Project Topics
Get Complete Project Material File(s) Now! »
Data Collection
Data collection refers to gathering primary and secondary data (Hox & Boeije, 2005). Primary data is original data that are collected for a specific research problem, by conducting the right research for the problem (Hox & Boeije, 2005). Secondary data, on the other hand, are material created by researchers and are suitable to reuse in other research (Hox & Boeije, 2005). This study uses both primary and secondary data. The secondary data was collected from academic databases such as Google Scholar and Primo. For this thesis, two different primary data collection methods were conducted to fulfil the purpose, firstly by semi-structured interviews to understand in what way influences affect consumer impulse buying behavior. Finally, a questionnaire to test possible factors that influence consumers mobile impulse buying behavior. How these methods were used will be explained below. Further, the primary data gathered from the methods were beforehand explained for the participants and respondents in this study to be anonymous and confidential. The participants approved the interview to be recorded and were also informed that the data was only used for research purpose, as well as the recorded files to be deleted when the study was finalized. Additionally, since the study was executed in Swedish the Appendixes, Tables and quotes were translated into English. Implying that the meaning of words could differ from the original interpretation. Due to the timeframe for this thesis, a back-to-back transcript was not completed. However, to ensure trustworthiness, the authors put emphasis in the transcript process when translating the words and finding underlying meanings.
Semi-structured Interview
Method, semi-structured interviews were used to establish the initial findings of this study. In exploratory research with an inductive approach, in-depth interviews help to find important information and underlying motivations to get a deeper understanding in the context of the topic (Malhotra et al., 2012; Saunders et al., 2016). In comparison to other qualitative methods such as focus groups, in-depth interviews allow for a more direct and personal structure as participants are interviewed one at a time (Malhotra et al., 2012). This also results in easier coordination of the participant, and take away any social pressure, as in-depth interviews result in the free exchange of information (Malhotra et al., 2012). Semi-structured interviews allow for the researcher to address a specific issue and responses are usually easier to interpret than of other qualitative approaches (Zikmund Babin, 2016). Semi-structured interviews enable the researcher to develop a set of key questions to be covered, also called an interview guide. Though, depending on the flow of the conversation the questions may vary as well as the possibility for additional questions (Saunders et al., 2016). The main purpose of the interview guide is to guide the interviewer through the topics that need to be covered, but also as additional questions may be asked it allows for the elaboration of topics that are of interest for the participant. The semi-structured interviews were found to be the most suitable method to use for the first part of the study. Since the topic of influencer marketing in relation to impulse buying behavior is emerging, in-depth interviews allowed the authors to gather deep insight and knowledge about this phenomenon. It provided the opportunity to examine key factors for the purpose of this thesis and was suitable with the first research question, that aims to understand in what way consumers are affected by influencers when it comes to impulse buying behavior. Additionally, since the participants were interviewed one at a time and no pressure from other participants could affect each other’s answer, the authors were able to collect different individual views and compare their opinions. The semi-structured interviews helped to explore responses and to identify key factors for the quantitative method, in order to answer the research questions (Saunders et al., 2016). With in-depth interviews, an interviewers interpretation of data could result in difficulties to analyze the data as well as acknowledge hidden messages (Malhotra et al., 2012). To overcome this, it was important in this thesis to connect findings with theory to make sense of the data. Also, there is a chance that the interviewer could influence a participant’s answers because of the lack of structure (Malhotra et al., 2012). This challenge could be avoided, as the study was based on semi-structured interviews and an interview guide was used.
Participant Selection and Sampling Technique
Some requirements were set up for this study to select the right sample of participants. Firstly, the base selection required the participants to belong to the population of Instagram users, in order for them to understand the platform. Secondly, since the most popular age span for Instagram users is between 18-29 (Pew Research Center, 2016), only participants between 18-29 were asked to take part in the interviews. Thirdly, as the research was based in Sweden, only Swedish citizens participated. Lastly, participants were required to have made an online purchase of endorsed products from influencers on Instagram, as the purpose is to investigate influencers impact on consumers mobile impulse buying behavior. To conclude, the target audience of this study are Swedish Instagram users between 18-29 years old that have made an online purchase of products endorsed by influencers on Instagram. In this study, non-probability sampling was chosen as sampling technique due to lack of representation of the targeted population in a sampling frame. Since the population of m-shopping Instagram users in Sweden are not known and no list could be obtained from the population. Rather than selecting sample elements on chance as in probability sampling, non-probability sampling relies on the personal judgment of the researcher (Malhotra et al., 2012). For in-depth interviews, non-probability sampling does not provide any guidance for the sample size and rather focus on the purpose in order to explore the research and gain theoretical insights to the number of samples (Saunders et al., 2016). Accordingly, a sample size of six participants was interviewed in order to find key factors to build the questionnaire on. The findings developed from the semi-structured interviews were later tested in the quantitative method to ensure high credibility of the findings regardless of the small sample size. It is suggested to continuingly collect samples until a saturation point is reached, meaning until enough data has been collected in order to provide new information and themes (Saunders et al., 2016). This implicates that additional interviews could be essential if six interviews were not sufficient enough to collect suitable data. Though, after conducting six interviews the saturation point was met and adequate findings were collected for the purpose and key factors, indicating that the sample size was sufficient. Further, there is a range of different non-probability sampling techniques available, and in this study a judgmental sampling technique was chosen. This study had a smaller sample size as the topic at hand covers a considerably large amount of information (Saunders et al., 2016). The target audience of this study concludes of participants with an Instagram account between 18-29 years old that have made an online purchase of products endorsed by influencers on Instagram. Thus, the authors then chose to include participants based on their judgement of those matching the requirements. Though, the likelihood of the sample to be representative is low when it comes to judgmental sampling (Saunders et al., 2016). Using a maximum variation sampling enables a more representative judgement sample, where diverse characteristics are identified to ensure maximum variation within the sample (Patton, 2002). Therefore, prior to the selection of the sample, this study used a list of sample selection criteria’s. This to ensure a mix of age, occupation and difference in Instagram activity. The interview questions were of non-sensitive character, as the topic did not bring up any upsetting issues that intimidate, discredit or incriminate the participants (R. M. Lee, 1993). Considering that the nature of the interview questions was non-sensitive, the location of the semi-structured interviews were in an informal setting (Saunders et al., 2016). The semi-structured interviews were conducted at the participants’ homes since they lived in different towns and had different schedules. It was important that the participants’ felt relaxed and comfortable in order for them to be honest and share their views on the topic. This to encourage the conversation to find possible codes for the second research question. The average interview lasted for 20 minutes, a sufficient timeframe due to the concise nature of the questions and for the first research question. All the interviews were held in Swedish correspondent to the target audience and were conducted during the time period between 8th to 14th of March 2018. To be able to discover themes and codes, the interviews were recorded to ensure proper transcribing of the findings. Additionally, notes were taken during the interviews, as the authors took turn interviewing the participants one at the time. A summary of this information is outlined in Table 1. Before the interviews started the participants were asked for permission to record as well as ensuring them anonymity and confidentiality. Further, the participants were given a short information about the topic and purpose of the research. The questions in the interview guide (see Appendix 1 & 2) were conducted to get an understanding of consumers perception of the influencer marketing phenomenon. In terms of the participants’ view of influencers, as well as participants’ shopping behavior through Instagram impacted by influencers and the usage of technology when purchasing. The interview guide included open-ended questions, with the intention to ensure discussions and avoid yes or no answers, which is preferable when conducting semi-structured interviews (Malhotra et al., 2012; Saunders et al., 2016). Further, when anything recognizable from literature or reoccurring factors from the different interviews about influencers impact and its impact on impulse buying behavior appeared, probing questions was appropriate to use. The participants also had different views that enabled the authors to detect and compare different factors. First, the interview started with introductive questions to establish foundation for the discussion and to make the participants feel comfortable. These questions were about a participants’ Instagram usage to make sure they matched the requirements for the population sample and to get introduced to the topic for further discussion. Essentially, the introduction questions were about identifying if the participant was a match to fulfil the purpose of the study. The second part involved questions about the participant’s different views on influencers and their collaborations with brands and their products. Since the first part was to establish previous experience rather than concepts (Saunders et al., 2016), this part captured the relationship between the participant and the influencer. Thirdly, questions were asked about the participant’s buying behavior when it comes to purchasing products endorsed by an influencer. This, to see the connection between an influencer and the process of buying, what kind of products one buy, how frequently they buy products, as well as for what reason. Moreover, the fourth section brings up the technological use of Instagram and the buying process. Discussing the issue of where participants buy products and why. Hence, this section aimed to identify for example, if they go through the whole buying process on their mobile device or their computer, and why the specific participant think that process is the most preferable one. Lastly, when the discussion had gone on for a while and the participants felt comfortable and had discussed some cognitive behaviors, a few questions about spontaneous behavior were asked. Fundamentally, the objective was to identify what factors might have caused the participants impulsive buying behavior.
Pilot Testing
Before conducting the semi-structured interviews, a pilot test was completed on the interview questions. The purpose of the pilot test was to assess the questions, environmental settings and the flow of the discussion. The authors chose one participant within the targeted population of this study to take part in a simulated interview, that later were excluded from the sampling frame of the semi-structured interviews. The outcome of a pilot test is to capture how a participant would perceive the questions and where the nature of the conversation will lead the discussion. The pilot testing was very useful, especially since the authors of this study were new to semi-structured interviewing. The participant in the pilot test was a 23-years-old female student. The test interview took place at Jönköping University in a quiet area and lasted for 15 minutes. The result of the test interview revealed some changes that were necessary. It concerned changes of the flow during the interview, such as not to ask repeating and similar questions back and forth in the interview, but also change the wording of some questions in order to lead the discussion in a natural flow. These improvements helped the authors to minimize the chance of collecting biased data from participants that might not understand the questions or think they are confusing in some way. Otherwise, the authors of this study felt they had incorporated all the necessary question to fulfil the purpose of the semi-structured interviews.
Data Analysis
The qualitative data from the semi-structured interviews were analyzed using a Thematic Analysis. When researchers want to find themes and patterns in their data, this type of analysis is beneficial and flexible to use when seeking to understand factors of human attitudes and exploring interpretations of a phenomenon (Saunders et al., 2016). The purpose of the first research question was to understand in what way influencers affect consumers mobile impulse buying behavior via Instagram in order to answer the second research question. Therefore, the interviews were constructed to find factors influencing mobile impulse buying behavior impacted by influencers on Instagram, that also confirmed and expanded the authors’ thoughts and experience of this phenomena. Since previous research method of semi-structured interviews was mainly used to gain an understanding of the findings. Moreover, Thematic Analysis was suitable since the authors wanted to gain insight and knowledge from the data gathered in order for the data to be tested in the quantitative method, this was also suggested by Saunders et al. (2016). The process of analyzing the data was examined by both the authors together throughout the whole process. Initially, it started by becoming more familiar with the data and compiling the results from each interview. Thus, by listening and transcribing the six recorded interviews. Thereafter, the notes and transcripts were read thoroughly and further analyzed. To understand the answers and to determine the most valuable information for the purpose, codes were identified. The codes were summarized and interpreted from the meanings of the answers and divided into themes (Saunders et al., 2016). After the analysis, a total of 22 codes were identified and divided into four different themes. The codes were distributed to different themes since they were of different characteristics. The characteristics were grouped and divided into these different themes and identified as; Instagram Influencer, Reason for purchase, Emotional Influence and Technological Use. The codes in the theme of Reason for purchase were partially found in previous researched and used to identify consumers different types of impulse buying behavior. These are further explained in detail in the findings, and the codes were tested in the quantitative study.
1. Introduction
1.1 Background
1.2 Problem Definition
1.3 Purpose and Research Questions
1.4 Contribution
1.5 Delimitations
1.6 Definitions of Key Terms
2. Literature Review
2.1 Social Media Marketing
2.2 Mobile Shopping Behavior
2.3 Impulse Buying Behavior
2.4 Ethics and Society
3. Methodology
3.1 Research Perspective
3.2 Research Approach
3.3 Research Design and Strategy
4. Method
4.1 Data Collection
4.2 Semi-structured Interview
4.3 Questionnaire
4.4 Quality of the Research
4.5 Ethics in Marketing Research
5. Empirical Findings
5.1 Findings from the Semi-Structured Interviews
5.2 Findings from Questionnaire
6. Interpretation
6.1 Consumers Perspective of Influencers Impact
6.2 Investigating Factors with Influential Impact
6.3 The Process of Influential Impact
7. Conclusion and Discussion
7.1 Purpose and Research Questions
7.2 Implications
7.3 Limitations
7.4 Future Research
