This chapter shifts from theory to practice in qualitative causal inference. It addresses the central challenge of how to construct causal graphs that can support valid unit-level causal inferences. Building on the framework of bias-free graphs, hypothetical interventions, invariant causal mechanisms, and event-history maps, the chapter explains how to design causal graphs with credible front-door, back-door, and side-door paths. The first section examines the origins of causal graphs, arguing that while qualitative traditions emphasize alternative hypotheses, they rarely yield structured causal paths. To overcome this gap, the chapter introduces the method of iterated modularity: constructing front-door paths incrementally through sequential pair-wise analyses of adjoining nodes. This approach allows complex causal processes to be decomposed into manageable segments while maintaining methodological rigor. The discussion also engages with objections that causal relations in the social and political world are too complex to yield neat causal paths, countering that temporal sequencing and careful distinction between variables and events can often impose useful structure. The chapter concludes by discussing the role of critical junctures and event-history maps, clarifying how they contribute to evaluating the empirical implications of causal graphs without succumbing to common inferential pitfalls.

This chapter introduces causal graphs as the foundation of qualitative causal inference. The approach begins with a bias-saturated graph that incorporates potential sources of systematic error, including endogeneity, spuriousness, confounding, and measurement error. Through theoretical and empirical testing, researchers assess whether this graph can be simplified into a bias-free representation, enabling more valid unit-level causal inferences. Beyond their methodological role, causal graphs offer clear communicative and heuristic advantages: they reduce ambiguity in visual representations, provide a shared language across methodological traditions, and require systematic reflection on potential causal relationships. The chapter proceeds in four stages. First, it introduces the basic elements of causal graphs and their role in defining paths relevant for causal inference. Second, it outlines three path-specific criteria – front-, back-, and side-door paths – that identify sources of bias and establish conditions for valid inference. Third, it situates qualitative causal inference relative to the detective model and identification-based inference, highlighting the hybrid nature of the proposed approach. Finally, the chapter develops two new concepts: hypothetical interventions and invariance as tools for causal interpretation, and a reconceptualized view of data-generating processes that reconciles probabilistic graph theory with qualitative traditions.

This chapter introduces event-history maps, the third conceptual innovation of qualitative causal inference alongside hypothetical interventions and invariant causal mechanisms. Event-history maps link the theoretical domain of causal graphs to the empirical domain of observed events, case studies, and hypothesis testing. They represent the realized values of the random variables in a causal graph, taking the form of a sequence of events structurally isomorphic to the graph. While causal graphs depict variables and potential relationships, event-history maps depict specific outcomes and temporal sequences. Only when paired with a bias-free causal graph can such maps credibly represent chains of causal relationships. The chapter illustrates how event-history maps can be used to strengthen within-case analysis. It reconstructs implied graphs and maps from published studies to show how existing research often reverts to the detective model, missing opportunities to test additional empirical implications. By explicitly pairing causal graphs with event-history maps, researchers can probe causal claims more effectively and avoid invalid inferences. The chapter concludes with a detailed example of evaluating the three path-specific criteria of valid unit-level inference, showing how event-history maps serve as essential tools for empirically validating causal graphs and improving the rigor of qualitative causal analysis.

So far, we have focused on predicting an outcome Y from a set of features X. In this chapter, we turn to causal inference where we ask: What would Y be if we set X to a particular value x? This question concerns the distribution of the outcome Y after some hypothetical intervention. Answering such causal questions requires new tools and stronger assumptions than prediction.

The chapter shows that more sophisticated difference-making theories of causation that draw on so-called causal models can accommodate mental causation too. Causal modelling theories invoke more complex relations of difference-making than the simple principle about causation that was used in previous chapters. These relations of difference-making are represented by causal models. Accommodating mental causation – either in the non-reductive physicalist case or in the dualist case – calls for some heterodoxy in model-building. If the heterodox models are allowed, however, they prove useful not merely for explaining mental causation, but also for capturing the distinction between higher–level causes that are explanatorily relevant and higher-level causes that are not. The chapter also discusses the interventionist theory, an especially prominent member of the causal modelling family, in relation to mental causation.