4. Considering Research Methods
4.4 Data Collection
CGTM rely on rich and sufficient data alongside credible analysis. During this section, some general concerns and questions are raised about data collection and data reliability. This outlines the types of data collected, the decision to use NVIVO software, and demonstrates how these methods address the concerns raised.
4.4.1 Ensuring Robust Data Collection
The collection of robust data is foundational to the reliability of findings within CGTM. Charmaz provides an excellent checklist to ensure rich and sufficient data is gathered (figure 4.1).384 This checklist was used throughout the three stages of research, both in planning and evaluating data collection. These questions helped guide how data was collected and the extent of data gathered.
The research design includes the use of different types of data; selection documents, statistics on church attendance, and semi-structured interviews. During stage 1, to ensure enough background data was collected, information about candidates contained in registration forms was gathered as attributes and collated in computer assisted qualitative data analysis software. This allowed a ready recall of background data which could be analysed alongside narrative accounts. During the interviews in stage 3, further information about current
ministry was gathered to add to this. In developing the research with Ministry Division, access was given to a range of participants, covering the full range of traditions and locations across the Church of England.
Detailed descriptions of a range of participants’ views and actions were gathered from the registration forms and sponsoring papers. These gave access to candidates’ previous experience and detailed narratives of their views about vocation, future ministry and belief. In addition, the later semi-structured interviews gave the opportunity to dig beneath the surface, investigating themes emerging from stage 1. Furthermore, the analysis of narratives given at interview, compared with narratives at selection, gave the opportunity to reveal changes over time.
4.4.2 Concerns over Data Reliability
Using qualitative methods, focusing on meanings expressed by participants, can act as a limitation. Townsend notes the weakness inherent in relying on participants as experts, making the research vulnerable to the effects of role ambiguity.385 Attempts were made to address this limitation in three ways. Firstly, in using archived selection papers, standardised and historical narratives from participants were collected. Secondly, the use of BAP reports
385 Townsend, 'Research Report: A Grounded Theory Description of Pastoral Counseling', 13.
Figure 4.1: A Checklist for data evaluating whether data is rich and sufficient
1. Have I collected enough background data about persons, processes, and settings to readily recall and to understand and portray the full range of contexts of the study?
2. Have I gained detailed descriptions of a range of participants’ views and actions? 3. Does the data reveal what lies beneath the surface?
4. Is the data sufficient to reveal changes over time?
5. Have I gained multiple views of the participants’ range of actions? 6. Have I gathered data that enables me to develop analytical categories?
7. What kinds of comparisons can you make between data?
and sponsoring papers provided multiple views of participants, within a standardised process of assessment. Thirdly, the inclusion of Church of England growth statistics provided reliable markers of growth and decline, and will be discussed in more depth later in the chapter. If there had been more time and resources, interviews with members of churches about their minister’s vocation and practice would have added further views and insights, but this was not possible within the timeframe and resources available.
4.4.3 Flexible Methods of Data Collection and Analysis
As outlined in 3.8 CGTM involves a process of initial data collection, initial coding and focussed coding. This is followed by theoretical sampling, secondary data collection, secondary coding, using the constant comparison method throughout to evaluate and develop theoretical categories and generate theory. Coding within CGTM allows for more flexibility, with different approaches being adopted throughout, as the research is shaped by the researcher in response to the data.386 Critics of this approach emphasize evidence of its misappropriation, in which flexible research is used as cover for unthoughtful research design.387 However, it does not necessarily follow that flexibility equates to poor research design. As outlined, initial sites of research were broadly established, while allowing for the initial data analysis to influence sampling and the interview schedule in stage 3. CGTM relies on this consistent and yet flexible method for data collection and analysis, to construct a theory grounded in the data. Active iterative strategies are employed between data and analysis, through the constant comparative method and ongoing interaction between data and emerging analysis.388 The coding of participants’ selection files allowed for the development of initial categories, which were then developed further through theoretical sampling and through subsequent coding and analysis. This process allowed a comparison between data: within selection documents in stage 1, within interview transcripts in stage 3, and also across these stages, generating ideas and giving the opportunity to construct a grounded theory of ordained vocation in the Church of England.
386 Mills, Bonner, and Francis, 'The Development of Constructivist Grounded Theory', 33. 387 Denscombe, The Good Research Guide, 110.
4.4.4 Computer Assisted Qualitative Data Analysis
Theorists take different views on the use of technology, and specifically computer assisted qualitative data analysis (CAQDAS) within qualitative research. Stern offers caution about programs, such as NVIVO, as they can hinder theory generation. Jamie Harding recommends first time researchers avoid it, as it can overshadow the development of the researchers’ intuition, leading researchers to quantify their findings.389 This can easily
happen, as NVIVO helps to group transcripts together, which can result in researchers losing sight of the details of each case, or losing sense of the narrative as a whole when scripts become fragmented.390 Having previously completed two qualitative research projects without CAQDAS, I wondered about its usefulness for this study.
There are cases of good practice391 and clear advantages in using CAQDAS. For example,
using CAQDAS can be more efficient than manual analysis, reduce the risk of human error and facilitate more complex analysis through hierarchical coding systems. Used well it can enable clearer thinking about the data, as it allows the researcher to view all parts of the data set clearly.392 In addition, as codes can be easily retrieved and reworked, it can deal with complex systems of coding.393 It is also helpful in identifying negative cases where two codes do not appear together as expected.394 Importantly it can reveal how codes are developing and the relationship they have with each other.395 Given the research design, a large amount of data was expected from fifty or more candidates, each with registration forms, providing extensive attributes alongside narratives, plus data from sponsoring papers and BAP reports, growth statistics and interview transcripts. With the sheer volume of data the advantages of CAQDAS were further emphasized. Manual analysis would be both limiting in holding together the data and restrictively time consuming. In addition, I lacked a defined study space, so any analysis procedure which could be held electronically, rather than on highlighted printouts, had practical advantages. Harding’s advice for researchers becoming confident in qualitative data analysis is to experiment and make their own
389 Harding, Qualitative Data Analysis from Start to Finish, 145.
390 Bryman, A., Social Research Methods, (Oxford: Oxford University Press, 2008), 566. 391 Stern, On Solid Ground: Essential Properties for Growing Grounded Theory, 120. 392 Harding, Qualitative Data Analysis from Start to Finish, 145.
393 Bazeley, P., Qualitative Data Analysis with Nvivo (London: SAGE, 2007), 3-7. 394 Bazeley, Qualitative Data Analysis with Nvivo 3.
judgment about whether their work is enhanced.396 Taking this advice, and having previously conducted a couple of qualitative studies without CAQDAS, I decided to use NVIVO during the pilot study and make a decision on its use for the rest of the study.