Suppose I have two statements:

Statement 1: Random variable Z is the common cause for random variable X and Y (Z causes both X and Y)

Statement 2: Random variable X and Y are (conditionally) independent given Z.

What's the relationship between Statement 1 and Statement 2? Does Statement 1 imply statement 2, or does statement 2 imply statement 1, or they have other relationships?

Example: Z could be the grade level(or age) for a primary school student, X could be his height, Y could be his math ability. Thanks!


1 Answer 1


Statement 1 is a scientific statement rather than a mathematical relationship. The idea of one random variable "causing" another doesn't have a strict mathematical definition, rather the role of the mathematician or statistician is to posit a mathematical model that captures the relationship in a way that is relevant to a particular scientific problem. For example, a mathematician might suppose that X and Y both have functional relationships with Z but not with each other, and that would capture the scientific idea of "causation".

Statement 2 is a mathematical statement, and it is one sensible and useful way to translate the first statement into mathematical terms. It is a very strong statement because it implies that Z accounts for all possible dependencies between X and Y. There cannot be any other common causes of X and Y that are not mediated by Z or correlated with Z. On the other hand, Statement 2 is more general than Statement 1 in the sense that X and Y could be conditionally independent given Z without being direct functions of Z.

  • $\begingroup$ I don’t think your assertion that any causal model where Z is a common cause for X and Y is plausible. In particular, even within the standard model of causality, if Z’ were another common cause between X and Y, then X would not in general be conditionally independent of Y given Z $\endgroup$ Commented Oct 30, 2022 at 2:06
  • $\begingroup$ @stats_model Yes, obviously. OP's Statement 1 says "the common cause" rather than "a common cause" so I assumed that no other common causes are present. Anyway, I have deleted the last paragraph of my answer so that it doesn't depend on such a strict interpretation of OP's words. $\endgroup$ Commented Oct 30, 2022 at 3:17
  • $\begingroup$ @GordonSmyth Thank you very much! This is very helpful! $\endgroup$ Commented Nov 2, 2022 at 21:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.