ABSTRACT
The situation of this section is again one in which I am interested in the causal relation x2 —* x\ [F], as shown at the left in Figure 1 below. As in Section 20 when I discussed reversal, I assume I am in the situation where the data at hand permit me to obtain the parameters of the reversed, probability diagram, at the right of Figure 1. The critical issue here is the relationship between the sampling that was done to obtain the data and a random sample from the population of opportunities. Since I have used P for a random draw from Q, let me use P' to denote the probability under which the right side of Figure 1 was obtained.
