rounded theory is characterised by the risk of data fragmentation. To mitigate against this I coded chunks of text (typically full participant answers) rather than partial text. I further formed a grid to account for participants coded answers under different nodes to prevent decontextualisation from their overall interview and Excel demographic data.

When coding I used the two techniques Strauss & Corbin refer to as the ‘mainstay of analysis’ (2008: 7): asking questions and making comparisons. I isolated key sections and asked sensidising questions of these. These questions helped identify the potentially different actors shaping participant’s views about austerity politics: partners, management at work, the public sector profession, the university sector. It also helped identify differences in definitions, for example in how participants understand the Southwark Council’s function. This form of questioning also identified the different consequences with which participants are acting, e.g the differing consequences of lost profit at work for P38 and P30. While within code constant comparisons brought out different aspects of the same phenomenon. For example, participant’s identification of being ‘liberal’ ranged from lapsed party affiliations to identification with the area of Peckham.

Analysis generated the following concepts, which I did further theoretical sampling from the transcripts to refine and populate:

• Positive or negative relative insulation
• Constant present & offset future
• Defence of consumer choice
• Censoring of excessive consumerism & low productivity
• Resilience to inequality (as a form of liberalism)
• Self-sufficiency from bureaucratic systems

I used theoretical questions to make connections between concepts. These included what is the relationship between relative insulation and does this affect how the consumerism of others is understood. I conducted negative case analysis finding aspects exceptional to these concepts, many of which were found in transcript P38. I found I did not reach the theoretical saturation at the concept level given the small size of the sample and degree of variation between participants. This hindered my move from concepts to categories and development of substantive theory. Though the analysis techniques supported the testing of hypotheses and answering of research questions.