• Keine Ergebnisse gefunden

4. Methodological Process: Data Collection and Analysis

The previous theoretical chapter outlined how framing can contribute to a more precise explanation of civil war onset by uncovering micro-mechanisms. It was argued that structural factors do not automatically lead to the outbreak of violence, but have to be articulated and channelled into collective action by framing agents. Therefore, framing functions as an intervening variable, that is, ―[a] variable framing intervening phenomenon included in a causal theory‘s explanation. Intervening phenomena are caused by the [independent variable] and cause the [dependent variable]‖ (Van Evera 1997, 11). Framing analyses require in-depth knowledge of social movements, their communication, and the cultural background of the targeted communities. Quantitative methods cannot provide for this, but a qualitative approach is most suitable. Therefore, a comparative case study design was chosen. Two cases of self-determination conflicts, namely in Casamance (Senegal) and Barotseland (Zambia) were selected according to the following criteria: Structural factors at the national and regional level (e.g. political stability, institutional capacity, economic performance, and identity-related variables) are relatively similar in both cases. However, the outcome – that is, the way in which conflict is waged – varies: While armed conflict escalated in Casamance, Barotse activists continue to use non-violent means to claim separation from Lusaka. Since both cases have a comparable propensity for the eruption of civil wars, structural factors cannot account for the variation in behaviour. But different protest strategies are assumed to be induced by framing, i.e. strategic communication by movement entrepreneurs and its effects on constituencies. Thus, the cases will be methodologically examined from a framing perspective.

In the following, the concrete methodological proceeding will be displayed. First, methodological considerations at the cross-case level will be outlined (4.1.). Here, the case study method will be theoretically justified and case selection will be explained. Since qualitative methods are uncommon in the field of conflict research, the merits of such an approach will also be exposed. Second, the procedure at the within-case level will be focused on (4.2.). Here, the different steps of frame analysis, i.e. data collection, data analysis as well as frame identification, and methods of measuring frame resonance, will be summarised. In this context, methodological challenges will also be taken up.

4. Methodological Process: Data Collection and Analysis

4.1. Methodological Process at the Cross-Case Level: Justifications and Procedure of a Comparative Case Study

4.1.1. The Case Study Method in Theory

In social sciences, case study designs sometimes appear to lack systematic methodological proceeding. However, this view ignores that good case studies require careful decisions and their rigorous and consistent implementation. Generally speaking, a case study is defined as an ―empirical analysis of a small sample of bounded phenomena that are instances of a population of similar phenomena‖ (Rohlfing 2012, 27; emphasis in the original; see also Blatter and Haverland 2012, 19-20). This definition implies that the scholar exemplarily focuses on a small number of selected cases in order to identify explanations that can be generalised to the larger population of cases. One can distinguish between different types of case studies, namely case-centred or theory-centred analyses (Rohlfing 2012, ch. 3; see also Levy 2008). Case-centred (also referred to as ideographic-inductive; Levy 2008) studies prevail in academic disciplines such as history, ethnology, or area studies. Their objective is to describe a unique event in a very detailed way. However, they are not necessarily inspired by theoretical considerations. This means that neither do case-centred studies apply a theory in order to examine the phenomenon, nor do they aim to contribute to existing theories or generalise their findings beyond the single case.1 In political science, this type of case study is also occasionally used in order to explore so far unknown events. Yet, the theory-centred case study approach is more prevalent in this discipline and also at the basis of the present thesis. Theory-centred case studies are ―explicitly structured by a well-developed conceptual framework that focuses attention on some theoretically specified aspects of reality and neglects others‖ (Ibid., 4). This means that the entire research process is guided by explicit theoretical considerations and their rigorous implementation. This is important as it helps to identify potential weaknesses and biases in the research process and review the validity of the empirical results (Ibid., 5). The analytical process can be summarised as follows: First, the cases that are to be studied are systematically selected from the universe of potential cases, which is defined by scope conditions. Selection criteria vary depending on the objective of the study (Blatter and Haverland 2012, ch. 2.3.; Rohlfing 2012, ch. 3). Then, the phenomena are analysed from a specific theoretical perspective by using a method (e.g.

process tracing, qualitative comparative analysis, etc.) that corresponds to the type of causality and the level of analysis. If several cases are looked at, data collection and analysis must be identical in all cases and subsequently, the outcomes are compared in a structured and focused way. Finally, empirical results are re-connected with theory (George and

1 A researcher examining the French revolution without referring to a theory or trying to compare it to other revolutions can be considered to carry out a case-centred analysis (King, Keohane, and Verba 1994, 42). However, even in these seemingly non-theoretical case studies, theoretical considerations are implicit. Otherwise, a phenomenon could not be identified as exceptional (Rohlfing 2009, 134)

4. Methodological Process: Data Collection and Analysis

Bennett 2005, 67-69; King, Keohane, and Verba 1994, 43-46; Levy 2008, 2-5; Rohlfing 2009, 139-140; 2012, 7-9, 66-80, 125-133). Theory-centred case studies can have different purposes. Not only do they provide stringent historicised descriptions and thus, an in-depth understanding of social phenomena. But they are also relevant beyond the unique or few cases studied because they lead to theoretically relevant results that can be generalised to the population of possible cases provided case selection was systematic.2 The method is useful to study how variables precisely function as well as interact and lead to an outcome.

Moreover, it can serve to build, test, or modify hypotheses as well as to probe plausibility and thus contribute to developing and advancing theories (Blatter and Haverland 2012, 68-70;

George and Bennett 2005, 110-111, 119-125, 130-131; Rohlfing 2012, 9-12, 40-43, 61-62).3

4.1.2. Criteria for Case Selection

As previously mentioned, case selection is not random or arbitrary, but has to be systematic and theory-guided according to the purpose of the study in order to provide valid results.

Therefore, the procedure of case selection that is applied in this thesis will be outlined and the cases will briefly be introduced hereafter.

The separatist conflicts in Casamance and Barotseland were respectively chosen as cases of violent struggle and non-violent protest from the population (or universe) of existing possible cases. This population included all violent or non-violent self-determination conflicts in post-colonial Sub-Saharan Africa.4 The geographic restriction is beneficial since a prerequisite of case studies is that cases are sufficiently similar in order to allow for systematic comparison and avoid distortions. The concentration on a single type of conflict, namely territorial conflicts further enhances the degree of comparability. If cases are too different, it is impossible to clearly identify variables that cause an outcome (Lijphart 1971, 688). Comparative case studies can follow different logics depending on the studied phenomena and the theoretical perspective that is chosen.5 In the present study, the research question and the theoretical proceeding require a most-similar design that is based on the method of difference exposed by John Stuart Mill who argued:

―If an instance in which the phenomenon under investigation occurs, and an instance in which it does not occur, have every circumstance save one in common,

2 Unlike in natural sciences, generalisation in social science is more limited and contingent (George and Bennett 2005, 131). Nevertheless, one can ―draw the most useful kind of theoretical conclusions from case studies, as they build on and go beyond improved historical explanations but present limited risks of extending these conclusions to causally dissimilar cases‖ (Ibid, 110-111).

3 For more details on the potential of small-N studies for theory building, see George and Bennett 2005, 119-125; Leuridan and Froeyman 2012; Rueschemeyer 2003.

4 Due to the specific historical context, self-determination movements in South Africa, which emerged mainly in townships, are not included in the universe.

5 There are, for example, most-similar and most-different system designs, univariate comparisons, or comparing relationships. For a detailed discussion of other types of comparison, see e.g. Klotz 2009;

Przeworski and Teune 1970, ch. 2; Rohlfing 2009, 134-137; 2012, ch. 3, 4.

4. Methodological Process: Data Collection and Analysis

that one occurring only in the former; the circumstance in which alone the two instances differ, is the effect, or cause, or an necessary part of the cause, of the phenomenon‖ (1843, 455).

Thus, in this design, all relevant variables are constant except for one. At the same time, the outcome differs. As a consequence, the variation in the outcome of the two cases is explained by the difference in the variable (Bennett 2004, 38-40; Blatter and Haverland 2012, 42-44; George and Bennett 2005, 153-156; Levy 2008, 10; Lijphart 1971, 687-689;

Przeworski and Teune 1970, 32-34; Rohlfing 2009, 134; 2012, 115-124; Van Evera 1997, 23, 84-85; see figure 2).

Figure 2: The logic of the method of difference