App-based textual interviews: interacting with younger generations in a digitalized social reality

ABSTRACT

Textual and visual app-based communication is standard everyday interaction. I argue that, as researchers, we have to be better equipped to include various ways of interacting digitally with participants in qualitative research interviews. In this paper, I do a methodological focused analysis of app-based textual interviews (n = 98) with young people selling illegal drugs online. I challenge the earlier skepticism towards interviewing through text, while also placing the interview in a new context of encrypted mobile phone applications. App-based textual interviews prove to be highly flexible and create a socially informal character of semi-structured interviews that interviewees perceive as safe. At the same time, the interviewer’s position is decentered as the interviewee takes full control of the interview context behind the screen. App-based textual interviews proved highly useful when interacting with a younger generation, also on a sensitive topic like illegal drugs. It also introduces several ethical dilemmas for discussion.

KEYWORDS:

• Qualitative
• interview
• digital
• ethnography
• online

Introduction

A majority of everyday-interaction takes place through the myriad of mobile phone communication applications that exist. Whether textually, visually, or in other forms, apps have created ways to communicate that overcomes both space and time. Regardless of continuous development of various digital communication tools, these have yet to become common platforms for doing qualitative research interviews. This despite the fact that it has always been important for social research to make connections to new technologies within its field (Fielding, 2014; Hinchcliffe & Gavin, 2009) and to reflect the day-to-day experiences of participants (Jones & Woolley, 2015; Taylor, 2002). Not using digital communication tools is in many cases a missed opportunity to reach out to people through their ‘go-to’ communication form, especially younger generations and others heavily embedded in the digitalized society.

For a long time, research literature on interview methods focused mostly on face-to-face interviews, underscoring that the non-verbal language was crucial in understanding what the interviewee says and keeping the conversation flow going (Bryman, 2016; Gubrium & Holstein, 2002; Kvale & Brinkmann, 2009; Rubin & Rubin, 2011). Despite how advantageous it might be to meet people face-to-face, this is not always possible. The latest worldwide pandemic with COVID-19 also shows that certain situations require social distancing that forces researchers had to move to online tools for communication with research participants (Nind et al., 2021). Most of the literature on interview methods and technology brings together both face-to-face interaction and the use of digital tools by taking advantage of video interviews, such as through Skype and Zoom (e.g. Deakin & Wakefield, 2014; Gray et al., 2020; Janghorban et al., 2014; Lawrence, 2022; Oliffe et al., 2021), especially after the COVID-19 pandemic. However, a big part of app-based communication happens textually. It is also asynchronous in its form.

The few contributions to textual online interviewing commonly discuss the use of e-mails (asynchronous), or chat forums (synchronous) (see, e.g. Barratt, 2012; Cook, 2012; Dahlin, 2021; Enochsson, 2011; Hinkes, 2021; Saarijärvi & Bratt, 2021; Shapka et al., 2016). However, the methodological research literature on the use of online platforms for app-based qualitative interviewing is still in its very early days. Only more recently has the use of mobile phone applications gotten some attention (see, e.g. Aligato et al., 2021; Gibson, 2020; Manji et al., 2021). In order for researchers conducting qualitative interviews to be able to grasp a digitally embedded reality, these methods also need to be developed accordingly (Davidson et al., 2016; Hine, 2004, 2015; Liamputtong, 2010). Especially with the new interview context of apps that, differently from earlier online and/or textual interviews, includes both aspects of asynchronous and synchronous conversation.

In this article, I discuss the advantages and challenges met when using a textual-based messaging application to do semi-structured interviews, and examples of possible solutions. The aim is to present a number of methodological and ethical take-home points concerning the use of text and apps for interviewing that may prove useful across populations and applications. The study in focus used the encrypted messaging application Wickr¹ to interview young people selling and buying illegal drugs through apps (see, Bakken & Demant, 2019; Demant et al., 2019). I use the study as an example of how researchers interview young people on the same digital communication platforms that they used on a daily basis. It is also an example of using technology to interview hard-to-reach populations on taboo topics, where factors like anonymity and encryption are important. The analysis concerns three main topics: 1) the role of technology within the interview situation, 2) the decentering of the interviewer due to ‘high context flexibility’, and 3) the role of emotions and bodily reactions, also in a textual interviews. These suggest various ways in which the use of mobile phone applications changes the interview context, both by communicating through text and by creating a new interview context.

Using technology for qualitative interviews

General advances in the society continuously reshape social research methods, and the development of technology is no exception. The main topics concerning digital interviews within the literature are the development of computer-based aids to record, transcribe or analyze interviews, as well as the use of video or audio tools to interview on a distance (Chen & Hinton, 1999; Edwards & Holland, 2020; Fielding, 2014; Palys & Atchison, 2012; Paulus et al., 2017). Early use of textual-based applications or digital tools rather warned against its use; Palys and Atchison (2012), for example, saw it as challenging and time-consuming. They argued that it could be difficult to establish a clear sense of trust and rapport, and that it was difficult to achieve a natural conversation that produced rich data. This perception of text-based digital interviews added to the previous research that questioned whether it was possible to develop good rapport with participants who you have never seen or heard (Bryman, 2016; Elmholdt, 2006; Mann & Stewart, 2000). Since then, communication within the general society has been increasingly digitised, not the least during the COVID-19 pandemic forcing people to converse online (Thorndahl & Frandsen, 2020). This led to a massive increase in the use of online platforms for research practice, especially video conferencing services (e.g. Gray et al., 2020; Lawrence, 2022; Oliffe et al., 2021). Also, digital technologies and Internet websites are increasingly being successfully used to recruit participants such as hard-to-reach populations and millennials (Archer-Kuhn et al., 2021; Chenane & Hammond, 2021; Dalessandro, 2018).

Digital interviews, both text-based and via video/audio, are in many ways different from in-person interviews (Gibson, 2020; Shapka et al., 2016; Thunberg & Arnell, 2021). On the positive and practical side, such interviews have minimal costs, few limitations of time and space, and greater anonymity and distance between the interviewer and interviewee making it attractive for use with vulnerable populations and for asking sensitive questions (e.g. Bargh et al., 2002; Barratt, 2012; Bryman, 2016; Cook, 2012; Davis et al., 2004; Deakin & Wakefield, 2014). Ethically, it is claimed that digital interviewing has an increased possibility of recruiting independently from traditional gatekeepers since it is easier to reach out to people (Janghorban et al., 2014; Marland & Esselment, 2019) and that such interviewing might reduce the interview effect as most visible bodily clues are non-existent (Hinchcliffe & Gavin, 2009; Manderson et al., 2006; O’Connor et al., 2008). Digital interviewing has also been shown to protect the interviewees by making it easier to withdraw from the interview and to have a choice about when and where the interview takes place (Deakin & Wakefield, 2014; Pearce et al., 2014). Researchers also argue that the visibility of the conversation encourages data that are more accurate by producing a mutual confirmation of meaning and understanding (Hinchcliffe & Gavin, 2009; Stieger & Göritz, 2006).

Another common topic within the subject of digital interviewing is the divide between synchronous and asynchronous conversations (Salmons, 2012; Schiek & Ullrich, 2017). Synchronous communication takes place using an immediate exchange of words and is often characterized by language, spoken and written (Baym, 2015; Schiek & Ullrich, 2017). Asynchronous communication includes e-mail interviews that takes place over time and in which the interviewee and the interviewer are not necessarily sitting ready to answer at the exact moment they receive a message. Both the flexibility of time and space give the participants some time to think and alternate their questions, which may be challenging, but may also allow for more reflective responses and help the interviewer handle surprises (Madge & O’Connor, 2004; Sands & Krumer-Nevo, 2006; Schiek & Ullrich, 2017). Authors of previous literature look at this as an either-or scenario, in which the communication is either synchronous or asynchronous, both of which have their advantages and disadvantages. However, conversing through mobile applications is, rather, an example of being both synchronous and asynchronous at the same time and is a mix of having both oral and written language elements. Understanding these different nuances in the use of new technology is important to the interests of both the research results and ethical concerns.

Researchers also discuss other challenging aspects of digital interviewing, such as overall topics like designing the interview, building rapport, and handling ethical issues digitally (Bryman, 2016; Dahlin, 2021; O’Connor et al., 2008). For example, how to handle the possible disruptive environments of digital (video/audio) interviews that take place at home or at work (Deakin & Wakefield, 2014). Another discussion is how to gain informed consent in a digital environment where the uncertainty is said to be higher due to factors like lack of signals, silence, and the blurry lines between public and private space (James & Busher, 2009; Paulus et al., 2017). Researchers should also be attentive to the design of the information technology, as it might influence both the communication and the relationship by setting boundaries and forming the layout of the conversation (Edwards et al., 2013; Kaufmann & Tzanetakis, 2020; Paulus et al., 2017; Ruppert et al., 2013). A number of recent publications challenge the earlier established characteristics of digital interviewing as being disembodied and lacking emotions (e.g. Bryman, 2016). Synchronous video communication, such as Skype and Zoom, have proven to be personal, emotional, and intimate despite being conducted online (Baym, 2015; Deakin & Wakefield, 2014; Janghorban et al., 2014; Lawrence, 2022; Oliffe et al., 2021). The latest research also shows that asynchronous interviewing on e-mails or apps can be more natural than once claimed (Dahlin, 2021; Gibson, 2020; Lawrence, 2022). Despite the various literature on the topic of digital and online interviewing, the use of textual mobile phone applications needs to be further explored in research literature to not leave a significant gap between the real world and the ability to research it.

The case of digitally embedded youth engaging in deviant activity online

This paper is a result of reflecting back on experiences with using a text-based application for interviewing young people. The interviews for discussion are drawn from Nordic Drug Dealing on Social Media (NDDSM) – a project on the use of digital communication technology to deal and buy illegal drugs in the Nordic countries. The study lasted from June 2017 to December 2018. We first got familiar with the field of social media drug dealing by observing and exploring on various social media platforms (Demant et al., 2019). We then used our observational insights to recruit interview participants active in buying and/or selling illegal drugs using digital tools like social media or darknet. Local students and the project coordinator (author of this article) conducted 106 interviews across five countries, whereof 98 were conducted on the encrypted text-based application Wickr. The participants were already familiar with Wickr and perceived it as a safe platform to talk about both their illegal and legal activities. The participant group was relatively young, with an average age of 23.1 (ranging from 16 to 45 years old).

We contacted several of the interviewees directly on Wickr, as it was common for them to list their Wickr ID in their profile or posts. By initiating the conversation on a secure platform that the interviewees already considered a safe-zone, we hoped to have more success when recruiting participants and to encourage openness concerning their illegal actions. This was also a platform that ensured participant anonymity and researcher protection, which made it suitable for this particular study. We offered the participants a choice to have the interview in other contexts, such as on Skype, by phone, face-to-face, or on non-encrypted communication platforms. Still, 92.5% of them chose to remain on Wickr, or to move on to Wickr if contacted on other platforms. They all used a variety of technologies in their everyday lives and expressed to be relatively comfortable with being interviewed on an encrypted, text-based phone application.

All interviews were made on project-owned smart phones to not involve any personal information of the students and to protect them from possible unwanted contact. We informed the interviewees about the study and their option to opt out at any point before starting the interviews, and asked them to confirm that they agreed with participating. They were also given the full name and affiliation of the project coordinator and project leader in case of any questions or further information. None of the project members knew any of the real identities of the interviewees, as the contact was run through anonymous accounts. At the time of the interviews, Wickr allowed screen shots to be taken of the conversation, which made it easier to fully engage in the conversation when it took place and then to transcribe the interviews at a later point, before deleting the images. Many transcripts include additional information about the interview process, such as interviewer’s notes, time of each message sent, time estimations from first contact, and images or symbols, such as emojis.

To discuss the use of app-based textual interviews, I use the transcripts of the 98 Wickr interviews as well as documented discussions held between the project participants in Slack.² In Slack, the students could reach out to the coordinator and other interviewers with questions or concerns during the data collection, and get immediate response. All interviewees had courses in qualitative methods for social sciences, including interview techniques, however, none of them had any training in using mobile applications for interviewing, since this, back in 2017, was not that known. Slack became a platform for continuous methodological discussions throughout the project. This included separate discussion threads for each country and two general discussions as well as one-on-one conversations with the coordinator. We also arranged a method-related workshop debriefing the data collection with all the involved researchers, where the author of this paper and the study principal investigator led discussions.

Due to the author’s participation in both the data collection and the meta-analysis of it, ‘I’ and ‘we’ will both be used throughout the article. ‘I’ refers to the author, used in such contexts as when reflecting on the method or the analysis, while ‘we’ refers to the research group or data collectors in general, which also includes the author. All the quotes used in this article have been securely anonymized to protect the identities of the interviewee as well as the specific geographic area, such as the city. Quotes will only refer to the position of the interviewee (drug seller or buyer) and the country, and all screen names or other personal information are removed.

A technology-led interview context and its challenges

Conducting interviews through a mobile phone application created an interview situation with almost no spatial or temporal limits. To participate, and to conduct the interview, the interview participants used their smartphones – bringing the interview with them wherever they went; being alerted at whatever time of the day. This created a new interview context that was both depended on the online and offline setting of the interview. Here, I will present some of the main challenges we met by focusing on the impact of the technology in use, the increased flexibility of the interview situations, and how emotions and bodily reactions remained an important factor throughout the written conversation. These three analytical topics are presented through different scenarios faced during data collection, as well as examples of how we handled them, both successfully and unsuccessfully.

The role of technology

Being familiar with the target group’s communication preferences and safety considerations can be quite useful to researchers (Tiidenberg & Cruz, 2015). In our case, we knew Wickr was a popular application among young people engaged in deviant activities, as the drug sellers we observed during ethnography often use their Wickr IDs in ads that enable buyers to contact them. Wickr was popular among members of this population due to its encryption. It was therefore considered safer than phone numbers and other non-encrypted applications, as expressed by this Norwegian seller:

The first time I heard about digital communication was when I was little;) After a while, I learned that encryption was something necessary to secure communication. (…) Since society chooses to battle against the use of these medicines [MDMA and psychedelia]—because that’s what it is: medicine—it is totally natural to choose a secure communication channel like Wickr to communicate. (…)

They often supported these security claims by referring to discussions on Reddit or other drug-related discussion sites where users express their opinions. Thus, because we hoped to encourage the interviewees to open up about their world, Wickr became a natural interview setting (Kvale & Brinkmann, 2009). Meeting them on their preferred platform demanded great research flexibility and willingness to learn other technological means but had the advantage of establishing a certain foundation of trust early in the interviews.

The use of a text-based app like Wickr for interviewing affected several factors of the interview, one being the practicality of typing one’s answers. Some applications might have limited number of characters within each message. However, even without such limitations, like on Wickr, the answers tend to be short and concise. This influences the conversation flow, which is as important in text-based interviews as when talking face-to-face. Achieving a flow depends both on the interviewer and on the interviewee and the way in which they adjust their communication (Roulston, 2014). A great deal of flow comes from managing rhythms and patterns of communication, which can be especially challenging within text-based interviews without any visual contact. Being familiar with app-based communication forms and the way in which a fruitful interview may consist of short answers is an advantage. An example of this is evident in an interview with a Swedish drug buyer in which the questioning and answering process flows quite well and the responses are short but informative:

Interviewer:

Can you tell me a bit about that first buy?

Interviewee:

I bought 100 ecstasy of a guy in a park in Gothenburg. It felt really unsafe.

Interviewer:

How did you come into contact?

Interviewee:

He announced it in a Facebook group and posted his Wickr.

Interviewer:

How did it feel to meet him in person?

Interviewee:

He stood with a group. He gave me information about his looks, and then I asked him

if it was him and got the goods, and he got what he wanted.

Interviewer:

Okay. How did you feel before you met him?

Interviewee:

Quite all right.

Interviewee:

It wasn’t worse than a (a famous big roller coaster in Sweden) ride.

This occurred early in an interview in which the conversation flowed from the first few questions. But there is no doubt that earlier experience with such communication form of sending limited messages among both the interviewee and the interviewer made the conversation flow better than if one of them expected longer answers. Although some research traditions are more interested in the stories people tell than the details (e.g. Riessman, 1993), in our project, we aimed to describe drug dealing on social media, so the information itself was important. We therefore appreciated short and concise answers to some of our questions and rather asked for further explanation when needed. It might be discussed if the text-based interviews limits the possibility to get thick and detailed answers, however, this comes down to the interviewee and their willingness to write extensively.

The overall dynamic of app-based textual conversation also changed the rules of timing. An important part of any conversation is being an active listener by using body language and confirmative sounds (Roulston et al., 2003). Emojis or smileys and textual comments such as ‘Interesting’ or ‘I see’ replaced these components. However, for example, reacting to a message, even with a confirmative ‘Aha,’ usually stopped the recipient from answering further or confused the conversation partner with an incorrectly timed answer. Timing answers carefully became a balancing act that we had to adjust for each interviewee, depending on their conversational rhythms. For example, some interviewees gave short answers in multiple messages, such as in this example:

Interviewee:

They’re a real cannabis activist nowadays, too.

Interviewee:

They’ve seen how much cannabis has helped me in my condition.

Interviewee:

And of course in hundreds of other patients, too.

Others sent longer replies and took more time but still replies into separate messages. This made timing follow-up questions and comments challenging. Developing an awareness of such necessary adjustments may demand more reflection than in face-to-face interviews, as textual communication is a relatively new way of communicating that is constantly changing. It also left the interviewer less in control and having to adjust to the new position in the digital context. Just as in face-to-face interviews, the interviewers regulated their responses and response time based on the interviewees’ rhythm to make the conversation seem more comfortable.

Decentering the interviewer

An additional factor to the use of technology when interviewing was the uncontrollable external factors that could affect the interview negatively. In such situations, the interviewer became decentered within the interview by leaving much more up to the interviewees, who now fully controlled the physical interview context on their side. The preferred context would have the interviewee feel safe and encouraged to answer the questions, and with minimal unwanted or non-ethical disturbances. Many interviews took place late in the evening and on weekends, and we did meet some surprises. At one point, an interviewee suddenly replied, ‘Can we continue this tomorrow? I am too drunk now and probably should not be answering.’ There were no indications of the interviewee’s intoxication until he mentioned it, which could pose ethical challenges. The use of asynchronous communications gave us the flexibility of time and space that made it possible for the interviewer to reach out to the project coordinator or leader for advice on how to handle such surprises. In the case of the drunk interviewee, the interviewer (being the coordinator) quickly ended the interview and got back in touch the day after. At other times, the interviewees may have been in situations that restricted their answers, or friends or others may have influenced them, without us knowing. Being somewhat familiar with the everyday life of the interviewee was an advantage, such as knowing if they worked regular jobs during the day and answered us in the evening. In our case, it was also important to consider their relation to buying and selling illegal drugs, which for some was a recreational activity and for others a full-time job.

In several ways, using such text-based technologies like Wickr removes power from the interviewer and to the interviewee. In addition to the already mentioned physical context of the interviewee, the informality of conversing at any point of the day challenged the way researchers communicated with the interviewees to retrieve as much information on the topic as possible, as they were always on the job. Continuing the conversation is important in any interview format but knowing how to do so through text was challenging, especially given that mobile phone interviews can be both synchronous and asynchronous. Online text-based conversations can involve an immediate exchanges of messages, but they may also take place over longer periods without interrupting the conversation. This is like silences in face-to-face interviews (Bengtsson & Fynbo, 2018). Researchers want to gather as much information as possible and obtain answers to all their questions, but they should also respect silence. The aim of completing all the interviews led the researchers to answer as quickly as possible to keep the interviewees’ attention alive. However, an immediate answer from the interviewers could break the conversation’s fluidity because /they seem too eager. Listening to the rhythm of the conversation on the part of the interviewees was therefore important, even though this meant letting them remain silent for a while.

Being adaptive and flexible proved quite useful for maintaining the flow of the text-based application interviews. However, some interviews also became too flexible. In the first interviews, the interviewers asked questions immediately after the participant agreed. They also told interviewees to answer when they had time to avoid putting pressure on them. This resulted in long waits for the interviewers, sometimes weeks or months, and the expected 1–1.5 hour-long interviews ended up taking between 0.5 hours and 2 months to complete. This made it necessary for the interviewers to keep track on the conversations by taking notes or watch the screen shots, so that same questions would not be asked several times, and that they remembered the earlier information provided by the interviewee. However, it was challenging waiting for a reply, not knowing whether the interviewee stopped answering at all or if they just got distracted for a period of time. One interviewer ended up sending multiple reminders:

Interviewee:

Sorry that it took so long to answer. *laughing emoji with a drop of sweat*

Interviewer:

No problem. *smiley* Could you tell me a bit about the first buying and selling? Did

you prepare in any way, take any precautions, or something like that … ?

–Two-day break –

Interviewer:

Hi! Are you still interested in participating in the research? *smiley* If not, that’s okay, too!

Interviewee:

Hi! Yeah, sorry, I totally forgot about this.

Additionally, the increased flexibility also led to interviewees answering at all times. This occurred in the morning as often as in the afternoon, but it occurred most often late at night. Like most other people, the interviewees were available after work, school, or afternoon activities such as going to the gym. This did not necessarily work best for the interviewers, who sometimes felt the interviews invaded their lives. Fortunately, the interviewers took necessary breaks, such as when eating dinner:

Interviewee:

I found a girlfriend 18 months ago.

Interviewer:

Nice to hear! I’m just going to eat some dinner. Then I’ll get back to you. Is that fine?

Interviewee:

Sure.

Balancing the interviewers’ and interviewees’ needs and boundaries was challenging, and the interview process definitely invaded the interviewers’ lives much more than in face-to-face interviews. Conducting interviews digitally or online means the interview can be continued, which resembles immersive fieldwork in which the researcher remains in continuous fieldwork over long period is in the situation.

One advantage of recruiting and interviewing through mobile applications was that interviews often started immediately after the participants agreed to take part in the study. However, after a while, the number of participants increased, so the research team decided to begin scheduling interviews in order to regain some control. Scheduling a time to continue a conversation was not a common way of managing this particular social space and led to some challenges. In the following Slack discussion, two interviewers describe their progress:

Interviewer 1:

Hi. It is going very slowly with the interviews. The few people who have agreed to

participate have been very difficult to get in touch with at the time of the scheduled interview.

Interviewer 2:

Same here. I have asked more than 10 people during the last few days, and no one has

answered or wanted to participate.

Scheduling introduced more structure into the messaging period and led to challenges. As in face to face interviews, some participants did not ‘show up’ at the agreed times. When this happened, it meant a considerable break in what began as an ongoing conversations in which people connected and communicated through messaging applications. Had we begun all the interviews immediately (some were rescheduled due to practical reasons of interviewing many at the same time, as discussed), we might have recruited more participants because we had captured their interest in participating. When we had to reschedule, some participants’ interest seemed to fade, whereas others remained eager to participate and did not mind the rescheduling.

Embeddedness through textual emotions and bodily reactions

An important limitation in textual interviews is the lack of body language and emotional reactions (see, Bryman, 2016). However, the way people communicate through text on various mobile applications is close to a spoken language in the way they formulate their answers, and there are several ways of showing emotions and reactions to other people’s messages. Through emojis, even bodily-distant communication can contain smiles, tears, embarrassment, or laughter. This proved useful when establishing a connection and showing personality, as did written cues such as using ‘hahaha’ for laughter or using exclamation points to emphasize a reaction. The person actively revealed these reactions when choosing to write them, so they differed from facial expressions. However, they helped set the tone in the interview and shaped the conversation flow for the interviewee and the interviewer. For instance, in an excerpt from an interview with a Danish seller, both parties included reactions such as emojis within the conversation:

Interviewer:

Do you talk with [your customers] for a while before you meet them? To check them

out?

Interviewee:

References are good enough. ☺

Interviewer:

Ah, okay. Who is your typical buyer?

Interviewee:

All kinds of people. It is just like asking what kinds of people there are in Danish

society.

Interviewer:

Haha, good point. ☺ But that’s interesting. Is it mostly students or a similar age group?

Interviewee:

Yes and no. ☺

(Interviewee then continues explaining different types of buyers)

The interviewers’ use of emojis changed to adapt to the interviewees’ use, as a way of reflecting the interviewee’s language. If the interviewees used a lot of emojis the interviewers increased their use. On the other hand, if the interviewees used no emojis, the opposite was true. When using emojis, both parties must be more aware of their feelings and make active decisions about which emotions to show. Achieving empathy and obtaining valuable, informative, and comfortable interviews was much easier when the researchers could also show their personality and convey their mood. However, as in the example above, the use of emojis was not always straightforward, and it did not necessarily imply bodily reactions on the other side of the screen. The emojis did add an extra layer to the statements and set a certain tone to the conversation, like the smirk one can imagine the seller has in the quote above when answering ‘Yes and no. ☺’. It was all about the context and the conversation at a whole that gave meaning to the emojis, and in such an interview session the interviewer would ask about the specific meaning behind the use of an emoji if it created confusion.

In some interviews, early communication difficulties almost ruined the entire interview. Especially in our case with a hard-to-reach populations who are alert to any negative triggers, particularly among such online communities where the complexities of negotiating access depends on collaborative efforts from the involved human participants as well as technologies, cultural codes, and other factors (Kaufmann & Tzanetakis, 2020). For example, in the interview excerpt below, the interviewer’s use of emojis seemed to trigger a negative reaction:

Interviewee:

Hello. What is it exactly you think I can help you with? I neither buy nor sell any kind of drugs on social media, or any other places for that matter. By the way, types like ‘us’ prefer to be informed about how you came to possess my Wickr ID, so maybe we could clarify that as well. (…)

Interviewer:

I just thought you were connected to some of the groups where selling of drugs happens. The Internet is full of clues, so if you know where and how to look, you’ll find information rapidly. In this case, Wickr IDs. We don’t have a normative purpose with our research. Our assignment is just to uncover ‘truths’ and gain knowledge about the world. We are geeky academics associated with the university. I thought you might find it exciting to tell your story.

Interviewee:

So you choose to start out by being a smartass? Maybe you would achieve better results with a more humble approach instead. Then you might actually achieve what you want. (…)

The interviewer went on to excuse himself and to say that he did not mean to be ‘a smartass.’ The conversation then continued quite successfully, even though the rough start caused the interviewee to be skeptical. This incident underscored how important it was for interviewers to reflect on their language and use of emojis when interviewees cannot read any bodily clues or hear their voice and what they say. Misunderstandings can occur easily, especially when attempting to establish access and rapport with new people that you have little background information on, not least when communicating on textual-based phone applications.

Discussion and conclusion

In this study, conducting interviews using a mobile phone messaging application was a necessary and timely way of conversing with the study participants. Conversing through text proved to be commonplace for the interviewers and the interviewees, which made the conversations natural and led to several major methodological strengths. The interviews seemed like regular conversations with an easy flow, and emotional reactions became part of all interviews through the use of emojis and short words or expressions. In addition, interviewing synchronously and asynchronously also increased the interview situation’s flexibility, which enabled the interviewees to answer when they had time, even across long time distances. Most importantly, the communication form was one with which the interviewees were comfortable. Which in our case was an encrypted messaging application that provided extra security concerning the illegal activities in question. Familiarity with textual conversations and the use of emojis to express feelings made the interactions less static and more personal than earlier literature suggests.

However, interviewing through a text-based phone application changed the connections to space and time common in face-to-face interviews. The new factor influencing both the result of the interview and the interview context is the technology itself. It proved important carefully to deliberate which application to use, both in terms of what the interviewees were comfortable with and that the interviewer could handle, as well as other practicalities like security and character limitations. Any technology will influence the style of communication, which the researcher then needs to take into consideration. On one hand are the limitations or options set by the medium itself, such as the use of visual aspects, time limits, or character limitations. On the other hand are how the interviewees use the application, or mobile phone itself. The study discussed throughout this paper already included a selective group with known digital competencies, where we knew they were familiar with the application used. It was an extra concern for us that our interviewees felt safe enough to open up on their illegal activities. This was helped by establishing a good report with the interviewee, such as by showing emotions and engage in the answers given by the interview using emojis and written confirmative sounds. To avoid misunderstandings, we paid attention to the way in which the interviewee talked; both the way they constructed their sentences, but also how they used emojis and reacted to questions or comments. Conversing back and forth on a platform familiar to the interviewee sometimes made the interviews similar to any other conversation with friends or others.

In some sense, these app-based interviews are similar to holding conversations during ethnographic fieldwork and not as a more formal interview situation with set limitations of time and space. When using a mobile phone application for interviews, the interview situation itself embeds into the everyday life of the interviewees, as well as for the interviewer. In some way, the high flexibility of the interview situation made the interviewer lose much of his/her power, both to a desirable and less desirable extent. Instead of being in a particular situation, often in a space and at a time chosen or at least agreed to by the interviewer, app-based interviews take place anywhere and at any time within the everyday life routine of the interviewee, leaving it to the interviewee to choose when and where to engage in the interview. It is also up to the interviewer to choose whether to be fully flexible or to schedule a time.

Despite several benefits, the interview setting was, at times, so informal and so flexible that in some sense it resembled a continuous field visit, which presents various ethical issues. For example, the interviews took place whenever and wherever. Interviewing someone when they are at a party or surrounded by friends requires ethical considerations that we were unable to take account of because the interviews’ contexts remained hidden. This also raises questions as to whether the interviewees remained fully aware of the formal intent to collect data for a project or if they came to view the communication as a more personal conversation, especially when the interviews lasted several days or weeks. Interviewing through applications is a new interview setting not only for researchers, but also for participants. Future researchers need to build methods to protect interviewees in these regards, despite the possible consequence of not obtaining as many interested and active participants.

Despite the practical and ethical challenges of text-based and also app-based interviews, our experience shows that app-based textual interviewing is valid, and timely necessary. Interviewing through an app was a great way for reaching young people, as they were already familiar with such communication forms and comfortable with communicating that way. Being behind a screen also increases the interview context’s flexibility, which may make it more comfortable for young people. Other populations such as older people may find it challenging for now, but the digitization of society will continue and make it a more common form of communication. Meeting the interviewees in a context and on a platform where they are comfortable contributes significantly to successful interviews, and considering whether a less traditional place such as mobile phone applications might serve this purpose is important. This proved especially valuable in our focus on a hard-to-reach population of drug buyers and suppliers, where anonymity and safety is important. In our case, using an encrypted platform that the interviewees already used for talking about similar topics as within the interview proved extra useful. This platform was a way to reach out to the target group on their home ground, which made it easier for them to take part and to continue the conversation across time. The success, however, does depend on the aim of the study, as text-based conversations might not always offer the long, continuous answers that researchers sometimes look for, but rather a frequent exchange of interaction between the interviewer and the interviewee. This worked very well in our rather descriptive, exploratory study of a new phenomenon. However, it is important to remember that app-based textual interviews are in their early beginning and that it requires continuous reflection from the researcher on the benefits and challenges, as well as further discussion of ways to improve this method and related ethical dilemmas.

Acknowledgments

Thank you so much to all the local students for allowing me to use our conversations for this methodological discussion. I would also like to thank all the anonymous reviewers for constructive feedback, as well as to everyone who gave me comments on early drafts, especially my supervisor Jakob Demant, Mareile Kaufman, Margaretha Järvinen and fellow PhD students.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by The Nordic Research Council for Criminology under Grant [20700034].

Notes

1. Wickr is an end-to-end encrypted messaging application with a user-defined burn-on-read settings for you to decide what happens with your communication when sent. They promise that all user content is removed from the device after it expires, and no communication is stored on Wickr’s servers (as well as any unique device identifier). See wickr.com.

2. Slack is an application created for easier collaboration. One can create different work spaces and invite people to join different discussion threads. It is used to communicate through chat within the thread or one-on-one, as well as to share files. Within our project, each country had their own discussion thread where they could contact the project coordinator or principal investigator, as well as a general platform for all project members. See www.slack.com