Dear all,
We are now reading and thinking about the issue concerning causal direction. According to the paper “Automating Path Analysis for Building Causal Models from Data: First Results and Open Problems” (Paul R. Cohen, etc, published in Knowledge Discovery in Databases Workshop 1993), “The estimated correlation between two nodes is the sum of the weights of all legal paths between them. A legal path goes through no node twice, and, for each node on the path, once a node has been entered by an arrowhead, it cannot be left by an arrowhead. The weight of a path is the product of the coefficients along it”. I am not sure if the estimated correlation here is just the total effects in smart pls. As for the two nodes which are linked indirectly, the value calculated in the way provided by Cohen is exactly the same as the value of total effects in smart pls. But for the two nodes which are linked directly, the two values are not the same.
So, would you please explain how we should calculate the “estimated correlation” between two directly linked nodes? Moreover, in order to calculate total squared errors, we also need to calculate actual correlation. What is the meaning of actual correlation here? Is it referring to the correlation between the constructs in smart pls?
Thanks very much!
Yalan Yan