A solid QCA does not end with the analytic moment. Researchers must make several analytic decisions at various stages in the analysis, some with more confidence than others. Researchers might also be confronted with data that are structured in analytically relevant ways. For example, cases might group into different geographic, substantive, or temporal clusters, or there might be relevant causal dependencies or sequences among conditions.

This chapter introduces the different robustness and diagnostic tools available in R to assess QCA results. It enables the reader to investigate to what extent their QCA results are robust against equally plausible analytic decisions regarding the selection of calibration anchors or consistency and frequency cut-offs. We present possibilities to assess robustness in R. Moreover, we introduce tools for cluster diagnostics and discuss strategies for dealing with timing and temporality, including ‘coincidence analysis’ (CNA).

Learning goals:

- Basic understanding of different approaches to diagnosing and assessing QCA results.

- Familiarity with how the robustness of QCA results to different analytical decisions can be assessed.

- Familiarity with proposals on how to assess QCA results in the presence of clustered data.

- Familiarity with how to model sequences and causal chains in R.

This chapter discusses the relation between causation and moral responsibility. We generally hold adult human beings morally responsible for their actions, yet those action are also events in the natural world, enmeshed in causal chains that extend backwards in time long before the agent’s birth. If the causes in those chains necessitate their effects, it would appear that we must either give up the view that humans are morally responsible for their actions, or embrace the paradoxical view that humans are morally responsible for actions necessitated by events over which they have no control. Tuozzo argues that Aristotle’s causal theory avoids this dilemma by recognising two distinct types of causal chain or nexus. In one of these, the links between cause and effect are indeed necessary, from beginning to end. But chains of this sort are necessarily finite, with a definite beginning and end. Each of these necessary, finite causal chains are also enmeshed in a different sort of causal nexus, one that does extend indefinitely into the past. But this sort of indefinite causal chain is possible only because it contains links that are not necessitated. This enables Aristotle to account for moral responsibility by locating the necessitating cause of a human action in the agent herself. Nonetheless, Tuozzo concludes, Aristotle’s theory does have the paradoxical implication that, although the state of the world at a given time does not necessitate all subsequent events, a complete description of it would, in principle, allow all subsequent events – including human actions – to be predicted.