Chapter 3 Research Methodology
3.4 Data Collection Methods
3.4.1 Sampling
In general sampling is classed in one of two types, the first is probability sampling and the second, non-probability sampling. With probability sampling, each sample has a non-zero probability of selection, usually the larger the sample size, the more convincing the result (Cooper and Schindler, 2011, p.373/374), however, it is often time and cost consuming to achieve the desired number. Non-probability sampling, on the other hand, takes a more subjective approach with more practical considerations. It allows more flexibility in choosing samples that are considered to better serve the research purpose and the time and cost budget (Cooper and Schindler, 2011, p.384/385). Because first of all the purpose of this research is not to draw any statistical conclusion but to explore views, opinions and practices relating to forecast accuracy in the real supply chain environment, and second, since this is a student research project with limited time, budget, experience, and ability to access a large number of potential candidates, non-probability sampling has therefore been selected as a more realistic option. Of the various non-probability sampling methods, judgement sampling (Cooper and Schindler, 2011, p. 385) is considered the most appropriate approach based on the following considerations. Firstly, it gives a higher response rate with limited time and budget. Secondly, by selecting well-known businesses, a high quality input is assured. Lastly, because a follow up telephone interview is targeted, these subjectively selected participants are chosen with this in mind, so more expertise and experienced input can be obtained through the in-depth interview. The targeted sample frame has then been decided to mainly focus on the New Zealand market, and the potential participants are defined as those working in the supply chain or similar functions in these businesses.
Sample size involving qualitative interviewing is a difficult question to answer (Marshall, Cardon, Poddar & Fontenot 2013, p11), as the focus is the process of data generation not
Chapter 3: Research Methodology Lu
43 the numbers. It is recommended that the sample size is be determined by the nature and design of the study and to then build the sample as the research progresses (Edwards & Holland, 2013, p5&7, Trotter II, 2012, p399). As there is lack of guidelines as to the sample size in qualitative research, Marshall et al. (2013, p12) suggested that a reference to the sample size of similar research and peer views would be a good measurement. Nevertheless, some rough number guides are still given in this regard as quoted by Edwards and Holland (2013, p66) and Marshall et al. (2013, p21), i.e. a mean of 20 to 30 for masters theses, and 30 to 50 for grounded studies. It has also been proved that statistically a sample size of 30 or more will give a convincing result as it enables a close match to normal distribution (Saunders, Lewis & Thornhill, 2009, p218). In this research, bearing these guidelines in mind, similar researches have been consulted and the view from the supervising professor has been obtained, the sample size is hence decided to be around 30- 35.
The number of final interviewees is determined by the willingness of the participants completing the questionnaire, as the interview is directly related to the questionnaire and is intended to obtain more information on those topics. At the questionnaire stage a reasonable and achievable number of participants have been selected based on professional network in order to secure a good number of participants who are willing to complete a follow-up interview to provide sufficient information for the research topics. It must be acknowledged that it is challenging to obtain all of the desired participants and their inputs given this is a student research with limited budget, time and experience. The final sample result is considered acceptable given there are total 36 questionnaire responses received with 35 fully completed and one partially completed. Among those, 14 respondents agreed to participate in the follow-up interview.
3.4.2 Questionnaire
Research questions identify things the researcher wants to understand and they need to begin with certain goals and based on substantial experience and theoretical knowledge (Maxwell, 2013, p73&77). This questionnaire is developed based on the literature review completed which studied supply chain forecasting and forecast literature in general. The
Chapter 3: Research Methodology Lu
44 importance of forecast accuracy, the contributing factors and the improvement enablers to forecast accuracy are summarised and concluded from these academic studies and research. The goal of this research is to investigate the practice in the real business world, and then compare those with the academic studies, to identify gaps between these two sides and generate meaningful research topics to the researches, while on the other hand to convey the research findings and suggestions to the real business world to realize their value. The research starts with a questionnaire comprising relevant questions.
The questionnaire is divided into three parts. The first part contains general questions about the sector of the business, size of the business, the development of supply chain management function, business focus on forecast accuracy, and, the perceived impact of forecast accuracy and the related measurements. The purpose of these questions is to understand the current development of supply chain management and focus on forecast accuracy in the real business world, and to get a view from the practitioners on what are the impacts and what kind of associated measurements have been or will be set-up. The second part of the questionnaire focused on the contributing factors. Questions were asked on what are the contributing factors recognized by these practitioners, and if they hold the same view as the related literature. For each of the factors, questions were specifically designed to examine the opinions and practices held in the real business world. The third part of the questionnaire was all about the improvement enablers; similar to second part, questions were raised based on literature review findings aiming to capture the views and practices held in practice.
To best capture the responses and effectively obtain the most meaningful input, the questions were designed in different forms including single choice, multiple choice, rating, grading, weighing, and filling blanks. Many questions were given open-ended options for the participants to add their comments and describe their specific practices. In this regard, this survey is not a simple quantitative research questionnaire any more, with a considerable number of open-ended questions to capture ‘words’, it is already incorporating qualitative methodologies.
To accompany the questionnaire, a cover letter was presented to each participant taking the survey. The cover letter introduced the background and purpose of the research, explained the value and potential benefits of this research, and the reason for choosing the participant.
Chapter 3: Research Methodology Lu
45 Both the draft questionnaire and the cover letter have been reviewed by the supervising professor for error proofing purposes, suggestions and recommendations were given to improve the clarify of both. Adjustment were applied accordingly before the final version of questionnaire and cover letter were sent out.
3.4.3 Interview Questions
Interview questions collect the data a researcher needs to understand the topics (Maxwell, 2013, p77). In this research, interview questions were aimed to obtain wider and deeper information on the research topics to achieve a better understating of those subjects, it was also taken as an opportunity to obtain clarification on some of the answers to the open- ended questions in the previous questionnaire. The semi-structured interview has a prepared list of possible questions relate to each of the questions in the questionnaire, but not all questions were asked to each individual interviewee. What to ask is determined based on the interviewee’s response to the questionnaire, then interview questions are selected and adjusted accordingly in order to get a deeper understanding of the particular questions, or to obtain more information and clarification on those topics.