For a random vector $X=(X_1, ..., X_n)^\intercal$ the expectation value can be written as $\mathbb{E}[X] = (\mathbb{E}[X_1], ..., \mathbb{E}[X_n])^\intercal$ according to equation 2 in https://en.wikipedia.org/wiki/Multivariate_random_variable

My question is now how this would look like if we have a conditional expectation regarding another random vector $Y=(Y_1, ..., Y_p)^\intercal$?

Would it just be $\mathbb{E}[X\mid Y] = (\mathbb{E}[X_1 \mid Y], ..., \mathbb{E}[X_n \mid Y])^\intercal$?

If so, why?

Maybe someone could give a measure theoretic reasoning in terms of the generated sigma algebra.


1 Answer 1


Yes, $\mathbb{E}[X\mid Y]$ would be exactly as you stated. This is because any element of $X$ could be dependent on any combination of elements of $Y$.

I can't really see what the alternative could be, to be honest


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.