Methodological Approach

The methodological approach applied in this study is content analysis (CA) with a discourse analytic (DA) approach (Hardy et al., 2004). CA stems from a positivist tradition and consists of the formation of analytical categories which are used in constructing a scheme for coding, which is then applied to textual data (Hardy et al., 2004; Yin, 2014, p.254). In contrast, DA is

grounded in a social constructivist tradition and emphasises the socially constructed, intertextual quality of texts (see Chapter 4). As a methodology, discourse analysis covers a wide range of tools for textual analysis ranging from the highly detailed micro-level analysis of texts, to more societal level interpretative analyzes (more on the varieties of discourse analysis can be found in Alvesson and Karreman, 2000). This study aligns towards the latter form of interpretative discourse analysis, while Article I applies a micro-level textual analytic method. While having different epistemological assumptions, DA and CA are not opposites, but rather CA can be seen to benefit from the insights of DA (Hardy et al., 2004). As Fairclough (2003, p.16) has noted, it is impossible to make texts transparent by applying analytical frameworks, and the actuality of the text is determined by the perspectives and theories which are utilised to analyze it. In this study, CA is used to identify prominent themes in the data, while DA provides a theoretical lens to give interpretation to the emerging themes.

CA in the three articles was carried out by coding the interview and document data using the computer assisted qualitative data analysis package, ‘TAMS analyser’ (Weinstein, 2006). This software enables creating a hierarchical structure of codes for the textual material, and analyzing these codes through different analytical tools. The purpose of the coding was to identify prominent themes from the data, by counting the occurrences of the codes in the texts (Hardy et al., 2004). The theoretical framework introduced in Chapter 4 was used to inform the formation of the coding schemes.

In Article I, the coding scheme was based on Leeuwen’s (2007) framework of discursive legitimation strategies (see sub-chapter 4.1). This framework draws from the systemic functional linguistics of Halliday (2013). Coding process involved reading selected documents through and identifying sections which were related to the ‘purposes or goals’ of the project. These sections were broken down into clauses and sentences to which a coding scheme constructed from Leeuwen’s theory was applied. These codes were counted, and broader analytical themes were formed on the basis of the frequency of occurrences. In Article II, the coding scheme was informed by the theoretical framework of Lawrence et al., (1999) on the discursive

dynamics of collaboration. Coding process involved identifying passages in the interviews regarding issues (problematisations which the projects intend to respond to), and the interests which participants have towards these issues (see 4.2). These passages were were coded inductively, i.e. the codes emerged from data (Hardy et al., 2004; Yin, 2014, p.254). In Article III, the coding scheme was informed by the ‘positioning triangle’ (Van Langenhove and Harré, 1999) which consists of three angles: speech acts, story lines and positions. Coding was done by searching interviews for speech acts which indicated an attribution of rights or duties to the participating stakeholders.

Next, the identified speech acts were grouped together into story lines which reflected the broader context of the collaboration (see sub-chapter 4.3).

Themes which emerged as the result of the coding process were

addressed as analytical sub-sections in the articles, utilising insights drawn from the DA theory. These insights manifested in two analytical tasks, which both stem from social constructivist assumptions of DA: 1) social reality constructed through meaningful interactions between people, and 2) the importance of interrelated bodies of text and discourses for the construction of social reality (see Chapter 4). Hence, the discourse analytic interpretation of the themes focused on how language is used by people to construct the social reality, and which discourses are applied in the process. As a disclaimer, discourse(s) can not be fully captured by analytical methods, and in this sense DA does not consider textual data as a reflection of reality, but only as a sample of it (Phillips and Hardy, 2002, p.74).

In Article I, DA was utilised to analyze how language was used for legitimative purposes. The analysis drew insights from the critical variant of discourse analysis, which focuses on the critique of hegemonic

discourses and genres which bring about inequalities, injustices and oppression in contemporary society (Van Leeuwen, 2006; 2008). Hence, the aim of the analysis was to uncover the relationship between applied legitimation strategies and hegemonic discourses in the field of regional development in Finland. Article II instead focused on the way the steering group representatives used language in the collaboration, in order to reach common understandings regarding central dimensions, i.e. issues,

and interests (see sub-chapter 4.2). The central aim was to analyze how hegemonic discourses present in the discussions on one hand supported the formation of the common ground around these dimensions, and on the other hand, constrained the discussion by limiting the space for discussion. Finally, in Article III, the focus was on how the steering group representatives utilised speech acts to attribute positions (rights and duties) to themselves and other stakeholders in the collaboration, within co-created narrative structures. While not drawing directly from the DA tradition, applied positioning theory is also based on social constructivist epistemology, and therefore the central assumptions are compatible.

Qualitative methodologies such as discourse analysis have a different form of evaluating validity and reliability compared to that understood in the positivist tradition, in terms of how accurately the method applied represents studied phenomena and how consistent or repeatable a study is (Yin, 2014). In the DA tradition, the reliability of the study is evaluated rather than the possibility of replication, in terms of how the results of the study are comprehensible and plausible to the reader. In this sense, reliability is the measure of how well the researcher explains the procedure of how the findings of the study were arrived at. To this end, this sub-chapter has attempted to give a clear explanation on how the analysis of the study was carried out. Validity in terms of the DA tradition measures how well the patterns of meaning that the study evokes are constitutive of reality (Hardy et al., 2004). While this is a matter of judgement on the part of the reader, the validity of this study has constantly been put under stress by presenting the research process and its findings in national and international conferences, and reflexively applying the received feedback.

6 Results

This dissertation consists of three article publications which look into the stakeholder collaboration process in regional place marketing projects from three interrelated perspectives. The themes of the publications are theoretically connected to the discursive model of collaboration (Lawrence et al., 1999), which describes the essential stages of collaborative processes from a discursive perspective. The individual themes of the articles also have their own theoretical aspirations, which is most evident with Articles I and III which are not directly connected to the model of Lawrence

et al. (1999), but instead draw from other social scientific theories of discourse and communication, in order to extend the ideas of the model.

Hence, the study offers a more nuanced version of the model. Indeed, the model is considered due its general nature as a good base for theoretical experimentation, and applying different theoretical ideas concerning communication and discourse in the context of inter-organizational collaboration.