For complex logic, I always use a single exit point to ease code maintenance. When the code logic dictates different types of returns depending on the circumstances I try to use some sort of state machine and put my exit points at the bottom so atleast I don't have to hunt them down later.

Deeply nested unless..else statements with a few negated or clauses and recursive sub routine calls can be enough fun without the added confusion of multiple exit points :-)

For a sequence of tests like in your post though, there's no reason to perform any more processing if one test fails so return the status/error/result/whatever and get out of there. IMHO.