CHAPTER V: RESEARCH METHODOLOGY
5.4 RESEARCH TECHNIQUES AND PROCEDURES
5.4.4 Data Analysis
The collection of data in itself does not make any sense if it is not properly analysed and interpreted to reflect its meaning, which is the focus of this section of the work. The analysis of the data gathered for this study could be done in one or two ways, which are quantitative and qualitative data analyses; but the qualitative approach is chosen for this study based on the earlier discussions of the research design. This qualitative analysis involves the getting of meanings
154 by grouping of responses (Saunders et al., 2007) as a way of identifying, examining, comparing and interpreting of patterns and themes that come out of the data (Sekaran and Bougie, 2010).
In order to properly analyse qualitative data, one of the steps involved is to codify the data by picking out themes or words, reducing it to the needed quantity and displaying the outcome (Hair et al., 2007). The codification process could be done by choosing what will be known as the coding units, which could be words, sentences, themes or paragraphs and this leads to a reduction of the huge data gathered to a manageable size (Sekaran and Bougie, 2010). This could have been done in one of two ways, either by generating them out of the data inductively or having them deductively derived from the literature prior to this stage (Saunders et al., 2007). However, for this study the identification of themes was done iteratively between the literature and the data gathered, with the former used as guide to discover out of the data other themes that might be important.
These themes were identified by the use of in-vivo codes that emerged from the terms used by interviewees as well as in-vitro codes created by the researcher with guidance from the literature (Strauss, 1987; Carcary, 2011). The process of codification was concluded with a final decision on the 17 themes deemed important for the research out of the total 27 themes initially generated by a merger of some themes that shared a lot of similarities, with the choices being guided by the reference to such themes by both the literature and the participants involved in the study. These led to the establishment of relationships between certain key themes identified, as the researcher aimed to
155 test some propositions made in the course of the study. There were a number of options available to the researcher for the analysis such as narrative analysis, discourse analysis, analytic induction, template analysis, grounded theory, content analysis, conceptual analysis, relational analysis and thematic analysis (Saunders et al., 2007; Bryman and Bell, 2007; Sekaran and Bougie, 2010).
However, discourse analysis was chosen for the analysis of the data gathered as a way to aid the gaining of deeper meanings into the discussions that ensue in the form of interviews as it is deemed to deal with broad areas of social reality (Silverman, 1993). The focus of discourse analysis is on the achievement of talk, in terms of the actions embedded in such talk and how they affect other persons involved (Wood & Kroger, 2000). Hence, it is the analysis of discourse to see how it contributes to the production of social reality, even as it explores the relationship between discourse and reality (Phillips and Hardy, 2002) with the focus being on language, the structure of talk, text or interaction (Spencer, Ritchie & O’Connor, 2003). These discourses are made up of stories, narratives and symbols that connect different dynamics of organizations leading to our understanding of reality (Chia, 2000, cited in Phillips and Hardy, 2002, p.15). The discourse in this study revolves around the narratives and stories of the different stakeholders that are involved in the study within the context of the study (Phillips and Hardy, 2002), leading to the presentation of the contextual issues earlier discussed in Chapter II to see how they have contributed to shaping the discourse. This study is mainly concerned with the construction of interactions between stakeholders in the Nigerian oil industry as a way of
156 understanding the relationships existing therein, with this particular discourse identified via the literature. This bears in mind that the statements made by individuals cannot be analysed without a consideration of the context within which they are made because such context explains the specific events leading to these statements (Wood & Kroger, 2000).
The use of discourse analysis is based on its focus on the constructive ability of the words or languages employed by participants, especially considering this study’s philosophical stances of subjectivism (Saunders et al., 2007) and critical realism (Bryman, 2008). It could be argued that this is a linguistic method of analysis, but its use anticipates answers to social questions and not linguistic ones, despite its linguistic connections (Potter & Wetherell, 1987). The place of discourse in human existence cannot be over-emphasised as it is deemed to rule all aspects of our lives through the actions that are associated with it (Potter, 2004). This analytical choice also helps us with an understanding of how the interactions between stakeholders affect or are affected by these discourses that they are engaged in as they relate with each other (Heller, 2003). As a result, it is deemed to enable the researcher to explore the relationship between the discourse and the reality being studied (Phillips & Hardy, 2002) as the discourse that ensues from the interviews can further enhance the exploration of social reality, which in this case is the relationship between stakeholders. There is also the relationship between a particular topic or discourse and other discourses with regards to how they impact each other (Bryman, 2008). In this case, the discourse of the relationship between stakeholders in the Nigerian oil industry is deemed to have impacted on
157 community development and Corporate Social Responsibility (CSR) issues and vice versa. With the realisation that relationships are mostly what the actors or parties make of them, this method of analysis is deemed capable of making meaning of the relationships under study. In other words, the various activities that the oil companies and the government engage in as a way of interacting with the host communities such as CSR activities, MOUs and GMOUs (as well as their processes) could be referred to as discourses (Poynter, 2011).
There are different varieties of discourse analysis with Discourse Analysis in Social Psychology (DASP), Conversation Analysis (CA), Post-Structuralism (PS), Pragmatics and Critical Discourse Analysis (CDA) being a few of them that could be adopted for this study (Wood & Kroger, 2000). However, this research takes to what Singer and Hunter (1999) referred to as Thematic Discourse Analysis which explores “common threads and inconsistencies embedded in the narratives” in the text (p.66) as supported by previous works (Potter & Wetherell, 1987; Burman & Parker, 1993). This involved the use of broad thematic levels of analysis following the examples of Tracy and Carjuzaa (1993), Wood and Rennie (1994) and Singer and Hunter (1999), all of which involved interviews with respondents. The analysis involved the exploration of variations within and across texts, the use of rhetoric by respondents and looking for accountability in the texts analysed (Potter & Wetherell, 1994). The choice of this approach is influenced by the size of the sample involved (21 semi-structured interviews, 180 surveys) just as authors (Wood & Rennie, 1994; Singer & Hunter, 1999) had previously used similar sample sizes. Also, this choice is based on the focus of the study being on the deeper meanings
158 contained in words and phrases used by respondent which were best reflected by the themes identified. These were mainly in the function, construction and variation of such words and phrases in ways that were deemed to achieve certain goals in the text (Potter & Wetherell, 1987). The table below shows the elements of TDA used in this study as well as the authors from whose works they were adapted to fit this study.
Element of TDA used
Previous Literature Remarks
Common threads and inconsistencies
Potter & Wetherell, 1987; Burman and Parker, 1993; Wood and Rennie, 1994; Singer and Hunter, 1999
These common threads aided an identification of themes that ran through most of the discourses of the topic.
Broad thematic levels
Tracy and Carjuzaa, 1993; Wood and Rennie, 1994; Singer and Hunter, 1999
These broad themes were informed by more specific themes identified from the data, but the broad ones are used to link the discourses with the topic under study. Variation Potter & Wetherell,
1994; Singer and Hunter, 1999
This focuses on differences within and across interviews.
Rhetoric Potter & Wetherell, 1994; Singer and Hunter, 1999
This concerns the argumentative use of words and phrases, even repetitively by respondents as a way of emphasizing the importance of the discourse.
Accountability Potter & Wetherell, 1994; Singer and Hunter, 1999
When respondents refer to some position or experience just to underline their credibility and qualification to talk about the issues. Also, when there is an indirect appeal to the interviewer for support of their points of view.
Extreme Case Formulation
Pomerantz, 1986 The presentation of situations with some form of extremity, with such words as never, none, every, zero, etc.
159 Thematic Discourse Analysis (TDA) could be seen as being similar to Qualitative Content Analysis (QCA) as both methods explore underlying meanings and themes embedded in talk and text. It has even been argued that both different approaches to one method, with just variations in terminologies as they are viewed as being part of Content Analysis (Krippendorff, 2004). However, George (2009) argues that they are different methods as QCA is focused on the frequency of occurrence of the specific themes or words that interest the researcher concerned, which makes the method sometimes resort to statistical techniques. Another key difference that makes TDA more preferable for this study is its emphasis on the constructive abilities of the discourses that social actors undertake with each other which the latter does not do (Bryman & Bell, 2011). Also, the chosen method was iterative but not as much as QCA is presented as being, especially in terms of the movement between data sampling, collection and analysis (Altheide, 1996, cited in Bryman & Bell, 2011). For this study, the data deemed sufficient was collected once only within a specified time, and the analysis conducted iteratively.
The use of broad thematic levels (Tracy & Carjuzaa, 1993; Wood and Rennie, 1994; Singer & Hunter, 1999) aided the synergy of the specific themes earlier identified as a way of gaining better understanding of the relationships between stakeholders in the industry. Also, the use of this method makes it possible to interpret the different actions carried out by the words used by the respondent as an indication of stakeholder relationships in the region. The analysis was aided with the use of the NVIVO software which was set up with a pilot analysis of part of the research data, leading to the generation of categories that were
160 coded. These initial codings were reviewed by merging some of the codes together in order to reduce the number of themes to be discussed in the study before they were analysed, with the analysis involving validation by colleagues to be sure it was rigorously done. Following the validation by colleagues, another round of analysis was undertaken to further ensure that the data had been subjected to a rigorous process leading to the finally used analysis as presented in the next two chapters.
The data was loaded onto the NVIVO data management software for coding and management for analysis. The coding process started out with a list of literature guided themes such as Stakeholders, Stakeholder Relationships, Contracts, Agreements, Resources, Development, Interests, Influence, Government, Engagement, Communication, Trust, License to Operate, Reaction to Stakeholder Relationships and Dependence which aided the initial coding undertaken. Other themes that came out of the data are Negotiations, Host community Participation, Accountability by Stakeholders, Traditional Rulers, Responsibility, Gender Issues, Conflict and Violence. These themes (in- vivo and in-vitro codes) and others not highlighted above resulted in a total of 27 themes, which were further merged leading to 17 themes as reflected by the table of Free Nodes below. The coding further involved annotation of thoughts and insights by the researcher as the interviews were read and coded (see Appendices VIII & IX).
161
Name Sources References
Communication 14 29 Contracts or Agreements 21 98 Dependence 16 27 Development 18 59 Engagement 5 7 Government 20 76
Host Community Participation 16 57
Influence 9 10
Interest 6 11
License to Operate 3 3
Management of Stakeholder Relationships 15 31
Negotiations 10 25
Reaction to Stakeholder Relationship Management 15 56
Resources 9 13
Stakeholder Relationships 20 47
Stakeholders 20 87
Trust 9 14
Table 5.5: Free Nodes of NVIVO First Coding (Extract)
It is noteworthy that some of themes overlapped with others, as certain statements by the respondents referred to issues touching on more than one theme. For instance, the discourse of Contracts and Agreements were undertaken with reference to Negotiations, same as Communication and Trust, Communication and Reaction to Stakeholder Relationships.
As a follow-up to the above coding schedule, these Free codes or nodes were further coded as Tree codes or nodes that were grouped together under major themes that covered broader areas of discourse in answering the research questions set at the beginning of the study. This final coding based on the major themes is reflected by the table below.
162