A. Philip Dawid’s contribution to the Discussion of ‘Parameterizing and simulating from causal models’ by Evans and Didelez
A. Philip Dawid对《因果模型的参数化与模拟》的讨论意见
When discussing causal inference in sequential problems, the authors only consider interventions that set their target variables to pre-specified fixed values. A more practically relevant aim, as considered by Dawid and Didelez (2010), is to assess the effect of a pre-specified dynamic treatment strategy, which details how each decision is to depend (possibly even stochastically) on previously observed variables. Can the methods of this paper be applied to this case?