Discussion on Causality
Steffen Lilholt Lauritzen · Scandinavian Journal of Statistics · 2004
First, let me congratulate both authors on two fine papers which illuminate important aspects of causal inference. I have only a little to say about Professor Arjas ’ paper which specifically illuminates the aspect oftime and causality in an excellent way. I will therefore concentrate on the concepts described by Professor Rubin which seem to be more controversial, thus lending themselves directly to discussion. 1. Causal languages In the modern revival ofinterest in causal inference in statistics, a number ofcompeting formalisms prevail such as structural equations (Pearl, 2000), graphical models (Spirtes et al., 1993; Pearl, 1995a; Lauritzen, 2001; Dawid, 2002), counterfactual random variables (Robins, 1986), or potential responses (Rubin, 1974, 1978; Holland, 1986). Much energy has been used to promote the virtues ofone formalism versus the other, seemingly without coming nearer to a consensus; see the somewhat relentless discussion ofDawid (2000). Professor Rubin’s paper advocates the use of potential responses in contrast to graphical models, illustrated with a discussion of direct and indirect effects in connection with the use of surrogate endpoints in clinical trials.