3.2 Multiple Case Study Approach 49
49
3.2.3.1 Within Case Study Analysis
The method of data analysis chosen was the qualitative content analysis. According
to Mayring (2010), content analysis can have three different objectives: summarize,
explicate and structure. For this thesis a summarizing and structuring content
analysis was selected as an adequate method. The objective is to reduce the material
in a way as to capture the relevant content and to create a manageable amount of
data. Moreover, the structuring content analysis has the purpose of identifying cer-
tain aspects of the data or to enable assessment of the data by the means of certain
criteria (Mayring 2010). This approach ensured a process of structured analysis as
well as a systematical investigation. Furthermore, it is common to let theoretical
assumptions guide the coding process.
Three methods are possible to code data: inductive, deductive and abductivedevelopment of categories (Döring and Bortz 2016). Induction describes catego-
ries as derived directly from the material without incorporating prior knowledge
regarding theory or concepts. For the deduction on the other hand, the current
state of research plays an integral part in the creation of categories. Abduction is
similar to induction as the cognitive process starts with the data but the process to
formulate hypothesis is less structured and more creative. For this thesis a mixed
approach containing inductive and deductive methods was used.
A review of existing literature provided a full list of impacts as described incurrent literature (inductive approach). Based on this knowledge, a framework
with common circular economy approaches has been deductively developed. The
purpose of this framework is to make case study data comparable by providing a
structure to code the data material accordingly. In the cases studies, data analysis
codes for circular economy implementation were identified in each inductively.
Later, the codes were categorized according to the framework.
The coding process has been conducted with the support of the coding softwareAtlas.ti. As soon as the coding process was finalized, the cross case analysis was
conducted to identify common patterns, similarities and differences.
3.2.3.2 Cross Case Analysis
The last step in the data analysis phase is the cross case analysis with the purpose
of comparing cases in order to identify patterns. To avoid drawing false or prema-
ture conclusions during the cross case analysis it was important for the researcher
to approach the data from different perspectives. For this purpose, cases were
grouped in pairs and later in threes and fours so that it was easier to recognize the
similarities and differences between them. This helps to exceed initial impressions.
This procedure not only ensures the probability of developing a reliable new theory