Abstract:We describe and contrast two distinct problem areas for statistical causality: studying the likely effects of an intervention ("effects of causes"), and studying whether there is a causal link between the observed exposure and outcome in an individual case ("causes of effects"). For each of these, we introduce and compare various formal frameworks that have been proposed for that purpose, including the decision-theoretic approach, structural equations, structural and stochastic causal models, and potential outcomes. It is argued that counterfactual concepts are unnecessary for studying effects of causes, but are needed for analysing causes of effects. They are however subject to a degree of arbitrariness, which can be reduced, though not in general eliminated, by taking account of additional structure in the problem.

What problem does this paper attempt to address?

The problems that this paper attempts to solve mainly focus on two different areas of statistical causality: 1. **Effects of Causes (EoC)**: - Research on the possible effects of interventions. This usually involves evaluating the impact of specific interventions (such as drug treatment) on outcomes (such as disease progression) in general or in individual cases. - The paper discusses different formal frameworks, such as decision - theoretic methods, structural equations, structural causal models, and potential outcome models, which are used to study the effects of interventions. 2. **Causes of Effects (CoE)**: - Research on whether an observed outcome is caused by a previous intervention or exposure. This usually involves evaluating causal relationships in individual cases. For example, in a legal lawsuit, the plaintiff needs to prove that a certain drug has caused his or her health problems. - The paper points out that counterfactual reasoning is necessary when studying Causes of Effects, but there is still some uncertainty. This uncertainty can be reduced, though not completely eliminated, by considering the additional structure in the problem. ### Main contributions of the paper - **Distinguishing EoC and CoE**: The paper emphasizes the important logical and technical differences between these two types of problems and explores different ways of dealing with each problem. - **Comparison of formal frameworks**: For EoC problems, the paper introduces multiple formal frameworks, including decision - theoretic methods, linear models, non - parametric generalization, and potential - outcome - based methods. For CoE problems, the paper discusses the necessity and limitations of counterfactual modeling and proposes using the potential outcome method for analysis. - **Examples of practical applications**: Through specific legal lawsuit cases (such as the Lipitor - induced diabetes lawsuit), the paper shows how to apply these theoretical frameworks in practical problems. ### Conclusion By comparing different formal frameworks in detail, the paper emphasizes that different tools and methods need to be adopted when studying EoC and CoE problems. Especially in CoE problems, although counterfactual reasoning is necessary, its results still have some uncertainty. The paper provides readers with a theoretical basis for in - depth understanding of statistical causality and guidance on practical applications.

Effects of Causes and Causes of Effects

Actual causation and the art of modeling

Causal Inference and Related Statistical Methods

Statistical Approaches for Causal Inference

Reconciling Causality and Statistics

Causality without potential outcomes and the dynamic approach

The Cause-Effect Problem: Motivation, Ideas, and Popular Misconceptions

Statistics and Causal Inference

Remarks on the analysis of causal relationships in population research

Evaluating causes of effects by posterior effects of causes

Causality from the Point of View of Statistics

Causal inference: critical developments, past and future

Causality and Association: The Statistical and Legal Approaches

Towards Bounding Causal Effects under Markov Equivalence

Is causation deterministic or probabilistic? A critique of Frosch and Johnson-Laird (2011)

Causal inference for psychologists who think that causal inference is not for them

Confounding caused by causal-effect covariability

Missed Causes and Ambiguous Effects: Counterfactuals Pose Challenges for Interpreting Neural Networks

Practical Implications of Modes of Statistical Inference for Causal Effects and the Critical Role of the Assignment Mechanism

Causes and Explanations: A Structural-Model Approach --- Part 1: Causes

Retrospective causal inference with multiple effect variables