3. Methodology
3.3. Sample Selection
For researchers, understanding the theory and practice of the samples is a necessary skill, whichever research method or approach they are adopting (Ritchie, Lewis, Nicholls, & Ormston, 2014). In fact,
even if a study involves a small sample or a single case study, there are still decisions that need to be defined about the people, settings or actions to cover and select (Merriam S. , 2009).
A key characteristic of qualitative samples is that they are relatively small in size, therefore enabling an in-depth exploration of the phenomena under investigation (Ritchie, Lewis, Nicholls, & Ormston, 2014).
According to table 3 there are two sampling processes, the non-probability and probability types. The probability procedure has the characteristic that every individual of the population has equal chance of being selected and the selections are made independently, which means that the probability of a unit being selected is not affected by the selection of other units from the population (Teddlie & Yu, 2007). This method is mostly used in quantitative studies.
Table 3 - Comparisons Between Non-Probability and Probability Sampling Techniques
Dimension Of Contrast Non-Probability Sampling Probability Sampling
Overall Purpose Of Sampling Generate a sample that addresses research questions
Generate a sample that addresses research questions
Issue Of Generalizing Sometimes seeks a form of generalizability (transferability)
When The Sample Is Selected Before the study begins, during
the study or both Before the study begins How Selection Is Made Uses expert judgement Often based on application of
mathematical formulas
FORM OF DATA GENERATED Focus on narrative data (numeric data can be generated)
Focus on numeric data (narrative data can be generated)
Source: (Teddlie & Yu, 2007)
For the purpose of the current qualitative study, a non-probability sample, specifically a purposive sample was chosen. The main reason, as pointed by Maxwell (1997), is that particular people are deliberately selected for the important information they can provide that cannot be acquired from other choices. The main idea is to purposefully select participants as well as important documents and information that will best assist the researcher to understand why TAP became private.
Purposive sampling means what the name suggests – members of a sample that are chosen with a
“purpose”. The main purpose is to give insight of the company and the reasons that led to the privatization. For this reason, interviewed people are related to TAP, including the actual CEO, current and former members of the executive management council (which enables the study to have perspectives of people prior and/ or during the privatization process), the actual CFO and the coordinator of the TAP’s union SITAVA. What they have in common is that they were all in the company before the privatization to occur, which is why they all represent valuable sources to understand this process.
According to Rwegoshora (2014), sample size may or may not be fixed prior to data collection, depending on the resources, time available and the objectives of the study. The sample size was fixed prior to data collection, however, during the gathering it was determined based on theoretical saturation6 and availability of the respondents. In fact, the study assumes a certain degree of participants’ homogeneity because in purposive samples, participants are, by definition, chosen according to some criteria – in this case, are working or worked in the company, mainly at top level management positions. The more similar are the participants’ experiences in a sample, the sooner it is expected to reach saturation (Guest, Bunce, & Johnson, 2006).
According to Creswell (2013), the number of participants involved in a qualitative case study is characterized by being small. Even though there is no agreed answer on how many participants a case study should have, some authors present their opinion.
Guest, Bunce, & Johnson (2006) raised the following question: “did six interviews, for example, render as much useful information as twelve, eighteen, twenty-four, or thirty interviews?”. They
6Theoretical saturation is the point in data collection when new data no longer brings additional insights to the research questions (Rwegoshora 2014).
conducted a study, analyzing interviews from women’s in Ghana and concluded that after six interviews, 73% of the thematic discovery occurred. Nielsen & Landauer (1993) created a mathematical model based on results of six different projects and demonstrated that six participants can uncover 80% of the major usability problems within a system. In fact, they calculated that the highest return on investment was obtained with five evaluators. According to Creswell (2013), qualitative case studies should have between four to five cases.
According to the literature, the size of purposive samples is mainly determined by saturation.
Charmaz (2006) mentions that when gathering fresh data, no longer bring new insights or reveal new properties, then data saturation is occurring.
After all being mentioned, five interviews were performed, mainly due to theoretical saturation, availability of the respondents and to the fact that it was acceptable by several authors and studies.
Additionally, one type of purposing sample is the snowball sampling method, in which participants with whom the researcher contacted, suggested other members that could participate or contribute to the study (Mack, Woodsong, MaQueen, Guest, & Namey, 2005). In the current study, one of the participants was reached as a suggestion of other, which is a characteristic of the snowball method.