I know that conditional independence does not imply independence, and I know that the converse is also false.

Here I need to show for which events $C$ it is the case that $A$ and $B$ are conditionally independent on $C$ iff $A$ and $B$ are independent $\forall$ events $A, B$.

i.e., for which $C$ is the following true: $P(A \cap B | C)= P(A|C)P(B|C) \iff P(A \cap B) = P(A)P(B)$

I know that the answer is for all $C$ such that $P(C)=1$, but I'm stuck trying to show this.


Informally, equation $P(A \cap B | C) = P(A|C) P(B|C)$ tells you something about what happens when $C$ occurs, but says nothing about what happens when it doesn't.

$P(A) = P(A|C) P(C) + P(A|C^c) P(C^c)$; similarly for $B$ and $A \cap B$. For example, let's suppose $P(A|C^c) = 1/2$ and $P(B|C^c) = 1/2$. $P(A \cap B | C^c)$ could be anything from $0$ to $1/2$. The difference between the two scenarios changes $P(A \cap B)$ by $P(C^c)/2$, leaving $P(A)$ and $P(B)$ the same. If $P(C^c) \ne 0$, and $P(A \cap B) = P(A) P(B)$ in one of the two scenarios, it won't be in the other.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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