Analyses of each case study and the joint analysis of all case studies

During the analyses of case studies, I constantly contrasted interviews, meetings and documentary data with each other. In other words, I employed comparative analysis between different types of data (Mason 2002, 33). In each case study, the practical process of analysing interviews and meetings was conducted by coding the data by using the computer-aided qualitative data analysis system Atlas-Ti. There were two stages in this process. In the first stage, those instances were identified from interview

11 More detailed descriptions of these analyses can be found in each original article, under the sections on Research design or Research methods.

and meeting talk in which practitioners either referred explicitly to audit techniques, or mentioned them indirectly while discussing something else. After this descriptive coding (Gibbs 2007, 149), I moved on to the second stage of analysis to form analytic codes (ibid., 44). Here, I divided the instances into more specific codes, according to practitioners’ descriptions of the features of audit, and their own reactions and opinions towards audit techniques. In this stage, I further interpreted the codes within a conceptual framework adapted from the analytics of government (see section 2.2.2).

This second stage of data analysis can be identified as interpretive coding (ibid.,150).

My viewpoint of studying audit as a technology of government slightly shifted as the research progressed. In the early days of the research process, my “analytic gaze” was focussed on audit techniques primarily as forms of regulative governance (Foucault 2005; Foucault 1984, 179–213). This phase resulted in case studies 1 and 2 which emphasise how practitioners try to follow the norms of audit techniques. After considering audit largely as a means of control, I then became interested in looking at the possibility of practitioners acting otherwise. At this point, I started to analyse whether audit can be challenged by practitioners. This effort was founded on Foucault’s (2007) statement on the importance of raising questions about resistance, especially in such contexts where the conduct and government of individuals take place. As Gordon (1980, 255) puts it: “Human material operated on by programs and technologies is inherently a resistant material. Resistance is thus an essential aspect whenever people’s behaviour is regulated by diverse programs and technologies.” All in all, in the process of conducting five case studies my analytic focus evolved: I moved from the first two case studies on what I later call as “iron cage” view to audit, to respective case studies (3, 4) on what I will call as “discretionary space” view to audit. By the time I conducted the last case study (5), I included both views and analysed the aspects of control and discretion side by side.

Despite the shifting viewpoints in case studies, the analytic focus in each of them was on the affordances of audit techniques, i.e. the features that restrict or enable different actions of those involved. In this sense, the local affordances of EHRs and SPA offer various possibilities and obligations for practitioners. The term affordance is largely used in human-computer interaction studies to refer to the quality of an object which allows an individual to perform an action. Originating from the theory of perception by Gibson (1977), the term denotes “perceived possibilities of acting”. Even though I aimed to understand the implications of audit techniques on practitioners’ actions, no one-dimensional causal relation was constructed between a certain audit technique and a certain response. As Rose and Miller (1992) state, technical instruments in governmental practices intertwine in complex ways with a wide range of governmental programs and management models.

Synthesis of case studies

Even though I call the original five articles “case studies”, I do not utilise a traditional case study design or use the term in a strict methodological sense. In a classic case-study design, one looks at each case as a discrete part and then documents something about those parts specifically (Mason 2002, 165). Instead, in the synthesis I use the same lens to explore similar themes and problems across each case. According to Mason (2002, 166),

“you do not have to see yourself as doing “case study research” to be able to nevertheless to identify case studies, contexts or ‘wholes’ within your data set for analytical purposes”.

For me, sometimes one case equals one site (case study 1 on outpatient clinics; case study 2 on the supported housing unit). However, one case can also represent a certain combination of sites (case study 3 on outpatient and child psychiatry clinics; case study 4 on the child psychiatry clinic and English settings on child health and welfare¹²; case study 5 on outpatient clinics, child psychiatry clinic and supported housing). By exploring the relations between apparently diverse sites and different combinations of data, I have tried not to raise strong boundaries between different audit techniques, but to demonstrate the common impacts of audit under different circumstances (Silverman 2000, 85).

In the synthesis of completed case studies, the already published findings are brought up in a new constellation. I have combined the case studies according to the features of the responses practitioners seemed to employ to audit. This concerns generalisation in which the individual elements discovered in the case studies (i.e. impact of audit) are embedded in the forms of larger social organisation (i.e. professionals’ responses).

This can be called, according to Silverman (2000, 84, 85), as a comparative method which enables the comparison of cases, and views them from a wider social perspective.

Finally, responses were analysed in terms of their functions for practice, based on a certain experience of audit: Audit as an iron cage, as a provider of discretionary space, and, finally, as a mixture of the two. This division on audit techniques will structure Chapter 3.

12 See footnote 14 on page 47.