• No results found

4. Methodology

4.4. Phase 3: Semi-Structured Interviews

The third phase of the study involved explanatory semi-structured interviews.

The purpose of these was to explore qualitatively the quantitative results from the survey and to obtain a richer understanding of how and why people are purchasing medicine online in order to address research questions 4 and 5 (what drives online medicine purchasing; how do people engaged in the

behaviour view their conduct). The interviews were informed by the preceding phases, and covered views of past and present medicine purchasing from the Web, behaviours towards the obtaining of drugs and medicines, and how decisions are informed and made.

A list of questions, which covered some fairly specific topics, was produced from the analysis of the survey data and used as a guide throughout the interviews. In semi-structured interviewing, the interview process is flexible and the emphasis is on the way the participant frames and understands issues and experiences (Bryman, 2007). Following Leidner (1993:238), a topic guide was used to provide a degree of structure but also to allow room to pursue topics of particular interest to the participant.

The answers provided in the survey assisted the creation of this interview guide, and interviews were tailored for different respondents. The questions differed depending on whether the participant had indicated that they had or had not bought medicine from the Web.

4.4.1. Interview Sampling

In accordance with Fitzpatrick and Boulton (1994), it was necessary to ensure that sampling contained the full range of possible perspectives so that the concepts and categories developed provided a comprehensive

conceptualisation of the subject. The survey included the option for participants to agree to being contacted in order to take part in follow-up interviews by providing their email address. They were sent consent forms and study information sheets, which were signed and returned prior to the

interview. Convenience sampling was used, as much depended on which of the respondents to the questionnaire study provided their details for a follow-up interview and then consented to being interviewed. Snowball sampling was also used, as I advertised on social media and via contacts.

Most of the interview participants were recruited through the online survey;

however, six participants were recruited via word of mouth and the MOOC, and contacted the researcher directly to be interviewed. A breakdown of the route to interview and route to the survey (if applicable) is provided in Figure 4 Consort flow diagram: route to survey

Figure 4 Consort flow diagram: route to survey

n = 4 Total

n = 14

Word of mouth – including 3 who did not fill in survey MOOC

(including n=2 who

did not fill in survey

n = 28

FB friend posted link on their page

n = 1 Completed

survey

Recruitment to interview

Patient.co.uk page Twitter (RTs)

Bluelight Forum

Don’t know (pre- MOOC) 1 did not

fill in survey n = 3

Route to survey

n = 3 n = 2

n = 1

Facebook

4.4.2. Interview Methods

A range of different interviewing methods was used; these are displayed in Figure 5 Consort flow diagram: method of interview.

Figure 5 Consort flow diagram: method of interview

Method

n = 4 Skype

video

Telephone Instant

Messenger Online

n = 4 n = 15 n = 1 n = 4

Face to face Offline

Email

n = 28 Total

This study mediated between the virtual field and the real field by combining online and offline interactions and communication in the different methods of interviewing employed. These included more traditional means such as face-to-face and telephone interviews, but also innovative techniques benefitting from contemporary technology like Instant Messenger (IM), Skype video messaging and email. Some online techniques were not dissimilar to their offline

counterparts. Face-to-face interviewing is comparable with Skype video interviews, as both the researcher and the participant can observe body language and respond accordingly, while IM, telephone interviews and email interviews do not benefit from a physical presence.

According to James and Busher (2009), the inclusion of both verbal and non-verbal cues such as facial expressions, gestures, postures and emotional mannerisms in face-to-face interviews enhance the social interaction. The immediacy of social presence taking place in ‘real-time’ is a pivotal part of the research relationship (James and Busher, 2009). This is also true for Skype video interviews, where the researcher and the interviewee benefit from being able to read the other’s body language. In contrast, telephone interviews have been criticised for their absence of visual cues (Garbett and McCormack, 2001).

Henson et al. (1978) suggested that although the face-to-face method might facilitate openness, participants might be subtly induced to untrue admissions, whilst telephone participants have been described as relaxed and willing to talk freely and disclose intimate information (Novick, 2008). I found that different methods offered different benefits. I expected the topic might

prevent some interviewees from revealing some details from their private lives or health experiences. However, this was not the case, as interviewees talked about how they had not told their friends and family about purchasing

medicine online for fear of negative reaction. The stranger “often receives the most surprising openness – confidences which sometimes have the character of a confessional and which would be carefully withheld from a more closely related person” (Simmel, 1950:404). I generally found that face-to-face

methods procured a larger amount of data than email interviews, which tended to encourage short, focused answers; however, I was able to return to the email conversations and resume questions if I felt that the data needed to be elaborated. This was not possible with the face-to-face interviews, which were

usually under time constraints. The majority of the email interviews were iterative processes, not confined to single conversations but continuing on over several days.

By using different interviewing methods, valuable data were collected from people who may otherwise have not been able to participate, and the techniques were tailored to their needs. IM and email interviewing avoided interview fatigue, as well as concerns about the safety of the researcher, and increased the geographical reach of the research.

The chapter now turns to the specific issues I faced during the interviews.

4.4.3. Temporal Issues in Asynchronous Interviewing

Asynchronous interviews such as IM or email interviews are not conducted in real-time, so participants are able to reread what they have previously written, reflect on and consider their responses and amend their text. In synchronous interviews, responses are spontaneous. There has been considerable debate about the reflexive nature of the online medium (Markham, 2004).

Asynchronous email discussions can allow for an extended and deliberate sequence of conversations, and enable researchers and participants to digest messages before replying (Kanayama, 2003). Kivits (2005) states that

participants have the time and space to refine their own thinking without intrusion by the visual presence of the researcher, which allows the development of a more thoughtful and personal form of communication.

Johnson (2011) claims that the semi-anonymity of online communications encourages people to self-disclose more than it hinders them from doing so.

Bowker and Tuffin (2004) also note that the ability of participants to reflect on their thoughts and reactions can be enhanced as a result of the intimacy furthered by the informality of typing.

I encouraged participants to review their previous correspondence and to revisit earlier topics in order to help the development of their thinking about other issues. I returned participants’ messages to them as part of the normal email exchange, so I kept to the thread of email messages rather than starting each email anew. I did not erase any messages from the exchange, in order to ensure that the participants and I were able to interrogate earlier texts as our dialogues developed. This proved useful during one occasion where I was

accused of repeating a question, yet inspection of the past exchange proved that this was not the case.

My attitude towards having to wait for responses from my participants changed over the course of the interviews, from trepidation and concern that the

participant had lost interest in the study to anticipation of the possibility of obtaining exciting information. As Russell and Bullock (1999:134) put it succinctly, “one of the beauties of e-mail is that you never know when you will get a response…”

4.4.4. Interview Data Collection

28 interviews were carried out over a period of four months, from November 2013 to the end of February 2014. Experiences of healthcare and purchasing medicine online formed the main focus of discussion. The average length of time for these interviews varied, as different methods were used. The face-to-face and Skype interviews typically lasted for an hour, while the IM interviews lasted between one and a half hours and three hours, and the email interviews took place over several days. Towards the end of my interviews it was clear that I was not generating any new information, and so it was felt that it would not be worth pursuing any more than 28.

Interviews were recorded using a tape recorder for face-to-face, telephone and Skype video interviews, and the text of the dialogue was automatically

generated in IM and email interviews.

Table 5 shows the interviewees by age, gender and their purchasing status.

Table 5 Breakdown of interview participants’ ages

Men Women All

Age range

Purchaser Non -purchaser

Not known

Purchaser Non – purchaser

18-24 1 1 2

25-34 1 2 3 2 8

35-44 1 1 3 2 7

45-54 1 3 1 5

55-64 1 1 1 2 5

65 and above

1 1

Sub-total 6 3 1 11 7 28

Total 10 18 28

The interviewees were predominantly White British and primarily based in the UK. The over-representation in the survey of the MOOC participants, who were more highly educated, meant that most of the interviewees had also attained a high standard of education. In order to make the survey more representative I did try to seek out participants from different social groups (i.e. across the range of employment options) and ethnic groups (although this information was collected I did not include it within my analyses because of limited data), however, I was limited with my convenience sampling and where I was able to advertise and distribute the survey online. Upon reflection, the online spaces that I used – social media and forums, ended up attracting particular

homogeonous groups.

4.4.5. Analysis of the Interviews

The analysis drew on Mason’s (1996) cross-sectional and categorical indexing.

This approach was used to obtain an overview of the data and generate

themes. Mason (1996:54) outlines three approaches: “literal, interpretive, and reflexive”. Literal indexing focuses on the exact use of particular language or grammatical structure. Interpretive involves making sense of research

participants' accounts. Finally, the reflexive approach attempts to focus attention on the researcher’s contribution to the data creation and analysis process.

As per Mason’s (1996) suggestion, in practice I used a combination of these approaches. To begin with, I organised the data by coding text and breaking it down into more manageable chunks. I created initial indexes (see Appendices) consisting of in vivo codes identified in the data, and my interpretive codes.

During this process I regrouped and revised codes, and organised them into lists to see the connections between them.

I sorted and grouped categories together and wrote an overarching description for each (akin to a memo in grounded theory). The data was then coded and input to the computer-assisted qualitative data analysis software package NVivo, which helped to construct the themes.