample size—all users of the building were being invited to take part and this included all visitors, staff, students and services. It is fair to say that ethical practices, guidelines and decisions made by university ethics committees are not keeping pace with the changing nature of visual methods especially with regards to social media and ‘best practices’, since these guidelines themselves are still in their infancy. In our own practice, we have used guidelines set out by, for example, the International Visual Sociological Association (Papademas and The International Visual Sociology Association 2009) in order to ensure all project stakeholders were treated with appropriate respect and protection, particularly with regards to participant-led photography (Vince and Warren 2012). The main concern with such a project is how to maintain anonymity and privacy, and we worked closely with the university ethics committee to develop the final guidance. For example, we gave advice on how to blur faces in images that contained people, if prior permission to capture that person had not been granted, or if it was not possible to seek that permission. So, we included the following statement in our participant information online:

Taking care taking pictures:

If possible, ask people for permission if they are the subjects of your photographs. Do not take photographs of confidential material and take care not to photograph anything which invades another person’s privacy or contravenes your organizations’ confidentiality policy (for example, visible contents of documents or computer screens). We will ensure all information is confidential and privacy remains protected. Any private information about the individual will not be made public with personal identifiers. All sources will remain anonymous and we will ensure any public use of visual data will adhere to consent agreements made with participants. During the lifetime of the project – January 2018 to December 2018 – a selection of images posted on Instagram will be re-published on the pages of our dedicated website. These images will be selected and moderated and agreed by all Parties (UWE, ISG and Stride Treglown) before publishing, including anonymising any obviously identifiable individuals.

We also gave guidance reminding participants to adhere to the social media terms of use for both the organisations (UWE) and Instagram itself, and the issues surrounding anonymity. A such, we also included the following paragraphs in our participant information online:

You may only take part if you are over 18.

By posting your images and comments on Instagram or by emailing your images and comments to the address above, you are consenting to your images and comments being used in research and industry publications. You will remain anonymous i.e. your Instagram username/email/name/personal details will not be identified. You may be referred to as ‘student’ or ‘staff’, for example. See below for more information on confidentiality. If you have any questions about the project, please email us on myUWEBBSview@uwe.ac.uk or get in touch with one of the research team.

If you would like help setting up an Instagram account in order to get involved, please email us on myUWEBBSview@uwe.ac.uk or get in touch with one of the research team and we will happily help you!

The Instagram feed:

#myUWEBBSview will be viewed publicly and will be used as a tool for discussion in a number of research-based focus groups and interviews with members of e.g. the student body, academic staff, alumni, the Executive team, and external business partners.

Please adhere to UWE Social Media regulations and take particular care not to capture personal identity information in your photographs

Please adhere to Instagram.

You are free to withdraw from the study without needing to justify your decision and without prejudice. The fixed and final date for withdrawal is the 31st December 2018. Should you wish to withdraw for any reason, any data you have contributed will be excluded from the research.

In addition, and given that our participant group was potentially so large, we had to consider how best to communicate this participant information and our ethical guidelines and permissions. We used the promotional postcards (see Figs. 1 and 2) and the project website (see Fig. 3) to do this and gave detailed notes to this effect in our ethics application.

Encouraging Participation: The Social Character of Instagram

Participant take-up and use of Instagram to contribute to the project was initially encouraging and overall, the project did generate some 267 useable posts for analysis (although this is in addition to over 500 images received via the private project email account!). However, it soon became noticeable that student users of the building were reluctant to engage with Instagram (but they were sending images and comments to the project email account). We also found staff and visitors to the building seemed more willing to send their images and accompanying captions/commentaries to the dedicated email address than they did to post to Instagram. Whilst this was also part of a wider issue of student non-engagement with the project more broadly,Footnote 3 investigation revealed some observations that are pertinent to our discussion of visual social media research here. Some of these we did anticipate, but others were quite unexpected, quite possibly overlooked in our enthusiasm for this experiment.

When we spoke to students about why they were not using Instagram to share their experiences and feelings towards the building (and seemingly preferring to do this privately through email), they said the following:

…but Instagram is for the presentation of the best self

Why would we contaminate our feeds with pictures of work?

…students now are too cool for school. They aren’t going to do anything if no one else is doing it. It’s like, why would we do something out of the ordinary or different to what everyone else is doing

Well it would be weird – you are creating this online identity and then, well, I wouldn’t post a random picture of some chairs or something…

Instagram, therefore, is a site for the construction of a very particular type of identity. It is seen by these students as a place for presenting an aspirational and/or polished identity and the work that is put into this—through considered images, staged and posed posts, filters and so on creates a particular narrative. The notion that one would post something ‘alien’ and misplaced in relation to this narrative—like pictures of a building and their place of work/study—was defined as ‘contaminating’ ones carefully constructed social identity. Having to explain ‘odd’ posts to family, friends and followers was not something that was welcomed by these students.

We find these comments and insights from student fascinating. So, rather than see the reluctance from our ‘millennial generation’ participants to engage with using Instagram as a problem for visual social media research of the future, we prefer to think how we can turn it to our advantage as researchers. In the #myUWEBBSview study we are discussing in this chapter, we saw Instagram very much as a data ‘collection’ tool, albeit one that seemed particularly suited to generating data about people’s sensory experiences in space and time, and therefore fitting for a project evaluating experiences of a building. However, when seen from the position of the social role Instagram already plays in young people’s lives beyond its technological capacities as a tool for generating, recording, storing and sharing visual data, we can see Instagram offering very important possibilities for projects where self and identity are of particular interest. Asking people to create Instagram posts potentially surfaces the very processes you are investigating and is ripe for future development as a method.

Managing Data During the Project

There are several practical steps to consider in the management of project data posted to Instagram for a research project, not least in this case because the study had the potential to generate a large data set on account of considerable numbers of posts to the hashtag. We have already discussed decisions around appropriate hashtags, and subsequent steps included:

  • Curating the feed to the project website

Information and interaction with participants were managed through a central project website www.myuwebbsview.co.uk which housed information, as we note above, about the project rationale, data use, confidentiality and anonymity and consent. Here it is useful to note that having one focal point for the project simplified recruitment literature since we were able to keep project advertisements relatively uncluttered, instead directing participants to in-depth information elsewhere.

The architects, Stride Treglown, wanted to work with us on developing this site and suggested that we link the front page of the website to Instagram in order that we create a ‘live feed’ directly from Instagram to the website for the project (see Fig. 4). However, following discussions with our university ethics committee, it was felt great care had to be taken if we were to ‘repost’ Instagram data onto our university-and industry-endorsed website. This showed us that the live Instagram feed is not something a researcher can have direct control over—the content posted, is, by the very fact that it is a public platform, not controllable. But we were able to ensure the posts on our website were deemed appropriate and acceptable by all project stakeholders, as well as adhere to UWE social media terms of use. This did not mean we couldn’t have anything ‘negative’ or ‘bad’ posted about the building, simply that we needed to ensure we were not seen to be endorsing any content that represented anything illegal, explicit or inappropriate. We therefore installed a filter on the website in order that members of the research team could then log on and ‘approve’ (or not) the images that were posted using the hashtag. During the twelve months of the project data collection, this approval activity was done every 1–2 weeks.

Fig. 4
figure 4

The project website with the approved and filtered Instagram feed

Full size image

  • Data cleansing

Before we could begin to analyse the posts to the Instagram feed, we had to make decisions about what would count as appropriate data. One of the downsides to having a public Instagram feed, advertised extensively around a university campus, was the number of images that were hashtagged to the project without really being connected to the project’s aims, or following the instructions we have given. Firstly, we excluded all images posted without captions since there was no way we could undertake the first stage of analysis without some attribution of meaning by the photographer.

Secondly, the hashtag was taken up by the university marketing and PR departments, who used it indiscriminately to tag generic images used to communicate general Faculty announcements and events, and so all those posts were disregarded. Thirdly, we decided not to use images which, although posts from building users, were not actually about the building. Examples of this category included photographs of a social gathering for International Women’s Day where the photographer was only referring to a feeling of pride in the delegates. This last ‘cleansing’ decision was taken on account of the clear brief we were working to: to gather users’ felt, and emplaced experiences of using the building in their day-to-day working and studying lives.

  • A mechanism for converting the Instagram posts into an offline form amenable to analysis

Instagram feeds and images are non-downloadable from the website, and any app or software that claims to be able to download Instagram images are no longer permitted by Instagram’s terms and conditions of use. Similarly, the only sharing of images from Instagram that is possible is using the platform’s own sharing tool. To circumvent this issue, we displayed each image and its attendant comments full screen on a laptop and created a screenshot of each. We then printed these in full colour, on A4 sheets to facilitate the manual coding and Grounded Visual Pattern Analysis (GVPA) image-sets that we describe in more depth below (Shortt and Warren 2019).

Analysis and Making Sense of Instagram Posts

Throughout photographic studies, how researchers have made sense of their data has varied. For example, content analysis lays down a set of guidelines on how to categorise and draw inferences from the detail of people and objects depicted in the image (Rose 2013). Semiotics and iconography (including social semiotics) equip researchers with a framework to ‘read’ the grammar of images by interpreting the content of the image as signs which signify and produce social meanings (van Leeuwen and Jewitt 2001). Finally, critical visual analysis subjects the image to an interrogation of its composition, but also takes account of the context of its production and the effects it produces as it is circulated and ‘consumed’ (Schroeder 2002). Other analyses have interpreted visual materials in a more holistic manner, regarding the ‘aesthetic’ produced by images as irreducible to its constituent parts (Hancock 2005). Most of these studies privilege the researcher’s reading of the images rather than the audience for which the image was intended, although see Cho et al.’s (2009) study of viewer responses to online corporate social responsibility disclosures for a notable exception. In the #myUWEBBSview project we are describing here, we needed an analytical method that would allow us to take into account the intentions of the photographers as well as consider the subject and style of the photographs themselves.

With this in mind, we used an analytical framework that we have recently developed called (GVPA—Shortt and Warren 2019). This approach was created as part of our ongoing search for a robust and systematic analytical method for photographs generated during field study research. GVPA provides visual researchers with a means of analysis that mines both the attributed meanings given by the participants and the visual content of photographs themselves. This is a difficult tension in this kind of visual research since it is important to retain as much information about why the photographer took the pictures (since that is the point of asking the participants to take the photos in the first place)—what Meyer et al. (2013: 513) call ‘dialogical’ visual research, whilst recognising that there may be additional interesting information to be gleaned from studying the content of the photo independently—what Meyer et al. (2013) call the ‘archaeological’ approach. This is because there are always cultural precedents and aesthetic considerations that come into play when we take a photograph, and often we are not aware of their operation. For example, why do we centre our subjects when we take pictures? Indeed, to off centre them is seen as ‘arty’. However—and herein lies the rub—these cultural traces and subconscious image practices only make sense when viewed as part of the context within which they were produced. We explain further through a walkthrough of our process in the #myuwebbsview project. Although much of what follows is not unique to research using Instagram and/or other visual social media, this section is an important ‘how to’ commentary on an aspect of visual research which is often overlooked. For that reason, we have dwelled at some length on how to analyse participant-led visual field study data as a practical complement to our introductory paper on GVPA that appears in the journal Organizational Research Methods (Shortt and Warren 2019).

Stage 1: Dialogic Analysis

Working with the ‘cleansed’ data set printed from the main Instagram feed, the first stage of GVPA is to allocate a code that sums up the meaning the image seemed to have for the participant—hence the ‘grounded’ element of the approach. The aim of dialogic analysis (Meyer et al. 2013) is to identify a set of themes that describe why people took their photographs as they recounted them to us through the caption, they had added to accompany their image. Without the captions to ‘ground’ the images their value as signifiers about the participant’s lifeworld—in this case their experiences of being in a building—would be lost.

We discussed and coded each image in turn by adding a post-it note to each printout and collating on a summary spreadsheet. Figure 5 opposite is an example of an Instagram post to the project #myuwebbsview from @curiousview. This process proved harder than anticipated, because we are both used to interviewing people at length about the photographs they produce for our projects (e.g. see Warren 2008; Shortt and Warren 2012). Consequently, we kept finding ourselves longing for more context, or explanation, and generally felt uncomfortable that we may have been misconstruing people’s intentions. This was something we had not anticipated and is a feature of any social media format where lengthy commentaries on images are difficult to add to the platform.

Fig. 5
figure 5

Example of Instagram post with caption/attributed meaning

Full size image

However—and this is something that we have yet to mention—working with a large data set always necessitates a trade-off between richness (small n studies) and breadth (larger n studies) and visual research is no different. Therefore, we accepted that our data could never be as detailed as that produced from smaller samples, but we were realising a different set of benefits. These included generating a lot of pictures, over the course of a year, that were ‘snapped’ in situ by people whom we didn’t need to meet and who only needed to take a few moments out of their day to take part in the research (as opposed to a formal interview).

Working together helped our coding decisions significantly since we were able to sense check each other and discuss the various merits of focusing on different interpretations of what the participants had said in their captions. We tried to be as literal as possible as we show through the example of the sofa image—Fig. 6 shows how we coded each image with its caption.

Fig. 6
figure 6

Coding each image with its caption

Full size image

As we progressed through the images, we accumulated piles for each code. In line with traditional qualitative thematic and/or coding methods such as template analysis (King and Brooks 2017), codes that were too niche or similar to one another were clustered together, and more densely populated codes were expanded into more fine-grained classifications (King and Brooks 2018: 225; Saldaña 2009). Finally, after every image had been coded (a total of 746 useable images across both Instagram posts and emailed images and captions—see Fig. 7), we took the X most significant themes that resulted (e.g. quite literally, the biggest piles!) to take forward to Stage 2 of GVPA, pattern analysis.

Fig. 7
figure 7

Coding all images and captions and creating themes

Full size image

Stage 2: Visual Pattern Analysis

In Stage 1, the ‘grounded’ part of GVPA takes place through dialogical analysis. Once the most noteworthy themes are identified, as we have explained above, we can then turn attention to the images themselves—e.g. what they are ‘of’ and how they have been ‘taken’—in order to see if field-level visual patterns can be recognised across the themes. This proved to be a more exciting and energising process than Stage 1, as it is here that the analytical skill of the visual researcher really comes into play. The first step is to lay the images out next to one another so that all the photographs in each theme can be seen as an ‘image-set’ (Shortt and Warren 2019). The example we are using from the #myUWEBBSview project is the theme of ‘visibility’ and the image-set can be seen in Fig. 8.

Fig. 8
figure 8

Image-set for the theme of ‘visibility’

Full size image

A large flat space is required because it is the collective impression of the set as a whole that is the level of analysis here (Collier 2001; Shortt and Warren 2019). It is also important to ensure your image-set is logically cohesive, for example we made sure that all the photographs by students were grouped together, as were those taken by staff since we considered these were significant grounded details for this project (part of our brief was to explicitly investigate these two building user groups). However, we still retained both groups in the same image-set since all are building users and interesting patterns or divergences between the two might be missed if they were separated into two distinct sets.

As with Stage 1, recording of the impressions and patterns was joint, between the two of us in conversation. It involved getting into a ‘zone’ where we tried to suspend our habitual desire to look ‘through’ the images as representative scenes. Instead we tried to look at what was ‘actually there’Footnote 4 using techniques like turning images upside down, reordering them, looking at the set from different sides of the table and in a large mirror, and even asking Harriet’s 4-year-old daughter what patterns she saw!

We recorded the patterns in two columns—one for symbolism (the objects, subjects, places, events depicted) and one for composition (camera angle, lighting, aesthetic effects, point-of-view and so on.) We used a list of prompts to ensure we were thorough and systematic. It is important to bear in mind the technical affordances of the photographic medium and/or social media platform used in the project when considering aspects of composition, lighting and effects. For instance, Instagram includes a range of automatic filter templates that can be quickly applied to an image to change its look and feel, and it is also possible to manipulate the image further by changing presets such as brightness, saturation, sharpness, contrast, blurring and so on. However, these tools are not available to the user when using Facebook or Twitter. Furthermore, the cameras on smartphones can vary considerably, along with the skill and technical acuity (and desire!) of the user, so when recording compositional observations about your image-set, it is the general impressions and larger-scale impacts that are most useful since these are most likely to show field-level patterns across the sample. Table 1 sets out our Stage 2 observations for the image-set from the theme of ‘visibility’. The columns are mutually exclusive lists in the sense that the items in the same row do not necessarily bear any relationship to one another. Our actual work-in-progress notes were written on flip-chart paper and a photograph of this can be seen in Fig. 9.

Table 1 Stage 2 symbolic and compositional viewing record
Full size table
Fig. 9
figure 9

An example of our work-in-progress notes, written on flip-chart paper

Full size image

Stage 3: Theorising

As we noted down our symbolic and compositional observations, we added a third column to our record sheet to aid Stage 3 ‘theorising’ (see Table 2). In this space we wrote our analytical reflections on what the observed patterns might mean in terms of answering our research questions—but also beyond these to allow us to develop bottom-up theory about space, behaviour and identity in a business education setting. Thus Stage 3 ran (in part) concurrently with Stage 2, with interpretive notes and links to research questions and wider theory recorded alongside the patterns.

Table 2 Visibility theme with Stage 3 column added
Full size table

The unique contribution of GVPA is that theory is built from more than what research participants say about their experiences. It is also developed from what Meyer et al. (2013) called ‘sedimented social meaning’ apparent in the photographs. Rather like archaeologists developing theories about long-dead civilisations from the fragments of pottery and buried architectural traces, looking for visual patterns in image-sets involves interpreting multisensory, emplaced, spatial aesthetics as clues about how our research participants are working and studying in their space. At times this feels like ‘clutching at straws’ but this is a normal part of sense-making and we quickly learned to trust our impressions and think analytically about how and why our observations were significant.

Our analysis is still nascent at the time of writing, but some of our early thoughts about what the GVPA means for the theme ‘visibility’ are noted below (we have highlighted elements from our symbolic and compositional viewing in bold):

  • The visual patterns in Stage 2 show a dark, muted and bland aesthetic across the image-set, with an obvious lack of people captured in the images. This poses an interesting juxtaposition when we consider the ethos of the building design—a flagship space to attract international and home students, facilitate links with businesses, and provide collaborative spaces for everyone to work together. There is a sort of ‘mis-match’ here between the planned social identity of the building—one that encompasses community, versus the lived aesthetic identity of the building—one that, as the image-set suggests, lacks people and a sense of vibrancy.

  • In relation to the point above, the image-set also depicts materials and spaces that connote ‘visibility’—glass, windows, walkways, communal spaces. Our participants (users of the building) have therefore captured the materials used for transparency and visibility to capture their feelings associated with this very experience. So, we could argue that the design of the building has achieved its aim—to be as transparent as possible, where activities are made visible and the blurring of boundaries between staff, students and visitors is created. However, from these data, we see people’s experiences of this are ambiguous, and at times, paradoxical; for example, spaces are used for both casual social activity (people sitting in groups/casual-looking meetings), yet at the same time we see patterns across the images that suggest the building is bland and anonymous (no faces, muted colours, empty walls). Thus, our experiences of visibility and transparency are complex and not as homogenous as building designers and planners may hope.

  • The visual patterns in Stage 2 also show long shots, no close-ups, and images where photographers are at a distance from the subject and/or are standing on walkways, positioned as spectators. There is a sense that people are broadly surveying their internal landscape. There is a sense of surveillance, watching and having an ‘overview’ of the goings-on in the building. For us, this raises questions about ‘seeing’ and ‘being seen’/the ‘observers’ and the ‘observed’. Another paradox can be noted here as we see spectators and voyeurs watching daily activities of others (standing on walkways as spectators, looking into internal offices/teaching rooms), yet at the same time we see some of those behind the glass or in offices (often captured as the subject of the voyeurs’ image), seeking privacy and concealment (doors are shut, and paper is cello taped to glass windows from the inside). This encourages us to question the tensions that exist between almost exhibiting and showcasing work as a ‘performance’ to be seen by multiple audiences (working in a glass walled teaching room), whilst at the same time understanding the professional work that is done in a business school—teaching, learning, writing, deep thinking, reading, marking, and whether or not it is always appropriate for such activities to have an audience. This in turn, raises further questions about how we signal ‘do not disturb’? Being visible to others suggests we are therefore available to others, and if we are made available through the use of the fixed and permanent materials used in a building, then how do we temporarily change them when moments of solitude are needed or chosen?

Conclusions

In this chapter we have discussed how Instagram can be used as a tool in a visually-based, participant-led field study. We have set out an approach that researchers could use in order to gather visual data generated by participants and how such data sets, be they small or large, can be analysed using GVPA (Shortt and Warren 2019). In particular, we have argued that visually-led tools such as Instagram work well for those research studies that wish to investigate the socio-material, aesthetic, sensory parts of everyday working life and as such, broaden the scope of how and where data may be gathered; online platforms like this allow us as researchers ‘access’ to those participant groups we may not otherwise be able to reach.

We have also mapped out some key elements to consider when designing research that uses social media, such as ethical considerations, managing online data and how to engage participants. We have reflected on some of the challenges we have faced during our own Instagram project and the unexpected reluctance from particular user groups. This exercise has helped us question some of the assumptions we can make about the visual culture and social media practices of those we research.

At this juncture it is worth noting however that this form of research is constantly changing. Social media platforms and our own devices are adapting and their functions, features and settings are frequently subject to updates and improvements. Even since the design of our own research project discussed here, the use of Instagram Stories has dramatically risen amongst users (since this feature launched in August 2016) and, retrospectively, this could have offered an alternative mode of gathering visual data from our participants, or perhaps enhanced the approach we were already using. Indeed, Instagram is now working on a new ‘green screen’ filter, which may be called ‘Background’, that allows users to ‘select an image from your camera roll that you can then use as the background for the Stories camera. After you’ve selected a photo, your actual surroundings are replaced by the image’ (Social Media Daily 2019). Such developments pose new questions, considerations and challenges for researchers wishing to engage with social media platforms in terms of planning and designing research studies as well as viewing, analysing and making sense of their data sets.

We hope this chapter has helped those who wish to use social media as a methodological tool in their visual research, but also those who are already established users of visual methods—we hope our reflections enable you to think about how visual methods, and photographic methods in particular, are expanding, what alternative methods are available, and what exciting new opportunities for methodological advancement are possible.