The main proponent of the approach explains it through an example, which happens to be an IR one (Ragin, 1987). Suppose that an analyst wants to know whether strikes are successful or not. Three possible conditions are identified: a booming product market, labelled A; the threat of sympathy strikes, B; and a large strike fund, C. A ‘truth table’ is assembled showing the sixteen possible combinations of A, B and C and whether or not a strike succeeded. It can be shown whether, for example, success depends on both A and B being present but not C, or whether C alone is sufficient. This way of thinking encourages the analyst to identify logical possibilities and then go through the evidence systematically to see what combination of factors is necessary and sufficient for a given outcome. Some examples of the constructive use of QCA in workplace analysis can be identified (Roscigno and Hodson, 2004).
QCA is far from being a ready-made solution. Weaknesses include its need to simplify complex patterns into yes/no dichotomies and the difficulty of handling more than a few variables at any one time. There is also a tendency to determinism, with a condition always being associated with a given outcome. How far advocates of CR are sympathetic to it would also be a useful theme to debate. The point here is simply that it may help to think about the bundling of causal powers and about how to develop a programme of research. That is, QCA may encourage researchers to think through why certain results seem to occur in certain conditions. It may be more of a preliminary tool or means of posing new questions than an analytical device that is used directly in an inquiry. For example, the work of Mars (1982) was mentioned above. It is notable because its origin is very rich ethnographic accounts of workplace fiddles; yet it is able to provide a classificatory framework based on a simple two-by- two matrix and then to offer a set of five causal influences. Anyone developing this approach could seek out apparent anomalies or mixed cases in terms of the framework, or address situations where the causal factors seem not to operate as predicted, or both. Similarly, in relation to one of the examples given above, teamwork, if we find that certain conditions seem to lead to teams having certain effects, we may be in a position to seek cases where some conditions vary, and then to explore effects. For example, do teams thrive in apparently unhelpful conditions or fail under more propitious conditions, and if so why?
This is not the place to lay out a detailed substantive agenda. The objective has been to recover some underlying theoretical and methodological strengths in IR, not to propose specific ways in which they can be deployed. But it is not hard to see where they might be of use. In addition to the areas discussed above, IR research as recently made substantial progress in areas as different as the work-life balance, partnership, and the effects of globalization. In all of these areas, systematic comparative inquiry could be an important way forward, for example in identifying types of partnership and the conditions leading to various outcomes of the relevant practices.
Such inquiry is not easy. As mentioned above, it is rarely the case that the characteristics of potential research sites are known in advance. In international comparative research, the problem is attenuated to a degree, in that the basic characteristics of a national IR system can be readily established. But, even here, the more subtle features of how a system really works – Stinchcombe’s ‘guts of the causal processes of institutional influence’ – may be hard to define. And if we were, say, comparing teamwork in the UK and Germany, how would we establish that any observed features were truly British or German? Yet these questions are not unanswerable, as Locke and Thelen (1995) showed. The need now is to try to be more