I've got a question that I really think should be quite simple to answer, but I just can't see what I'm missing. We have the random variables $X \sim R(0,1)$ and $Z\sim b(1,1/2)$. I want to determine $\operatorname{Cov}(XZ,Z)$. However, as far as I can tell, this is the same as $\operatorname{Cov}(Z,Z)=\operatorname{Var}(Z)=1/4$. The listed answer is $1/8$. What am I misunderstanding here?

  • $\begingroup$ I changed \textit{~} to \sim. Thus you see $X\sim R(0,1)$ rather than $X\textit{~}R(0,1)$. ${}\qquad{}$ $\endgroup$ – Michael Hardy Aug 1 '14 at 16:35

We are presumably expected to assume that $X$ and $Z$ are independent. That really should have been explicitly specified.

We want to calculate $E(XZ^2)-E(XZ)E(Z)$.

By independence $E(XZ)E(Z)=E(X)E(Z)E(Z)=1/8$.

The first term is $E(X)E(Z^2)$. But $E(Z^2)=1/2$ because $Z^2=Z$, Thus $E(XZ^2)=1/4$.

Our covariance is therefore $1/4-1/8$.

  • $\begingroup$ Yeah, this is what happens when you forget your covariance formulas! $\endgroup$ – Elswyyr Aug 1 '14 at 15:45
  • $\begingroup$ You probably have used over and over again $\text{Var}(X)=E(X^2)-(E(X))^2$. The covariance formula has exactly the same structure, $\text{Cov}(XY)=E(XY)-E(X)E(Y)$, $\endgroup$ – André Nicolas Aug 1 '14 at 15:49

The $\operatorname{Cov}(XZ,Z)$ is given by $E[XZ\cdot Z]-E[XZ]E[Z]$ which is the same as $E[X]E[Z^2]-E[X]E[Z]E[Z]$. This gives us $1/2 \cdot 1/2 - 1/2 \cdot 1/4 \cdot\cdot 1/4$ which is equal to $\frac{1}{8}$.


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.