3 Research Methodology
3.4 Data Collection
3.4.1 Semi-Structured Interviews
Bryman and Bell (2015) notes that during the selection process of a data collection method, it is essential to first determine the type of information required for the study. For multiple case studies or any other type of study with a defined and clear topic to further explore, it is advised to adopt interviewing. Interviewing enables the researcher to build an in-depth understanding of the understudied topic and ensures comparability between the conducted interviews (Bryman & Bell, 2015). It allows the researcher to find out what questions are relevant or irrelevant, as well as expose patterns and connections that were not previously recognized in the literature (Dunn, 2005). Interviewing is mostly suited for this research,
especially because it makes comparability possible and reinforces the generalization of the population of the service design practitioners.
Semi-structured interviewing is used for this study. It is considered as a more suitable method since it maintains a balance between both structured and unstructured interviews. This method will give the interviewees the freedom and flexibility to respond and provide deeper and richer answers when compared to structured interviews, on the one hand (Bryman & Bell, 2015). While, on the other hand, it is still more controlled in comparison to unstructured interviews, meaning that it ensures that the researcher remains on track towards finding answers to the research question in a more efficient way (Bryman & Bell, 2015).
Thus, semi-structured interviewing serves as a more favourable choice for this study, not only because it has the possibility to obtain answers with specific themes, but also offers a high level of flexibility in the data collection. An interview guide is developed for this reason, which can be found in Appendix B.
However, semi-structured interviews have some disadvantages in comparison to other data collection methods. Critics argue that the use of such a method could potentially deviate from the constructed interview guide leading to loss of focus on the research topic during discussions. This, in turn, will decrease the comparability between the interviews (Bryman and Bell, 2013). To overcome this issue, the researchers firstly ensure not to include many questions in the interview guide, in order to give some space for flexibility to discuss more thoroughly about the responses connected to the themes. Furthermore, the researchers managed to return to the themes when needed during the interview, in order to ensure that all questions are answered.
The interview guide is sent to the interviewees prior to conducting the interviews to have a brief idea about what is being investigated and what questions are required to be answered. The interview guide mainly uses the formulated research questions and the relevant themes derived from literature review as a framework.
This is supported by Bryman and Bell’s (2015) recommendation regarding the interview guide development, which suggests relying on the theoretical findings when determining the themes in order to ensure that the research question will be answered. Furthermore, the questions are formulated in an open- ended manner to allow for more detailed answers that are rich with insights regarding the different themes.
This way will allow the interviewees to answer more freely about the subject, in order to ensure that all the different types of ideas and answers are captured and avoid the possibility of missing any data that can be valuable for this study.
This is also in line with Bryman and Bell’ (2015) advice in regards to question formulation in an interview guide, where it suggests that such questions should not be very specific or influenced by the researcher’s preconceptions about the research area, as this could potentially restrict the explorative aspect of the qualitative research. Interviewing is used here as the data collection method of choice. Interviewing enables the researchers to explore subjective meanings and reveal relationships in areas that are not thoroughly researched (Valentine, 2013a). It helps the researcher to discover what are relevant or irrelevant questions and reveals connections that were not identified in the literature. In particular, semi-structured interviews are chosen instead of structured and unstructured interviews. This interview style leaves flexibility for the participants to explore topics they consider important, while offering a structure that enables comparison across cases (Longhurst, 2013).
The average interview took 45 minutes, and they were conducted not only in presence but also over phone and video calls according to convenience and geographical constraints. Interview guides were developed based on recommendations from Dunn (2005) and Longhurst (2013). Developing the interview guide was regarded not as a linear process, rather, the questions and phrasing evolved based on past experiences. The questions were structured along various themes that came forward in the theoretical framework, guided with an introduction and conclusion. To ensure a conversational tone, follow-up questions were determined by the dialogue rather than the interview guide. Respondents were given the opportunity to talk freely, and interview questions were linked to the current topic of the conversation. At the end it was made sure that every question was covered.
3.4.2 Data Analysis
The use of thematic analysis to structure and analyse the derived findings from a multiple case study with multiple participants is well supported by Nowell et al. (2017), who claims that the thematic analysis is mostly valuable for the examination and comparison of the perception and contexts of different participants involved in a study, which could potentially produce new and unanticipated insights. This method is also useful for the identification of key components in a large set of data and reinforced by a well-structured approach to handle the data, which allows the researcher to generate a clear and organized outcome (Nowell et al., 2017).
In this method, the use of coding helps the research to convert unstructured data into structured one, as well as labelling the parts that are related to the theory or considered practically important for the study (Nowell et al, 2017, Bryman and Bell, 2015). The last stage of this method involves the process of organizing the codes into overarching themes which represents the key elements of the research question (ibid.). Despite that the thematic analysis is considered as an effective method for data analysis, it is argued that there are some disadvantages in regards to coding the data, in the sense that it leads to the de-contextualization and fragmentation of the material. This implies that the context of the material and the flow of the narration could potentially be lost. However, since this method is generally approved and considered important among researchers, coding is still advised as the initial step to take for a qualitative analysis (Bryman and Bell, 2013). The adoption of this approach enabled the researcher to structure the plenitude of gathered data into themes, making the data analysis process simplified.
3.5 Anonymization
For the research, a decision has been made previously to make all the interviews anonymous. The purpose for doing so is to allow the respondents to a larger extent and in more depth discuss and share their perspective about how service design application can contribute to the practice of entrepreneurial startups, whilst not jeopardizing that the gathered information will be negative or in any way harmful for the respondents. This is likely to give a more honest and transparent picture of the reality. Bryman and Bell (2011) explains how The Academy of Management (AoM) Code of Ethical Conduct demands the
researcher to tread carefully and reduce the potential harm that could be caused to the respondents. It is thus crucial to ensure that the empirical data that is obtained in no way can be traced back to the respondent nor the employing agency/firm, and information that can thus has to be discarded from the study. Refining the empirical findings and deleting associative information can be difficult, but still doable according to Bryman and Bell (2017).
In order to secure the anonymity of the respondents, each service designer has been given a codename.
Furthermore, information related to gender, age and design agencies' names have been discarded to assure total anonymity. By doing so, the hope is that the respondents have provided thorough answers to the interview questions addressed to them, not factoring in that their opinions can harm them, hopefully leading to less self-censorship. From the interviews, the transcriptions also have been subject to anonymization, in which all information that could be specific or directly related to a service designer have been withdrawn.