• Ei tuloksia

3 Theoretical perspectives

4.4 Thematic narrative analysis

As with the process of narrative inquiry, narrative analysis also has a diverse set of methods. In thematic narrative analysis, the story is kept intact but the focus is on the content, looking for themes across the narratives, and not on the process of the telling itself or the structure or form of the narrative (Riessman 2008, 53-58;

74). Narrative theorizing (narrative analysis) is made from the case, keeping the stories intact instead of segmenting them (Chase 2005, 663; Riessman 2008), and theorizing may also be conducted across cases as in paradigmatic theorizing (analysis of narratives: Polkinghorne 1995, 21). According to Polkinghorne (1995, 10) the analysis of narratives with paradigmatic reasoning builds conceptual networks across cases, and looks to establish relationships between the established categories.

Valkonen (2007, 73) argues that the crucial methodological difference between narrative analysis and the analysis of narratives is that narrative reasoning interprets and finds implicit narrative structures created in the narratives, whilst the analysis of narratives is based on what is explicitly visible in the data. In this study both means of reasoning were utilized, and as Chase (2005) suggests, we can analyse dimensions across narratives, without losing the context within the narratives, and this was possible due to the limited number of narratives involved.

In narrative methods, the interpretation is central (Riessman 2002, 218), and thus ties it to the hermeneutic tradition. However, one does not look for a single interpretation of a narrative, but rather assumes that the narrative can have different interpretations.

The knowledge we have about the world and ourselves is constantly being reshaped, and a narrative is an account told from one point of view in one situation (Riessman 2008, 187). Therefore the task of narrative research is to produce not one true interpretation, but an authentic range of perspectives of reality (e.g. Valkonen 2007;

Johansson 2005).

The interviews were arranged to form a chronological narrative of the adoption process. Most of them started with the motives for entering the pre-adoption services many years before actually doing so, and ended in the present when the interview was conducted, and including the time after the adoption had taken place or after the process had been terminated. Firstly the narratives were individually analysed, making storylines and performing an initial coding of interesting elements

and their relations within the narrative (Riessman 2008). The narratives were also summarized around the thematic meanings which emerged, and these summaries were then used to bring the whole narrative back to mind, in forming interpretations of shorter sequences of the interviews, as well as being used cross-case in consecutive analysis. In finding common elements and theoretical frameworks to explain the experiences, the analytic reasoning was extended to cover the broad subject area being considered (Polkinghorne 1995). Later, when the data was analysed against the chosen theoretical frameworks, the completed narratives formed a picture of the contextual and social situations of the narrator. These were then utilized in the synthetic interpretation, and analysis was also made on the thematic sequences of data which were compiled (Riessman 2008, 74). The computer software Atlas.ti was used for the initial coding of the shorter sections of the narrative data. This was then complemented with more traditional cut-and-paste techniques in a Word processor during the later stages of analysis. The excerpts used from data were translated into English after the analysis had been completed, and feature in the appended articles.

Inductive reading and analysis highlighted the importance of both emotions and power, and theoretical concepts (emotion, power and strategic interaction) were

derived from the data. According to Riessman (2008, 74), a thematic narrative analysis can also be guided by prior theory and so an abductive phase followed (Tavory &

Timmermans 2014). In this approach, empirical data is worked with in relation to a broad and diverse range of theories and types of knowledge. The analysis proceeded similar to the coding process described by Tavory and Timmermans (2014, 54) from open coding to axial coding, where “promising” themes are coded across the data, the conceptual dimensions are identified and the variations accounted for. The analysis and coding was conducted simultaneously with the gathering of more data (Tavory

& Timmermans 2014, 125), and hence resulted in slight changes being made to the additional questions in the interviews.

The sections of the interviews where the narrator talks about emotions or where the account was interpreted as carrying emotional content were coded in regard to corresponding emotions. These were each treated as equally important due to gendered (Fischer et al. 2004; Brody & Hall 1993) and individual differences in emotional experiences and expression. The accounts of emotional experiences were treated as analysis units (Riessman 2008, 60-61). The variation accounted for all of the primary emotions (sorrow, anger, joy and fear) and featured many combinations.

As the negative emotions were seen to be the most interesting and influential on the experiences within the institutional context, they were chosen for further in-depth analysis. This is supported by Kemper (1987) who suggests that negative emotions are the most influential in social interaction. As the concept of emotion is seen as being socially, contextually and relationally embedded, the complete narratives were kept attached to the coded excerpt throughout the analysis and used in the interpretation.

The unit of analysis thus shifts between the excerpts and the complete narratives. For example the analysis of fear was made on the basis of explicit accounts of fear and also accounts that I interpreted as stemming from fear. These were then re-interpreted in the light of the whole personal narrative, for example in the light of infertility or hardships during the adoption process, and served as an interpretation of what the sense of fear might be associated with in that particular narrative.

Power was utilized as a concept in understanding and explaining the client position of the prospective adoptive parents, as well as its emotional consequences. It was conceptualized as controlling and supporting institutional and professional practices as perceived from the client position, and furthermore as power negotiations through strategic interaction by prospective adoptive parents in the institutional setting. This was achieved by coding the actions of actors in the context, and also by interpreting their associations to emotions and relationships in the setting, again taking into account the whole personal narrative. According to Landman (2012, 32), narratives offer the possibility to uncover subjective experiences and emotions in relation to power dynamics, and thus the “institutional constraints” described by Landman (2012, 32) have been analysed as controlling practices.

In the analysis of strategic interaction, the data was coded descriptively identifying all sequences of narrated actions or thoughts about actual, intended or considered actions of the prospective adoptive parents within the institutional setting. This was then followed by a coding which utilized the conceptual frameworks of Goffman (1959;

1970; 1983): dramaturgy, strategic interaction and expression games. Emotion and power were running as intertwined concepts through every analysis and constituted important interpretative frameworks throughout the study.