Snowball sampling (Frankwick, Ward, Hutt & Reingen, 1994; Bryman, 1999; Salganik & Heckathorn, 2004) was employed to solicit participation in the research. A discussion of the merits and demerits of the snowball sampling technique is considered before describing the process by which snowball sampling was employed in this investigation.
Snowball sampling is essentially a technique for finding research subjects through those who have already been recruited for the research (Atkinson & Flint, 2001; Heckathorn, 2002). Snowball sampling is applied primarily for explorative, qualitative and descriptive research (Hendricks, Blanken & Adriaans, 1992) and for making inferences about a population of individuals (Faugier & Sergeant, 1997; Franwick et al., 1994). Atkinson and Flint (2001) argue that snowball sampling is an economical, efficient
and effective method for qualitative studies. For example, Franwick et al’s. (1994) study of strategic decision making employed this technique to identify members of the company who were involved in the decision- making network and processes that were linked with the introduction of core technology in the organisation. In the context of this research, senior managers and CEOs were asked to refer other senior managers and CEOs so that these can then be contacted to ascertain their interest and availability for this investigation. It can be argued that such respondent- driven sampling in a corporate environment would allow for a better sample selection in that it narrows the focus only to those who are likely to be engaged in leadership succession.
Nevertheless, there is argument that snowball sampling is subject to a number of biases. For example, in snowball sampling, it is possible that sample members might not be selected from a sampling frame; a respondent with many friends would be more likely to be recruited into the sample than another person with few friends. Furthermore, a snowball sample is not strictly random and therefore has limited use in quantitative studies but Bryman and Bell (2003, p. 107) argue that this is not a large or intractable problem where qualitative research is concerned. Another disadvantage associated with finding respondents and with initiating the chain of referrals is the matter of gatekeeper bias (Groger, Mayberry & Straker, 1999). In Groger, Mayberry and Straker’s (1999) research, data collection was hampered by nursing home staff through whom they were trying to obtain consent and access to their sample population. These staff acted as gatekeepers who often impeded access to
their intended respondents. In this research, this was also experienced in the form of secretaries and personal assistants to CEOs in family- controlled publicly listed companies who often appear to this researcher to make unilateral decisions that their CEOs would not be interested in this research. In a number of cases, these secretaries did not even suggest that they would ask their CEOs of their possible interest in this research. This could have been because these CEOs had been asked to participate in research but had given some form of ‘standing orders’ to their secretaries to not entertain such requests.
As the participants had to be senior managers who were involved in either the planning and/or the implementation of leadership succession processes, only those who were in appropriate senior management positions and above were contacted. The process is described below.
The CEOs and senior managers in ten Malaysian companies were selected from the network that I had previously established and were initially contacted by telephone and formally invited to participate in this research. The purpose of the research was explained and their voluntary participation was then solicited. Of the ten contacted, three responded positively while the rest indicated that they were not able to participate in the research for various reasons, chief of which was the problem of confidentiality. The three who responded positively were then sent complete details of the research along with the appropriate consent forms. They were also subsequently asked for the names of industry colleagues who they felt might be interested in participating in this research. From
these three, another 20 potential participants, including their own CEOs, were obtained and they attempts to contact them by telephone were made. Not all could be reached and of those who were contacted only two indicated their willingness to participate in the research. Using the snowball sampling technique a total of 65 potential respondents were contacted. In every case of a CEO in a family-controlled company, the invitation to be a part of the research was politely declined by the CEO’s secretary or personal assistant. The process of asking for referrals resulted in a total of 12 participants from a number of industry sectors. However, one of the 12, who was coincidentally the only female CEO who had agreed to participate in the research, subsequently declined to participate when I arrived at Kuala Lumpur. Of the remaining 11 participants, one was subsequently omitted from the sample pool due to the fact that although he had been identified using the snowball sampling technique, and he was highly placed in his company, the interview revealed that he was more appropriate for a marketing-oriented research topic than for this investigation. This is one of the weaknesses of snowball sampling as there can be control slippage over the precise nature of the population from which participants would be drawn (Frankwick, Ward, Hutt & Reingen, 1994; Bryman & Bell, 2003).
As a result, a total of ten participants, representing nine companies, were recruited for the research. There were two participants from one of the companies contacted who were both senior managers in the company in question and had heard about the research and had requested to be
included in the study. As this was a company for which I had provided training in the past, I welcomed the inclusion of both participants.
Of the ten participants, four carried one or more of the following titles: CEO, Managing Director or Executive Director – indeed, two of the four had the dual title of CEO/MD. The remaining six were senior managers in their respective organisations who were involved in policy and corporate strategy development and implementation. The participant list, along with a brief description of each company, is shown below in Table 4.1. A full description can be found in Appendix D.
In terms of market capitalisation, two of the nine companies involved in the research can be found in the top ten positions in terms of market capitalisation in the Kuala Lumpur Stock Exchange. It should also be noted seven of the nine companies are parent companies for a total of 450 subsidiaries – the exceptions being Cilicon Berhad and FineAds Berhad. Perusal of the annual reports published by the seven parent companies will reveal that this is so. It is therefore possible that the leadership succession practices of these parent companies will be adopted by these ‘child companies’ (the Malaysian equivalent of the Japanese kogaisha) which would be in keeping with the practice of ‘father leadership’ in East and South East Asian companies.
Table 4.1: Participant list and company description
Company description and pseudonym Participant
code
(pseudonym)
Number of
participants from the company and status
Company A:
Pseudonym: AmSure Berhad
Number of employees: Over 20,000
Industry sectors: Finance, banking and insurance Listed on the Kuala Lumpur Stock Exchange (hereinafter known as the KLSE)
Status: Government-linked Company
Amar Singh 1 (Senior manager)
Company B
Pseudonym: Berani Berhad
Number of employees: Over 2,000
Industry sectors: Manufacturer of industrial products, health foods, hotel development and management, construction
Listed on the KLSE
Status: Family-controlled publicly listed company
Bernard 1 (Managing
Director)
Company C
Pseudonym: Cilicon Berhad
Number of employees: Approximately 2,000
Industry sectors: Manufacturer of components for the electronic industry
Listed on the Nikkei Stock Exchange
Status: Non-family-controlled publicly listed company Charles and Chester 2 (Human Resource Director and a senior manager).
Company D
Pseudonym: DesignerPens Berhad
Number of employees: Approximately 500 (in
Malaysia)
Industry sectors: Manufacturer of stationery and writing instruments
Listed on the KLSE
Status: Family-controlled publicly listed company
Daniel 1 – (Human Resource Director). In addition to providing views on DesignerPens Berhad, this
participant was also able to give views on Cilicon Berhad as he had recently resigned from Cilicon Berhad
Company E
Pseudonym: Ellaconstruct Berhad
Number of employees: Approximately 1,000
Industry sectors: Manufacturer of construction materials and construction and development Listed on the KLSE
Status: Non-family-controlled publicly listed company
Ezra 1 (CEO/MD). The
respondent carries both the CEO and MD titles.
Company F
Pseudonym: FineAds Berhad
Number of employees: Approximately 100
Industry sectors: Advertising Status: Privately owned company
Frederick 1 (CEO/Executive
Director). Like Ezra, Frederick carries two titles.
Company G
Pseudonym: Great Flours Berhad
Number of employees: Over 2,500
Industry sectors: Flour milling, manufacturer of animal feed, fast-moving consumer products Listed on the KLSE
Status: Family-controlled publicly listed company
Company H
Pseudonym: HanSoon Berhad
Number of employees: Over 1,500
Industry sectors: Manufacturer of fast-moving consumer good
Listed on the KLSE
Status: Family-controlled publicly listed company
Henry 1 (Senior Regional
Manager)
Company J
Pseudonym: J-Power Berhad
Number of employees: Over 5,000
Industry sectors: Construction, property
development, water treatment, power generation Listed on the KLSE
Status: Family-controlled publicly listed company
Jeremy 1 (Senior Division
Manager)
In order to maintain confidentiality, pseudonyms were employed for the companies and for the participants involved in the research. The word ‘Berhad’ that follows each company pseudonym is the Malaysian equivalent of ‘Limited Company’; for the sake of brevity, the word ‘Berhad’ may not always be included when discussing the companies in this thesis. Furthermore, the pseudonyms are cross-referenced and the identities of the participants are maintained under lock and key.
Of the ten participants, nine offered opinions on their current companies while one, Daniel, was able to offer opinions about two companies as he had resigned from Cilicon to join DesignerPens between the time when he was first contacted and the time when the interview actually took place.
Eight of the participating companies are very large publicly listed companies – seven of these are large conglomerates in Malaysia while one, Cilicon, is listed on the Nikkei Stock Exchange in Japan. Only one of the participating companies, FineAds, is a privately owned company. A number of privately owned companies were contacted but apart from FineAds, they all declined participation in the research, ostensibly for confidentiality reasons. It might be argued that the inclusion of only one privately owned company is unrealistic in that it gives only a narrow perspective of how CEO succession is implemented in one, rather than in a number of privately owned companies. When FineAds agreed to be part of the research, I considered that it would be advantageous to include it in order to compare and contrast the way in which a privately owned company implemented CEO succession with the methods used by publicly listed companies.
The collective views of all the participants reflected succession practices of the organisations that they were leading (that is, the parent companies) as well the practices of their subsidiaries. As mentioned earlier, seven of the nine organisations involved in this research have subsidiaries akin to what the Japanese call kogaisha or child companies. For example, Gary who is the managing directors of Great Flours is a director in the holding
company which has approximately 140 subsidiary companies involved in the following sectors: manufacturing, entertainment and hospitality, property development, shipping, commodity trading and waste management. Another organisation, AmSure Berhad is the parent company of about 150 subsidiaries while J-Power has approximately 90 subsidiaries. Approximately 40 percent of these subsidiaries are publicly listed companies. Berani, Ellaconsruct and HanSoon have a total of 70 subsidiaries among them. In total, the views and opinions gathered from these ten participants would be representative of leadership succession practices in the more than 450 Malaysian subsidiary companies that they collectively represented. As can be seen in the findings, like their Japanese counterparts, the Malaysian version of the Japanese kogaisha looked to their parent companies for instructions in strategic areas such as leadership succession planning and implementation.
There were two reasons for using the ten participants selected. Firstly, they represent a broad spectrum of Malaysian businesses spanning the following sectors: finance, electronic manufacturing, consumer products manufacturing and sales, advertising, power generation, property development, construction and construction materials and pharmaceuticals. Furthermore, as mentioned above, with the exception of Cilicon Berhad and FineAds Berhad, the other seven companies are parent companies that collectively have 450 subsidiaries. That these subsidiaries emulate the leadership succession practices of their parent companies was confirmed during the interviews with the respondents. As such, it is reasonable to argue that the opinions of the respondents in this
investigation represent the leadership succession practices of their subsidiaries. That this is so can be seen in a perusal of company documents from, for example, AmSure Berhad. The leadership succession planning processes of two of the largest subsidiaries in this company were made available to this research. For confidentiality reasons, while I was allowed to study the documents, photocopies were not allowed. Nevertheless, it could be seen that the documents were essentially the same and that the practices in these subsidiaries mirrored those of the parent company. Secondly, securing the participation of senior managers for research of this nature was not an easy task due to their extremely busy schedules.
Having secured the consent of these ten senior managers to participate in the research, appointments to interview them were made and I then flew to Kuala Lumpur to conduct the interviews. The actual interviews were spaced over a one-month period (between June 10 and July 10, 2007) in order to accommodate the schedules of the ten participants. A 90-minute interview was negotiated with each participant. In retrospect, I am very grateful to each of the participants as most of them actually allowed for a two-hour interview. During these interviews, many of the respondents made available pertinent company documents such as annual reports, human resource management policies, leadership development training programmes, organisational charts, press releases and leadership succession policies for closer scrutiny. While these documents were made available for perusal, only copies of press releases and annual reports were allowed to be taken out. Additionally, published annual reports for
the past ten years, and magazine, trade journal and newspaper reports pertaining to these companies had been scrutinised.
Each interview was audio-taped as consent was obtained from the participants to do so. The audio-taping of interviews has been cited by Heritage (1984, p. 238) as being advantageous in that it provides an accurate record of what had been said and helps to correct the natural limitations of our minds and memories. The accurate transcription of such recordings allow for a more thorough examination of the responses provided. Personally I found that recording the interviews allowed me to concentrate on asking questions and avoid having to spend time taking notes. The participants were informed that each audio-tape would be securely stored and would be destroyed upon the completion and acceptance of the thesis. Each of the interviews was then personally transcribed by me as there are many nuances and peculiarities in Malaysian conversation that I did not want to lose during transcription. Three participants were contacted during and after the transcription process in order to obtain more information or clarification of the points they had made. Permission to do this had been obtained during the interview sessions. A copy of the consent form used to solicit their involvement can be found in Appendix C.