3.4 Research Design and Paradigm 49
3.4.2 Phase Two: Semi-‐Structured Interviews 53
Phase Two consisted of semi-‐structured, face-‐to-‐face interviews. The shortest interview lasted five minutes and two seconds and the longest, 25 minutes and
56 seconds, with a mean duration of 12 minutes and 39 seconds. Interviews were audio recorded and transcribed in verbatim. Participants had the right to decline to be voice recorded, with two participants exercising that right. The interviews were chosen to build upon the information recorded in the online observations. This approach elaborates on the methodology of Pan and Fesenmaier (2006) and used navigation paths as a basis for the direction of the interview and questioning regarding observed online behaviour (see Figure 3.0).
Participants were firstly asked to explain the factors they considered or that influenced their decision when purchasing airline tickets when either travelling to New Zealand for study or returning to their home country following study. This line of questioning was aimed at determining how participants believed they evaluated different airlines’ offerings and came to their final purchase decision. Following this, participants were asked to explain why they went to each location (e.g. web page or significant link clicked on) on the path to their final purchase decision in Phase One. The aim of this was to determine whether their purchase decision process was consistent with answers provided to previous questions, which asked participants the factors that they believed to influence their purchase decision. Participants were then asked to explain the specific reasons for their final choice of flights.
Asking participants the factors they considered and then asking them why they made certain decisions in the online simulation, was aimed at determining whether they utilised any underlying or subconscious criteria without being fully aware of it. Additionally, this method was used to see if there would be any contradiction between how participants perceived their own behaviour, how they explained their own behaviour and what their behaviour illustrated to the researcher.
In the final stage of the interview, participants were asked to rank what they believed were the four most influential/important factors in their final choice of airline tickets. Again this was to strengthen the understanding of the factors at the forefront of their evaluation process. They were then asked to briefly explain
the reasons for the order of factors they had given. This was intended to extract reasons for the prioritisation or valuation of particular factors.
Semi-‐structured interviews were used as they provide greater depth and explanation to the observations recorded in Phase One. This is because semi-‐ structured interviews offer a solution in situations where greater depth and explanations are required to justify a given response (Harris & Brown, 2010; Phellas, Bloch & Seale, 2011). According to Byrne (2004), interviewing is useful for understanding the experiences of participants and how they made judgments in those experiences. As this thesis is centred on understanding the search and evaluation processes of international students when purchasing airline tickets, semi-‐structured interviewing was deemed to be the most appropriate method for extracting in-‐depth, yet concise information.
The benefits of semi-‐structured interviews can be seen in prior research into online behaviour. Erdoğmus and Çiçek (2011) used semi-‐structured face-‐to-‐face interviews to analyse the online behaviour of Turkish consumers who engaged in group buying. The study claimed interviews were required due to the lack of existing literature from which to form any, more objective means of analysis. They also claimed that semi-‐structured interviews were an effective means of better understanding consumers’ behaviour in their respective online environments. That being said, the methods adopted in Erdoğmus and Çiçek (2011) are reliant on retrospective accounts of participants’ own behaviour. This introduces the possibility of responses being biased by memory limitations or desires to conform to social norms (O'Gorman, Wilson & Miller, 2008; Tversky & Marsh, 2000).
Zhang (2013) investigated the effect of information preferences on consumers’ search for health information, their perceptions and needs of search tasks and user experience with search systems. This involved observing participant search behaviour using Camtasia by TechSmith, followed by a questionnaire asking about the perceived ease-‐of-‐system-‐use, how the system worked, levels of enjoyment and engagement and intentions of future use. This mixed-‐methods
approach is limited by the questionnaire, which is constrained by the predetermined confines of the questions presented. These questions may not accurately represent the views of the participants, nor fully explain the entirety of their beliefs (Pawar, 2004). Thus, an interview with questions based upon the observations made in Phase One aimed to avoid any biases or assumptions in the interview design.
Many studies have either observed online behaviour or used qualitative tools such as questionnaires and interviews. However, an extensive review of the literature failed to find any that combined observation with follow-‐up qualitative methods as adopted in this thesis. Such a methodology unobtrusively analyses behaviour whilst also understanding the reasons behind that behaviour. Each method alone is limited, either by lacking the in-‐depth understanding of causal factors or by retrospective or prompted accounts of one’s own behaviour. However, the triangulation of both methods helps to create a more objective and accurate analysis approach.