Qualitative studies approach the concept of sampling differently to that of more
traditional, quantitative orientated research methodologies. In quantitative research, the desire is to obtain a sample that is representative of the population of interest to allow
Methodology and Method 83 generalisations from the research findings to similar populations. Because qualitative
research is more focussed on the depth and richness of information produced by the research participants rather than the ability to generalise conclusions to other populations, sampling is guided more by concerns surrounding the ability to obtain a sample that is most appropriate given the research aims and can supply sufficient data to allow the research question to be adequately explored (Fossey, Harvey, Mcdermott, & Davidson, 2002).
Before commencement of sampling and recruitment, I attended a meeting with the local police Area Commander and District Intelligence Manager. The purpose of this meeting was to explain my proposed study and gain their written consent to interview police officers in the local district. A research proposal outlining all pertinent information regarding the proposed study was presented and discussed. After obtaining written consent to interview officers in the district, I was provided with a list of potential police officers considered to be relevant for participation in light of the issues discussed during the meeting.
On first conception, the intended sample for this study was limited exclusively to front-line police officers. It was believed that police officers involved in front-line duties would have the most specific and detailed experience with negotiating situations and relationships involved in policing the mentally ill in a practical, material manner. However, as Fossey et al. (2002) note, sampling in qualitative research is often a process that can be influenced and re-evaluated as a result of engaging in the research procedure itself. As such, sample population definitions can be subject to change in order to better address the research aims. The list of potential participants provided by senior police management was not limited to front-line duty officers: some were involved in community and youth projects
84 Methodology and method and other various job-roles not exclusively ‘front-line’ in nature. I respected the senior
officers’ knowledge and understanding of the research area and reflected upon the previous assumption that exclusively front-line officers would be relevant to the study at hand. I began to appreciate that by utilising various police officer positions of duty I would gain access to varied experiences, knowledge and understandings of interactions with the mentally ill and therefore decided to broaden my sample population to include any officer position that had knowledge of, and interaction with, those with mental illness in the community. Furthermore, after conducting early interviews I was able to identify particular areas of policing and the relevant job titles that I believed would provide informative and comprehensive data, utilising purposive sampling through consultation with the local police station manager to recruit participants involved in the identified areas of policing. This extending of the sample population served to ensure a wide, detailed and informative body of data was collected that was best able to explore the various practices and issues present in the field of policing the mentally ill.
Police officers on the provided list were contacted via email and supplied an information sheet detailing the study (Appendix A). Of the 6 officers contacted, 3 consented to participate in the study. The local district police station manager assisted the recruitment process by locating 8 further officers to participate, resulting in a total of 11 participants recruited.
Although this sample may appear small by traditional research standards, therefore inviting criticism of the research’s power of generalisability, Wood and Kroger (2000) note that small samples in discourse analysis do not necessarily prevent generalisability of the
Methodology and Method 85 research. They note that while traditional, quantitative research strives for generalisability
through large samples in order to examine relationships between variables of interest, discourse analysis has as its central goal the exploration of the production and function of those ‘variables’ (i.e. discourses). By adopting a Foucaultian approach to discourse analysis, the emphasis is not on discovering ‘truth’ as conceptualised as discovery of fact within the data collected, but instead explores the processes behind knowledge and knowledge production (Hook, 2001). As Wood and Kroger (2000) note “...the interest in discourse analysis is in language use rather than language users: the units of analysis are
texts...rather than participants” (p. 78). Because the focus of examination is on language and language processes instead of the individual participants and the population they are reported to represent, the sample size needs to be adequate enough to provide a
comprehensive and detailed sample of discursive processes concerning the research subject and this does not automatically require a large sample population. In fact, Coyle (1995) notes that large samples can be cumbersome to discourse analysis in consideration of the large amount of data collected that is needed to be transcribed, analysed and interpreted.
Furthermore, the number of participants recruited in qualitative research is often guided by the concept of data saturation (Fossey et al., 2002). This entails that the data collection process is continued until the point at which no new systems of meaning are emergent and the researcher is confident that continued data collection would contribute no further relevant information. It was hypothesised that between 10 and 12 participants would be required to reach data saturation in the present study. On completion of the 11th interview it was noted that no further variation or range of discourses and experiences were
86 Methodology and method emerging, therefore I was confident that no further interviews were needed in order to
conduct my analysis.
3 of the 11 participants were women and officers ranged in number of years working in the police force from just under two years to over twenty years in service. Although gender did not appear to have an effect on the nature of data collected, it should be noted here that those with shorter lengths of working experience within the police force expressed greater frustrations with the mental health system than the more experienced officers. The participants ranged in title and duties within the police force. While some officers had more direct interactions with those with mental illness, others had more indirect relationships, such as senior management officers. Whilst participants involved more specifically with front-line duties were able to comment on and discuss current interactions and concerns involved at the ‘coal-face’ - the immediate experience when attending call- outs - those who were somewhat removed from front-line duties were able to discuss more systemic and interagency relation-based systems and considerations. Positions and duties represented were: front-line officers; community constables; armed offenders squad members; youth service officers; domestic violence team members; and senior
management officers. Some officers held more than one position and others had been involved in a range of positions over their career. Because of the small and local nature of the potential participant pool, I have chosen not to provide descriptions of titles and duties in further detail for fear of inadvertently identifying those who chose to take part in the study. No further demographic data was to be included in the analysis and was therefore not collected.
Methodology and Method 87