Chapter II: Theoretical Framework
3.3 Justification of the selected methodological Approach – the Case Choice and Comparison
The starting point of our methodological justification will naturally refer to the importance of highlighting the unit of analysis of our research, which determines the pro and contra for the selection of a particular methodological toolkit. As we have mentioned, we need a method best for confirming the hypothesized relationship between the institutional performance (condition) and the policy - outcome. The dilemma of choice between the individualistic and collective scope of analysis should not be dismissed while taking into account the peculiarities of defence ministries as key units of our analysis. We need to have a good account of mechanisms working behind the scene, often very much person - related and unobservable, significantly affecting the course of action within the defence ministries. In this regard, the decomposition of a social structure (defence ministry) into micro - foundations cannot be bluntly rejected, but might turn to be very beneficial in understanding the individual related basics of social phenomena – the transformation of defence sector. On the other hand, it will be hard to deny that institutional transformation as any social phenomenon is of a very complex nature difficult to explain solely on the basics of the individual behavior. The distinction between individual and collective becomes a tough mission, and the international dimension makes the whole effort much more complex and difficult to model.1
3.3.1 Opting for qualitative Approach
Steven Lukes‘ statement that the institution of military – the army makes from an individual a soldier is equally valid and not less problematic than the assertion of opposite. Ultimately, we must admit that some social acts (if not all according to Emile Durkheim) have less if any individual source and possess the imperative of collective behavior.2 Following this logic the ministries of defence of Armenia and Georgia represent complex political institutions that are exposed to individual influence and serve as powerful source of influence too. Defence institutions reveal the quality of unitary actors, but are also deeply involved in the thick network of domestic national as well as international structures, through which various interest - groups and individuals act and have a significant impact on the decisions and
1 Andrew Bennett and Colin Elman, ―Case Study Methods in the International Relations Subfield,‖ Comparative Political Studies 40, no. 2 (February 1, 2007): 171, doi:10.1177/0010414006296346.
2 Durkheim, ―Social Facts,‖ 433.
58 processes. Thus we don‘t see any need to limit ourselves beforehand and adopt a single scope of analysis, significantly reducing by that the probability of alternative dimensions of causal linkages. The methodological complementarism we regard as a valuable tool for keeping the needed flexibility. Even if we‘d try to isolate a single mode of analysis to respond to a particular epistemological objective such as to describe, explain or predict, the inherent divergence of the resulting methodological toolkit would extremely endanger the ultimate validity of our claims.
The mentioned balance was less relevant for the decision we‘ve made in favor of the qualitative approach of the study. Although Evan Lieberman admits that comparative method is a weak approximation of the statistical method, the causal inferences derived from the qualitative analysis of cases across the time and levels seem to him to be very helpful to confirm ―hypothesized relationship between the institutional form and policy outcome.1‖ We agree to the argument that statistical analysis, and especially the cross - national statistical analysis is of great advantage in producing the preliminary information and the range of dependent variables to test.2 However, the nature of data available for our study, largely caused by the high sensitivity of the security and defence related issues, significantly reduces the probability of access to relevant sources and thus the probability of the full - scale statistical survey and analysis. Despite the high desirability of combining quantitative and qualitative approaches (Nested Analysis), we objectively have to resort to qualitative tools that offer distinct benefits that are even enhanced once applied in a comparative (cross - case) framework.
3.3.2 Arguments for Case - Selection and Comparison
The comparison of defence ministries of South Caucasus countries (Armenia and Georgia) that are exposed to the influence of external processes (NATO - integration) invites us in the realm of international relations and foreign policy that according to Bennet and Elman are too complex, unstructured and difficult to model and test statistically.3 This argument is persuasive and invites to accept a predominantly qualitative approach with the major effort to be devoted to cross - case analysis with the over - time comparison. Often a conventional
1 Lieberman, ―Nested Analysis as a Mixed-Method Strategy for Comparative Research,‖ 444–446.
2 Ibid., 436–438.
3 Bennett and Elman, ―Case Study Methods in the International Relations Subfield,‖ 171.
59 statistical approach is marginal to projects due to the serious deficits of the research design in observational settings.1 The contextual scope becomes very critical for the selection of cases, which in turn requires the clear understanding of its implications on research outcomes.2 Mostly, it is best achieved provided the analytical concept is able to illustrate a strong link between the hypothetical logic and the domain of argument. We believe our concept does it.
All units of our analysis (defence ministries and armed forces) represent countries (Armenia and Georgia) located in the same region that had common soviet past with no tradition of well - functioning democratic institutions and are now undergowing processes of democratic transformation, with NATO playing a significant role in national security and defence policy - making of each country. The Alliance as independent variable is less relevant for the choice of cases; however, the processes within which it applies all available elements of conditionality we are about to study, dictates a clear choice of case - selection, where similar developments took place. To avoid the selection bias caused by the inherent dependence on dependent variable, we evidently choose small number of case that share the similarity of
―causal and conceptual homogeneity.3‖ The contextual scope is, consequently, mirrored in the process of case selection, and reveals the strong relation to certain causal linkages. The heavy focus on the dependent variable prerequisited the within - case approach with a consequent application of the cross - case analysis of two countries, which we believe is best able to illustrate the elements of NATO - conditionality and their effects. Further, the empirical justification of the intentional limitation of cases by Armenia and Georgia, with exclusion of Azerbaijan, has to be explained by the political stance of Azerbaijan, that of neutral balancing between the NATO and Moscow led CSTO (Collective Security Treaty Organization), which makes the narrow comparison of only Georgia and Armenia with their radically different strategic policies more plausible and even intriguing. The plausibility of such limitation, i.e.
choice of cases that are not typical in their polical objectives to the Alliance, is largely explained by a higher benefit of finding causal relationships that such analysis would reveal as compared to cases (Azerbaijan) that show even more political contextual similarity.
Ultimately, the amount of information necessary for the chosen qualitative approach in case of Azerbaijan exposed a dangereously insufficient level, already limited by the nature of sensitive data – sources, so typical for the defence and security institutions.
1 King, Keohane, and Verba, Designing Social Inquiry, 475.
2 Geddes, ―How the Cases You Choose Affect the Answers You Get: Selection Bias and Related Issues,‖ 89.
3 Mahoney, ―Qualitative Methodology and Comparative Politics,‖ 129.
60 Touching upon the very basic question of the benefits and limitation that the method of comparison exhibits, we must at first acknowledge that political science is largely of comparative nature. Political scientists always make comparison, and that is not to say that we put equation mark between the comparison as method of variable measurement, and comparison as technique. We are aware of this distinction. The real problem, once dealing with national political systems, is the usual asymmetry between the small number of cases and large number of variables.1 It is further aggravated, as in our case, if the application of statistical data - gathering is not possible and the mere reliance of ―objective interpretation‖ of inherently subjective data generated from interviews seems problematic. The danger of over -reliance on (dependent) variables will be examined in detail in following chapter. Here, however, we take advantage of Collier‘s suggestion, by designing the comparison of cases that are matched on the dependent variable, but might reveal extreme differences in other respects.2 From our analysis we hope not only to distill a common set of explanatory factors but to discover new explanations. There is little knowledge about the given phenomenon, which is the institutional influence of NATO over the defence institutions of Armenia and Georgia. Our cases have the contextual and conceptual similarity, but vary on dependent variable exposing different outcomes of institutional (political and military) performance.
Thus, we believe that the comparison of cases we have selected is the most productive way to become deeply ―familiar with the processes under investigation‖ and to increase the validity of our conclusions.3