Chapter 4 Methodology
4.2 Methods (in general)
4.3.2 Focus Groups (specific)
Two opposing methodological forces exist at the point of extracting meaning from the focus group study. The principle of the focus group is to deliberately focus on what the researcher wants (Morgan, 1997). The principle behind open coding is for the researcher to extract concepts and categories of knowledge as they emerge from a sample of a real-life situation (Corbin and Strauss, 1990). Layder argues that ground up (bottom up) emergence cannot occur without some pre-existing theoretical influence (top-down) on the participants, the researcher and the situation itself. This research agrees with Layder and assumes that one can learn from concepts developed in either direction (Layder, 1998).
The focus group questions were evaluated using top down techniques. Further exploration of the textual data was expected to yield reasons for the values expressed and could be compared against the hypothesis that “trust” issues were central to the value of the rule module. While the focus group did not reflect the loose and open, real-life situation typical of a grounded study, the reasoning upon which the participants based their opinions was open. Extracting concepts from their text was found to be revealing. All audio transcriptions and written artefacts from the discussion board and quizzes were available for analysis.
Answers from the focus groups were synthesised into single, general statements for each question with ungeneralisable opinions added as qualifications. While uncoded, the generalisations were based on the same principles of concept extraction as those in grounded theory as espoused by Corbin and Strauss. (Corbin and Strauss, 1990). The twelve statements were then recirculated to participants for validation. Participants were asked if their views had been misrepresented or omitted, and whether they had any further comments to make. Because the qualifications and corrections were made by only one participant each, revision of the original general statements was not required. Had there been multiple similar qualifications for a given synthesis, revision would have been required. Participant qualifications additional to the synthesis were included in the
reported results in the next chapter. The general statements formed the core of the findings of the value of the rule module software.
The textual data were then open coded for concepts that were raised across the questions. The theoretical perspectives brought to the design of the software and the research methodology guided the search for further concepts. In particular the coding sought to check on the health of the focus group discussions, any failings or confusions with the research methods, any comments on software design and any reasoning that involved trust.
In keeping with Corbin and Strauss’s grounded method, each synthesised concept was considered provisional and had to earn its place as a result of recurrence in the data. Conceptual tags were created using TAMS Analyser software
(Weinstein, 2009) and discarded if they bore no relevance to the issue under scrutiny (Corbin and Strauss, 1990). In particular, concepts related to social banter, questioning and providing additional information were coded, but passed over.
To ensure that there was no breakdown of relationships between members and that they only influenced each other’s opinions through reason rather than the exercise of power, transcripts of the discussions were coded for signs of agreement and disagreement in keeping with both Bales’ classification of ‘decision’ interaction (Bales, 1950) and Morgan’s view of healthy focus group behaviour. Morgan demonstrates the way in which consensus and diversity of opinions are revealed, when participants are comparing and sharing experiences and perspectives with each other (Morgan, 1997). If compromise positions were suggested in this study, these were considered to occupy the ground between ‘agree’ and ‘disagree’ and were also coded. The content of these coded comments could inform any determination of undue influence, reticence or polarisation within each group. Coding responses for Bales’ ‘antagonism’ interaction sub- category (Bales, 1950) could also identify potentially dysfunctional focus groups – those with personal rather than conceptual conflicts.
The TAMS Analyser is shown below (Figure 13) with tags applied to a transcript. The content of the transcript provides insight into interpretation. One comment
can be classified in more than one way. Further, if a particular tag such as
“Commitment_Comment” is ascribed to one participant, another participant may agree with the idea without making a tagged comment. P194 in the example below says, “Just to endorse what has already been said.” This was tagged as “Agreement”, but the substance of the opinion held did not appear in the list of Commitment_Comments. Confirmation of the strength of the idea cannot rely on the frequency of tags alone. Validation of any conclusions had to come from asking each participant what they considered to be the general view of the group (for each question) and by asking participants to comment on the researcher- synthesised conclusion. Tagging was used primarily to identify comment patterns that existed across questions. The main benefit of viewing frequencies of
particular tags is to see that they came up in all focus groups (Figure 16).
Figure 13. A tagged TAMS Analyser transcript. Structural tags appear red while content tags are blue.
Figure 14. The TAMS Analyser search facility results for a single code associates it with all other structural tags (here “Commitment_comment”). Results are sorted here by Participant.