If $X_i$ are random variables, then $E(\frac{(X_i - 1)^2}{X_i})$ will be?

Question

If we have something like $E(\frac{(X_i - 1)^2}{X_i})$, then will:

$E(\frac{(X_i - 1)^2}{X_i}) = E(X_i + \frac{1}{X_i} - 2)$; or;
$E(\frac{(X_i - 1)^2}{X_i}) = \frac{E(X_i-1)^2}{E(X_i)}?$

I am getting a bit confused which one will be.

Answer 1

The first one pretty much has to be right... all you have done is multiply out the expression inside the expectation value.

The second one is wrong. You are not generally allowed to factor expectation values. For instance $$E(1/X) \ne \frac{1}{E(X)}.$$