4. CHAPTER FOUR: SNOWBALL SAMPLING PROCESSES
4.5. Sampling process in German banks
At the beginning of November 2006, many change agents were busy concluding their tasks for the end of the year. Therefore, they postponed interviews and were not able to recommend other change agents to participate. As Figure 10 suggests many attempts failed to build a snowball. As a result, many change agents at distinct organisational levels were approached who had worked with me in previous technological projects.
The first participant CA30 recommended CA31 who agreed to be interviewed. CA31 referred to a change agent who had extensive change agent experience within technological projects in a target bank in Australia (see Figure 8). After the interview, CA32 referred to the change agent who replied after I had completed the data collection in Germany. CA33 participated in my study and recommended CA34 who agreed to the interview. After the interview, CA34 referred to CA35. CA36 could not locate any potential participants. After the interview CA37 recommended CA38. I conducted the interview with CA39 who was not able to provide me with any contact details of other change agents. After the interview, CA40 recommended a change agent who said to her that he would participate in my study. However, he did not reply to my e-mails. So, another change agent who was not target participant referred to the last change agent CA41.
Figure 10: Snowball sampling in German banks
CA30
CA31 CA17 CA32
CA33
CA34 CA35
CA36
CA37
CA38
CA39
CA40 CA41
Figure 10 visualises the behaviour of change agents (CA30-CA41) in the snowball sampling process within target banks in Germany, while Table 7 details the evolution of the sampling process and illustrates 16 potential participants who were invited for interviews. Of the 16 invited people, 12 change agents participated, one change agent declined to participate, two potential participants did not respond to my invitations and one change agent answered after the sample quota was reached. The data collection with change agents in banks was occurring in the same time period as within public sector settings and took about three months.
Table 7: Sampling process in German banks
Individuals in sampling process
Behaviour Behaviour led to recruitment of
CA41 Active Reached sampling quota
Table 7: (continued). Sampling process in German banks
Individuals in sampling process
Behaviour Behaviour led to recruitment of
The sampling process has been a catalyst for change which is reflected in varying transformations. At the beginning, I transformed myself from European practitioner into
“Australian academic researcher” then to “Australian and German practitioner” at the same time and at the end of the journey back to “Australian academic researcher”.
These processes drove me to examine the patterns of behaviour of participants and widen my mental horizon. Nevertheless, acting outside an organisational structure in a new role as researcher in Australia and Germany involved uncertainty and required courage, patience and persistence. In that process, I needed simultaneously to apply old skills and learn new ones such as networking and strategic thinking regardless of different environments. After a well-planned rationalist strategy failed to recruit change agents, an incremental approach based on learning-by-doing was practiced.
A number of people at different organisational levels were approached to provide names of different change agents in Australia. The informants gave me potential participants’
names, nevertheless, they requested to remain anonymous in e-mails. So, invited change agents declined invitations or did not respond. As a result, I centred on active individuals who encouraged change agents to participate. Due to their pro-active behaviour towards this study, these individuals were labelled actors who led to the
recruitment of change agents in public and private sector settings in Australia. In contrast, in Germany, I already had many contacts with change agents from my previous work experience and did not require the guidance of actors.
The lack of a significant job title and corporate image made the entire sampling process messy and challenging. An effect of this was that I needed to rely on key individuals for the purpose of recruitment in both countries. The building of personal relationships was a prerequisite for the recruitment of change agents. The various divergent roles performed by change agents had a significant impact on the behaviour of potential participants. It was possible to identify patterns in the behaviour of the active and passive change agents. Those who agreed to be interviewed were more likely to be recommended by active participants. In contrast, those who did not reply to the e-mails or declined invitations were suggested by passive change agents. In most cases, the active behaviour led to snowball sampling, whereas, passive roles temporarily discontinued the entire sampling process and I needed to reflect and re-examine the sampling approach. The individual reflection model (see Figure 4) had multiple purposes. In the first place, the reflection helped to cope with running emotions, which triggered a learning process. The learning process built a basis for the recruitment of change agents within six target organisations in Australia and six proposed companies in Germany.
In Australia and Germany, 41 change agents participated, 14 did not respond, nine declined the invitation and two replied after the sampling quota was reached. Due to the fact that I have never worked in Australia, I needed the assistance of actors to recruit initial change agents while building relationships with various people in public and private sector settings; for this reason, the Australian sampling process lasted three times longer (about nine months) than the German process (about three months). The assumption that women would participate or support this research because of my gender could not be confirmed. Most participants were men. The majority of participants
preferred face-to-face interviews. Conducted interviews will be discussed in the next chapter.
5. CHAPTER FIVE: ANALYSIS OF PUBLIC SECTOR