In many applied fields of statistics the concept of causality is central to a scientific investigation. The author's aim in this book is to extend the classical theories of probabilistic causality to longitudinal settings and to propose that interesting causal questions can be related to causal effects which can change in time.
The proposed prediction method in this study provides a framework to study the dynamics and the magnitudes of causal effects in a series of dependent events. Its usefulness is demonstrated by the analysis of two examples both drawn from biomedicine, one on bone marrow transplants and one on mental hospitalization.
Consequently, statistical researchers and other scientists concerned with identifying causal relationships will find this an interesting and new approach to this problem.

Inhalt

1. Foundations of Probabilistic Causality.- 1.1 Introduction.- 1.2 Historical aspects of causality.- 1.3 Probabilistic causality.- 1.4 Different interpretations of probability in causality.- 1.5 Counterfactuals in causality.- 1.6 Causality in statistical analysis.- 1.7 Discussion.- 2. Predictive Causal Inference in a Series of Events.- 2.1 Introduction.- 2.2 The mathematical framework: marked point processes.- 2.3 The prediction process associated with a marked point process.- 2.4 A hypothetical example of cumulating causes.- 2.5 Causal transmission in terms of the prediction process.- 3. Confidence Statements About the Prediction Process.- 3.1 Introduction.- 3.2 Prediction probabilities in the logistic regression model.- 3.3 Confidence limits for ?t using the delta-method.- 3.4 Confidence limits for ?t fit based on the monotonicity of hazards.- 3.5 Discussion.- 4. Applications.- 4.1 Multistate models in follow-up studies.- 4.2 Modelling dependence between causal events.- 4.3 Two applications.- 4.4 Sensitivity of the innovation gains on hazard specification.- 4.5 Discussion.- 4.6 Computations.- 4.7 Further uses of the method.- 5. Concluding Remarks.- Appendices 12.