This chapter outlines the research methodology adopted in this dissertation. There are three subsections. The first subsection outlines the two common epistemological considerations in business research and provides the reasons for choosing one of them for this research. The second subsection explains the research design, including nature of the research, data collection and analysis methods. The third subsection explains two limitations regarding the research method.
Positivism versus Interpretivism
There are two main research paradigms in marketing (Hussey and Hussey, 1997 cited in Malhotra and Birks, 2007). They are positivism and interpretivism. These two paradigms would be explained below.
Positivism involves a natural science approach (Bryman and Bell, 2007; Malhotra and Birks, 2007), which emphasizes causality (Malhotra and Birks, 2007). Causality helps the researcher in explaining and predicting marketing phenomena (ibid.). In order to succeed in doing this, the research must rely on factual information (ibid.). The reliance on facts leads to an emphasis on objectivity (ibid.). The purpose of a positivist research is to verify theories from other sources (ibid.), which agrees with the deductive approach that either confirms or rejects hypotheses derived from existing theories (Bryman and Bell, 2007). However, although it focuses on validating facts or theories in an objective sense, it may not be appropriate in social science research because of its natural science approach (ibid.). Natural science emphasizes on objectivity (ibid.) and causality (Malhotra and Birks, 2007). There are many causes for marketing phenomena, and therefore it may not be suitable to focus only on the probabilistic cause and effect relationship (ibid.).
Interpretivism has an opposite stance from positivism, for which it emphasizes on understanding consumer behaviour and obtaining new comprehension of it (Malhotra and Birks, 2007). It is subjective in nature, which includes the view of the researcher (Bryman and Bell, 2007). The subjectivity makes it completely different from the natural science positivist approach (ibid.). This approach agrees with inductive theory, which is the generation of theory from the findings of research (ibid.). Bryman and Bell (2007) mentioned that symbolic interactionism, which is an element of the second research question in this research, is related to interpretivism. Symbolic interactionism is similar to phenomenology (ibid.), which asserts the interpretation of an individual’s behaviour from his/her own view (Bogdan and Taylor, 1975 cited in Bryman and Bell, 2007). However, despite the interpretivism’s more suitable framework for social science research, it has its own limitations. As the findings are more subjective (Malhotra and Birks, 2007) and personal (Malhotra and Birks, 2007; Rosenau, 1992 cited in Holden and Lynch, 2004), they may not be comparable or applicable to other settings (Holden and Lynch, 2004).
From the above brief analysis, both positivism and interpretivism are with its benefits and critics. However, as the aim of this research is to explore the consumption and interaction patterns of young female university graduates in Hong Kong, it has the tendency to focus on the understanding of their consumer behaviour. Therefore this research would adopt an interpretivist approach. This research is not aiming at a generalized application to certain theory but an understanding of their construction of the possibly new sense of self, thus the limitation of interpretivism mentioned above could be solved.
Research Design
In this subsection, three topics of research design are discussed. The first part outlines the approach of this research as a result of using interpretivism as the underlying philosophy. The second part discusses the data collection process, including the use of research tools, the handling of confidential issues and sampling techniques. The last part describes the data analysis process and gives an example of data interpretation using an excerpt from one of the participants.
Nature of the Research
In this research, an exploratory research design was adopted because of its nature in gaining insights and understanding marketing phenomena (Halman, 2002 etc. cited in Malhotra and Birks, 2007). As the aim of this research is to explore and understand the consumer behaviour of young graduates, exploratory research is a more suitable approach. Exploratory research is beneficial in being flexible in the research process or when the research does not have a structured background with the research field (Malhotra and Birks, 2007). This research uses only primary data instead of together with secondary data because the current issue is of an exploratory nature and thus it may not be able to be represented by secondary data which are mostly quantitative and are lacking further interpretation (ibid.). Using primary data is advantageous for its ability to address specific problems (ibid.). But it may be more time and money consuming in data collection and analysis (ibid.).
In view of this research design and the adopted philosophical stance, a qualitative approach is used because of its inductive and interpretivist nature (Bryman and Bell, 2007). Malhotra and Birks (2007) mentioned that exploratory research is used in instances where quantitative measurement is not appropriate. Also interpretivism focuses on understanding marketing phenomena (Malhotra and Birks, 2007). In addition, induction implies the formation of theory from an elaboration of findings from an unstructured framework, which agrees with one of the benefits of exploratory research (ibid.). These reasons constitute the usage of qualitative approach in this research. Qualitative research is especially beneficial in eliciting sensitive, complex issues and uncovering underlying causes of human behaviour (ibid.). However, the main critique of this approach is its use of non-representative sampling (ibid.). But as the aim of this research is to explore the characteristics of young female graduates’ consumer behaviour instead of establishing a generalized finding, this limitation would not constitute a problem.
Data Collection
As the current research is adopting a qualitative approach, in-depth interviews are used to collect primary data. The ‘ understanding’ nature of in-depth interviewing agrees with the aim of this research (Malhotra and Birks, 2007). Also, the process is more flexible that there is no formal structure guiding the flow of the interview (ibid.). Johnson (2002 cited in Malhotra and Birks, 2007) suggested that qualitative interviews could be used to understand at a greater depth about daily events, for which multiple reasons and reflections could be reviewed. Thus the consumption by the participants and their underlying reasons for their behaviour could be more deeply evaluated. However, fewer interviews could be conducted because of the longer time and higher costs needed for interviews (Malhotra and Birks, 2007). Thus analysis should be focused on the quality of the interviews instead of the quantity (ibid.). Also the collected data may be difficult for interpretation because what the participants mentioned may have some hidden meanings behind (ibid.). Therefore researchers have to be attentive to the transcripts and also the observations during the interview process (ibid.).
Semi-structured interviews were chosen because of its flexible nature and the help of a general interview guide with prompts or topics for the interview process (Bryman and Bell, 2007). Prasad (1993 cited in Bryman and Bell, 2007) mentioned in her research that the benefit of semi-structured interviewing is its ability to keep the interview focus but also allows flexibility on interview timing and the research issues. The interview guide (Appendix I) is designed in accordance with the three research questions and the interview questions are sorted with respect to the flow of the interview issues. Some questions initially not included in the guide may be asked as the participant mentioned some other related information of the research issues (Bryman and Bell, 2007).
It is essential to obtain the approval of participants for collecting their information and allow them understand the purpose of this research due to their confidentiality (Carson et al., 2005). Therefore, before the interview, a consent form (Appendix II) which explains the purpose of this research was provided for them to sign. A further oral approval from them for recording the interview was obtained before the start of the interview. Recording the whole interview is important for deeper and repeated interpretation of the conversation and elimination of any biased view of the research issue (Bryman and Bell, 2007).
Non-probability sampling technique was used to select participants for interviewing. Specifically, judgemental sampling was chosen because it is based on personal judgement of the researcher in selecting appropriate participants (Malhotra and Birks, 2007). This technique was chosen mainly because of the convenience of this sampling (ibid.) and limited time for this research. As this research is not aiming for generalized application to a wide population, this sampling technique should be more suitable in this context (ibid.). Ten participants were invited for the interview. All of them are graduates of last year and are currently having a job or were working in a full-time job for at least half a year. They are also friends of the researcher in order to have an easier access and a more convenient understanding of their situations.
The locations of the interviews were mainly friend’s home or restaurants where a silent environment is provided for clear recording of the interview and less distraction of focus of the participants. The interviews usually took about 40 to 60 minutes. The language used in the interview was Cantonese, which is the mother tongue of Hong Kong residents, in order to allow participants to speak more freely and conveniently and researcher to understand the issues more thoroughly. A brief introduction of the aim of research was first explained to the participants. Some relatively relaxing introducing questions were asked at the beginning such as ‘ Did you go shopping recently?’ or ‘ What products did you buy recently?’ for warming up and allowing participants time to be relaxed to speak more about their views. The original order of the questions in the interview guide would be changed according to the speech of the participants. This would allow the participants to feel the convenient flow of the interview and be able to tell more deeply about the issues she wanted to elaborate. Projective technique was used during the interviews to encourage participants to elaborate more deeply about their underlying feelings or beliefs about the current issue (Malhotra and Birks, 2007). Probing was also used during early times of the interview to motivate participants to allow them focus on their interested issues or an important point they mentioned (ibid.).
Data Analysis
After data are collected from the interviews, all ten interviews were transcribed and translated from Cantonese to English word-by-word into transcripts (Appendix III). The reason to transcribe all details of an interview is that the analysis aims at finding out not only the content but also the way they spoke their opinion (Bryman and Bell, 2007). Also, the full transcript allows a comprehensive and possibly repeated investigation on the data (ibid.). After all the interviews were transcribed, data reduction was used to organise the large amount of data and some unnecessary data would be withdrawn from interpretation (Malhotra and Birks, 2007). Coding was then used to manage the remaining data (ibid.). Separate ideas and points in the data were allocated to appropriate categories for interpretation (ibid.).
For data interpretation, the excerpt from Yvonne is given as an example. In this excerpt, she emphasized about her tiredness of work and thus she would like to go for relaxation:
“…maybe it’s because sometimes it’s tired for work, you work for so long, sometimes let’s say you can’t go shopping for the 5 days, so it remains holidays like Saturdays and Sundays, so you want to relax and find something to reward yourself. So I tend to for example buy clothes, that thing. Because of these reasons, for example always feeling great pressure, so buying clothes this activity becomes more frequent. Or it’s more frequent to do other things… Yoga, because I also, because after started working I think that I really want to find something to do, like doing exercise, for relaxation, so I especially go to apply for a Yoga course.” (Yvonne, Occupational Therapy, Interview 3)
She stressed twice about her desire to relax from work, and she also mentioned twice about the tiredness and pressure from work. Her excerpt implies that her extra spending on shopping or Yoga lessons is mainly because of her work. Therefore it could be deduced that these consumptions are caused by her work identity.
Limitations
Apart from the relatively subjective nature and the non-representative sampling technique common in qualitative research, there are two further limitations related to the specific context of this study.
The first one is about the number of participants recruited for interviewing. As time was limited for the dissertation, only ten interviews were conducted. Although the sample of this study does not need to be representative, more interviews would constitute a more thorough analysis of the general trend of the young graduates’ consumer behaviour, especially for some of participants mentioning that some of their friends are thinking at an opposite way. Also the occupation of participants was not sampled in an evenly distributed way. Thus the difference between two extreme occupations may not be drawn. However, the data collected from two participants having the same occupation may be used to verify Young’s (1991) suggestion that behaviour of two people acting the same social role may not be the same.
The second limitation is the research method used in this study. Only semi-structured interviews were used to collect primary data. Focus group was not used due to the limited time for this study. However, if more time was allowed, focus group would be a good way in drawing more insights from the interaction between participants (Carson et al., 2005). One benefit of doing focus groups is the unexpected ideas evolved from the group discussion which would not be possible for interviews (Malhotra and Birks, 2007). However, it is more difficult for the researcher to control the group dynamics and analyze the data (Bryman and Bell, 2007). But in the current context, a focus group discussion might be beneficial to generate more ideas about the possible underlying reasons for certain consumer behaviour.