Early on in the research I conducted a pilot study to to reveal issues and barriers related to recruiting potential participants, to explore the use of oneself as a researcher in a culturally appropriate way and to test and modify interview questions. Although familiar with Twitter as a participant, conducting a pilot study also allowed me to gain familiarisation as a researcher. I tested six different methods which I describe in more detail here. The table below reflects on the outcomes of the pilot methods.
From the pilot study, I took three methods forward – participant observation, semi-structured interviews, blog post analysis and interviews – but these were supplemented with additional methods which emerged during the study.
In a conventional ethnography, participant observation involves observing and participating in the activities of cultural members in their natural setting over an extended period of time. Observation can involve attending to the setting, the people who are present and how they interact, the events which take place, and the conversations which are exchanged. In a physical context, one might have a sense where attention should be drawn to observe those elements. With its different spatial configuration and distorted sense of time, Twitter obliges the researcher to think differently. Of course participant observation is not only about being present and merely observing; the act of participating enables a better understanding of the phenomenon. How one actually goes about participant observation needs careful explanation; there are those who claim participant observation on Twitter is not possible (Markham, 2013). In this earlier post I describe some of the aspects to which I might attend and the ways in which observing online is notably different from offline. In a subsequent post, I then set out in more detail the daily routines that this entails, including different points of entry, following interesting trails, recording field notes and how sometimes, interesting leads would actually be brought to my attention. I finally attempted to summarise my activity through the following visualisation which I discussed in this post.
In ethnographic observation it is usual to make mental or jot notes if appropriate, then subsequently work them up into more detailed field notes. Given how ‘sticky’ information is on the web, my initial notes were only brief, but included a link back to the original data. I later expanded these brief notes with more substantial observations and further reflections using MindView, a concept mapping application which I describe here.
I opened dialogue with participants in three different ways: in-depth, semi-structured interviews; ethnographic interviews conducted during observations; and asynchronous interviews using Voxer. In addition, I also briefly interviewed several blog post authors which I’ll discuss in a later section.
I invited participants for the semi-structured interviews where they had written a blog post or tweet which discussed their use of Twitter for professional learning. In some cases however, participants volunteered themselves. Given that most of them were distant, some even from different continents, interviews were conducted with Skype. Each interview was recorded, transcribed then offered back to the interviewee for comment? Where they gave their permission, and in keeping with the ethical stance taken in this study, recordings of the interviews were posted to the Edutalk podcast channel.
Following suggestions from participants, I trialled the use of Voxer, a ‘walkie-talkie’ app which some have described as an audio version of Twitter. It enables threaded, asynchronous audio conversations to take place for those who are subscribed. This was rather late in the study, and the outcomes were mixed. I didn’t press this hard through all the social media channels I had available because at this point, I was not short of data. As a consequence, the response rate was low and I only received responses from two people. It did however mean that one participant could become involved who had expressed a desire to participate, but had struggled to find the long slot of time needed for an extended interview. Voxer may therefore have potential in allowing some to become involved in research who might otherwise be unable. Since Voxer allows multiple people to be involved in an exchange, it might also be a different way of facilitating focus groups between remote participants.
‘Ethnographic’ interviews are spontaneous and respond to issues as they arise naturally. They are typified by their brevity, usually being of much shorter duration than other forms of interview, both in time and number of exchanges. It is one thing to observe activity during an ethnographic session, but switching role to that of engaged participant requires declaring one’s status as a researcher. <#ethics> How to declare one’s status and ask a question within the character limit is another matter. My solution was to include the hashtag #4MyResearch in any opening question I posed.
Conducting interviews in the ways just described runs the risk of once more privileging the humans, assuming it is only they who have things to say. It is easy to slip into the dichotomous world of humans and nonhumans as distinct entities and thereby attend solely to those who ‘speak.’ Consider the ‘Like’ button in Twitter. As a nonhuman, how might it be interviewed and allowed to ‘speak?’ What would it have to say? How can I report that?
The strategy I adopted to ‘interview’ the nonhumans was informed by a technique developed by Adams and Thompson (2016). I discussed this at length in an earlier post.
Blogs have sometimes been used in research as a form of diarying, however, rather than in this solicited form, unsolicited blogs offer access to publicly available, naturalistic data though one must be wary of the apparently ‘public’ nature of online content, including blogs <#ethics>. For this study, I only included blog posts discussing use of Twitter for professional learning and where the identity of the author was ostensibly clear; this amounted to twenty-four posts in all. Where commenting was enabled by the blog author, I took that as an invitation to discuss the post, and entered a discussion with the eight authors who responded. These were often not particularly lengthy exchanges; without visual cues, it is more difficult to read whether the interviewee is becoming tired of the exchange, so I never pushed for too long <#ethics>.
One evening, sitting on the sofa and scrolling through my Twitterstream, I was struck by the offline experience in which I was immersed, and how this was likely to be very different to that of other Twitter users. What kind of device were they using and what application? Where were they located? What time of day was it? Were they devoting their undivided attention to Twitter, or was it something to fit in alongside other activities? Given that they are likely to be using a device with a built-in microphone, perhaps they could narrate and record what they were doing? I tried this for myself and produced an example recording which I then posted to my research blog, together with a Guide, a Participant Information Sheet and link to a consent form.
Unaware that a similar technique called ‘thinking aloud’ was already established, I called them ‘audio arcs.’ Unfortunately no-one responded to the requests I put out through Twitter to take up this challenge, although once more, I didn’t apply too much pressure for the same reason mentioned earlier. I did however mention it during one of the interviews and the interviewee indicated that they simply didn’t use Twitter in that way. For them it was a series of brief encounters, rather than an extended visit and that they couldn’t think of a way to capture what they were doing when it involved irregular and fleeting activity.
During observation sessions one of the ways I bookmarked tweets to which I might wish to return later was by ‘Favoriting’ them. And then Twitter changed the ‘Favorite’ to ‘Like.’ This prompted me to think more deeply about how Liking might be perceived by the author of the Liked tweet and subsequently to change my strategy for bookmarking tweets. I discuss this and my solution at length in this post, but the upshot was that I assembled and deployed a Twitter bot which automatically Liked and retweeted tweets which satisfied particular conditions. The PLDbot sprang into action with tweets which contained the terms professional AND learning (OR development) AND twitter. This presented certain practical and ethical issues which I discussed in the aforementioned post.
Once again, an alternative method emerged in response to unexpected circumstances, further illustrating the importance of the adaptive and responsive approach encapsulated within flânography.
In the second part of Chapter 4, I’ll go on to discuss ethics within the study, how I managed and analysed data, and how I addressed what might usually be called rigour.
Adams, C., & Thompson, T. (2016). Researching a posthuman world: Interviews with digital objects. Basingstoke: Basingstoke : Palgrave Macmillan.