Probabilistic causality in longitudinal studies (Q1338841)

The author considers a dynamic approach to causal analysis in longitudinal studies along the lines of \textit{E. Arjas} and \textit{M. Eerola} [see J. Stat. Plann. Inference 34, No. 3, 361-386 (1993; Zbl 0766.60048)]. The sequence of causal events is formulated as a stochastic process, more specifically as a marked point process, and the influence of a cause is expressed by the difference in the prediction probabilities of the effect given that the cause has just occurred, compared to the hypothetical situation that it did not occur. The book consists of 4 Chapters. In Chapter 1 the literature of probabilistic causality is reviewed, and in Chapter 2 the concept of a prediction process and its connection to causality is treated. Statistical properties of the prediction probabilities are considered in Chapter 3. Chapter 4 is devoted to statistical analysis and sensitivity of the method to the model choice, and outlines for further uses are dealt with at the end.