Qualitative methods

To validate and further explore the variables portrayed in the Social

Attachment Model of Bushfire Preparedness, qualitative data were collected through conducting in-depth semi-structured telephone interviews with a sample of residents living in the four target Tasmanian communities.

4.3.2.1 Using narratives to assess people’s perceptions of, beliefs about, engagement with the social environment

Interviews were chosen as a method of collecting data because the present study was interested in the individuals’ perceptions of why they did or did not prepare for bushfires, and what factors influenced these decisions. Telephone interviews (see below) were conducted.

Understanding residents’ own perceptions of the reason for their behaviour is fundamental. It would be futile to base community education programs on others’ (e.g., experts, fire agencies) perspective of how residents think, as they would not address the interpretive processes that influence residents’ decision making. Furthermore, unless risk communication and education programs are based on the way residents’ reason with this information, agencies cannot expect positive behaviour change to emanate from the residents themselves.

Moreover, different people have different reasons for doing something, so observational data would not have been able to explain why people had decided to do something. In light of this position, the present study treated interviews with residents as narratives that describe their world, rather than ‘true’ pictures of ‘reality’. Due to this study’s proposition that individuals do not themselves always know why they act in the way they do, and that their behaviour is influenced by their social environment, interview responses were not be treated as providing direct access to ‘experience’ but rather as actively constructed ‘narratives’ involving activities which themselves demand analysis (Silverman, 2005). As a result, mere content analysis of the interview data would not suffice, and further analysis uncovering the reasons and processes leading to people’s attitudes and behaviour were conducted (see section 4.3.2.5).

4.3.2.2 Justification for conducting telephone interviews

Because of the geographical spread of target communities and the need to conduct all interviews within a finite period of time, telephone interviews were undertaken. This method was most appropriate for interviewing people who were widely dispersed within a time frame that ensured comparability. Telephone

are more economical than face-to-face interviews (e.g., less travel time, less time organising meetings, and can sample a wider demographic), and less obtrusive. Furthermore, telephone interviews are deemed an effective method of interviewing when the researcher has specific questions in mind (interview schedule) and when the researcher has already had face-to-face meetings with or built a rapport with interview participants during previous fieldwork (e.g., meet face-to-face during recruitment) (Berg, 2007).

Telephone interview participants for this study were recruited from bushfire information sessions held in their local communities where the researcher had invited all attendees to participate in the present research. As such, when community members approached the researcher at the conclusion of the event, the face-to-face encounter allowed the researcher to establish the legitimacy and implication of the research, and the potential participant’s role in the study. In many of these

encounters, lengthy discussions about bushfire and non-bushfire issues were

discussed, and thus prior to the interview, a rapport had been established. Arguably, this rapport played a large part in the lack of attrition from phase one interviews (n = 34), and phase two interviews conducted 12 months later (n = 34).

4.3.2.3 Interview schedule

The interview schedule was based on the schedule developed by Paton, Kelly, et al. (2006) and also used by Paton, Bürgelt, et al. (2008). The interview schedule was presented to participants in a ‘semi-structured’ interview format (see Appendix D). One of the benefits of semi-structured interviews is that the researcher has the freedom (and encouraged) to digress, to probe beyond the answers to their prepared questions (Berg, 2007). Similarly, they are permitted to adjust the language to suit the audience or particular interviewee. The benefit of this freedom is that the

interviewer, by adjusting the language of the question will ensure that the participant understands the question and thus provides the most accurate data. On the other hand, if the researcher asks questions that the participant does not understand, or uses a vocabulary unfamiliar to the participant, they may form negative attitudes towards the researcher and their involvement in the study, and thus not provide as much information or even terminate the conversation. Therefore, the added benefit of adapting the language of the question to suit the audience is the development of trust between the interviewer and interviewee, and a feeling of the participant that the researcher empathises with their situation. For example, intimate knowledge of the community or area the participant was from, and the issues they were likely to be passionate about, aided in the establishment of trust and increased the likelihood and their desire to divulge further information than that elicited by the prepared

questions. Even something as seemingly basic as adopting the nicknames

communities had ascribed to the certain groups, such as ‘firies’ for fire department people, ‘volunteers’ for volunteer fire brigade members, ‘shackies’ (shack owners), or ‘greenies’ (individuals perceived to be passionate about environmental issues), aided the establishment of a common language, and as a result the build of rapport. Unscheduled probes are another way for the researcher to understand the world and the issue at hand from the participant’s perspective.

4.3.2.4 Interview procedure

Telephone interviews were scheduled to be conducted before and at the start of the official bushfire season, between October and December (the bushfire season is deemed to be between November and March by the Tasmania Fire Service). This timing was selected so to engage with residents during the time that they should be thinking about bushfires and, if not already adopted, implementing bushfire

preparedness behaviours (Note: the researcher recognises that the telephone

interviews may have acted as prompts to adopting such behaviour which may have biased results; however, the researcher does not consider residents becoming more prepared as a result of participating in the present study as a negative result). As such, their ability to express the decision making process they were engaging in regarding the adoption of bushfire preparedness behaviours should be easier and more accurate at the time they were actually making the decision, as opposed to providing retrospective answers following the bushfire season. As classical memory theory explains, people’s ability to recall the reasons for making a decision, and most past episodic detail for that matter, is poor and inaccurate (see for example Loftus & Pickerel, 1995; Neisser & Harsch, 1992).

Prospective interviewees were contacted at the time specified by them on their consent form (day of the week, time, and sometimes specific dates). The interviewer introduced herself again and the reason for calling, and asked whether the resident was still interested in participating. If the resident indicated that they were still interested in participating but that it was not a convenient time,

arrangement for a future interview were made.

All interviews were conducted over the telephone and consent was again obtained for recording the interview. Participants were again assured that they would remain anonymous in the process, and that any individuals that they might refer to/identify (e.g., neighbour, brigade chief) in their interview would also be given synonyms to ensure their anonymity. All interviewees agreed to have their

interviews recorded. A telephone line adaptor (JNC Digital) was used to record the interviews to an 80 gigabyte iPod (Apple Inc.) with an attached MicroMemo (XtremeMac) voice recorder. As well as being stored on the hard drive of the iPod

for transcription purposes, interview files were also downloaded and stored on a personal computer.

4.3.2.5 Interview transcription and analysis

Interviews were transcribed verbatim into a Microsoft Word document. These documents were then transferred into the Computer Assisted Qualitative Data Analysis Software NVivo 9 (QSR International) for analysis. Thematic analysis was selected as the method of analysis for these data.

Thematic analysis is a widely used method of conducting qualitative analysis (Braun & Clarke, 2006). Thematic analysis is the exploration across a data set, be it interviews, focus groups, and/or texts, to find repeated patterns of meaning (Braun & Clarke). Braun and Clarke (2006) propose that thematic analysis is in fact what researchers carry out unknowingly when attempting more supposedly sophisticated or highly-regarded forms of analysis like grounded theory, but lack the required knowledge or skills to do so. In fact there seems to be a growing trend of researchers ascribing to a ‘light’ version of grounded theory, whereby they adopt the coding procedures but not the theoretical commitments. The authors argue that this is actually a very close lookalike to thematic analysis. Since qualitative data in the present study is not used to develop theory, but is rather employed to validate and further describe a theoretical model of bushfire preparedness, thematic analysis rather than grounded theory was selected as the analysis of choice.

Thematic analysis has on occasion been described as a research tool (e.g., Boyatzis, 1998) or a process performed within other ‘major’ analytical traditions (e.g., Ryan & Bernard, 2000), like grounded theory, rather than a specific approach in its own right. Braun and Clarke (2006) however, argue that it should be

considered a qualitative analysis method in its own right and advocate its flexibility as well as a relatively straightforward six-phase guide to performing thematic analysis (Table 4)

Table 4

Phases of Thematic Analysis (adapted from Braun & Clarke, 2006)

Phase Description of the process

1. Becoming familiar with the data

Transcribing data, reading and re-reading the data, noting down initial ideas.

2. Generating initial codes

Coding interesting features of the data in a systematic fashion across the entire data set, collating data relevant to each code.

3. Searching for themes Collating codes into potential themes, gathering all data

relevant to each potential theme.

4. Reviewing themes Checking if the themes work in relation to the coded extracts

(Level 1) and the entire data set (Level 2), generating a thematic ‘map’ of the analysis.

5. Defining and naming themes

Ongoing analysis to refine the specifics of each theme and the overall story the analysis tells, generating clear definitions and names for each theme.

6. Producing the report The final opportunity for analysis. Selection of vivid,

compelling extract examples, final analysis of selected extracts, relating back of the analysis to the research question and literature, producing a scholarly report of the analysis.

This six-phase guide was used to conduct thematic analysis on the qualitative data obtained from telephone interviews with residents living in bushfire risk

As thematic analysis adopts coding techniques often used in grounded theory methodology, a more detailed description of phases 1 to 5 (see Table 4) and the coding techniques commonly used is now provided.

(Phase 1) Although transcribing interviews is very time consuming, it provides an excellent way for the researcher to build knowledge of their data and is thus even argued to be a “key phase of data analysis within the interpretive

methodology” (Bird, 2005, p. 227). Full transcripts conducted by the researcher who will be analysing the data ensures that seemingly irrelevant expressions and

wordings, such as ‘umms’, repeated words, pauses, laughter et cetera, are not edited out (Bazeley, 2007). In a similar vein, incomplete sentences or breaks in

conversation can provide a deeper understanding of the participant’s belief and context. As such, all interviews conducted were transcribed verbatim by the researcher herself.

(Phase 2) Guidelines for conducting ‘initial coding’ include remaining open to exploring the data and unassuming by not applying pre-existing categories to the data. This way, coding will be as close to the data as possible. Codes represent the most basic segment or element of the raw data without losing context or meaning (Braun & Clarke, 2006). As such, codes should be simple and precise, and are best reflected as action so to avoid making conceptual leaps or adopting theories before analysis is complete. Where possible, in vivo codes, codes whose description includes words or phrases that the participants themselves used, helps preserve participants’ exact meaning and avoids removal of context (Charmaz, 2006). In this initial phase of coding, speed and spontaneity, as well as constant comparison of data to data ensure that the initial codes remain close to the data.

(Phase 3) Once all the data have been coded and collated the researcher goes through and sorts the codes into clusters, or potential themes, of similar ideas or phenomena. The result from this initial analysis is that all the codes will be collated under initial themes. Codes that do not seem to fit ‘neatly’ under other themes are collated under a ‘miscellaneous’ theme to be revisited later. This phase concludes with a collection of candidate themes, sub-themes and the extracts of text that have been coded and relate to them. These initial themes are then further refined in the next part of the analysis (Braun & Clarke, 2006). In this way, this third phase is similar to ‘axial coding’ employed in grounded theory which involves sorting, synthesising, and organising the fractured text segmented into codes during the initial coding phase into new categories and subcategories (Charmaz, 2006).

(Phase 4) Finally, once a set of candidate themes have been devised from phase three, refinement and scrutiny of each potential theme takes place. For example, the researcher may find that some ‘themes’ are not in fact themes (e.g., only one code represents it), and therefore can be collapsed into other themes or be broken down into further sub-themes. Themes should be clearly distinct from each other and the data within them cohere together meaningfully. Two techniques of refinement complete this phase. Firstly, the researcher should read all the codes that have been selected to represent the relevant themes. The researcher may find that some codes better represent other themes, represent new themes, or need to be discarded from the analysis altogether. Once the researcher is satisfied that themes developed best represent the data, and codes are allocated suitably, the researcher analyses the validity of each theme in relation to the overall data set. At this point the researcher should also revisit the data set and code any additional data within the themes that was missed during the initial coding phase (Braun & Clarke, 2006). This

phase therefore reflects the ‘focused coding’ stage adopted by grounded theory analysis (Charmaz, 2006).

Thematic analysis, utilising the rigid coding techniques also employed by grounded theory, was applied within a mixed-method action research methodology. In this way, the proposed Social Attachment Model of Bushfire Preparedness could be tested, through the use of structural equation modeling, on a sample of Tasmanian residents living in communities at risk of bushfire, and thematic analysis of the interview data obtained from a sub-sample of these residents, used to validate the appropriateness of this Model. The following chapter thus provides a summary of the rationale, construction, and distribution of the Bushfire Preparedness Questionnaire distributed to the sample areas.

Chapter Five – Assessing Bushfire Preparedness

5.1 Introduction

As outlined in Chapter Four, this study adopts a mixed-methods approach to firstly explore what motivates people to adopt preparedness and secondly, to

determine how these characteristics can be applied to develop an effective community engagement program (Chapter 9). This chapter thus presents the first step of data collection for the study which involved the development and distribution of the Bushfire Preparedness Questionnaire to a sample of Tasmanian communities identified as being at risk to bushfires.

This chapter will begin by detailing the communities that were targeted by this study. A detailed summary of the Bushfire Preparedness Questionnaire and included measures will then be presented followed by the results of the factor reduction analysis applied to the measures to ensure validity. The descriptive results of the sample then follow. The chapter concludes with a discussion of these findings and efficacy of subjecting these data to structural equation modeling analysis.