taking the resultant moment (and then calc Mr/I) is the better approach.
if you use moment components and add you run the risk of being (unknowingly) conservative. Mx*r/I +My*r/I is conservative (why?).
the "proper" way is to maximise the function Mx*r*cos(theta)/I +My*r*sin(theta)/I ... using the reulstant M is much easier !
by taking the resultant vector I mean C = sqrt (A^2 + B^2). Otherwise calculate the stress contribution from each moment individually and add them toghether
i think Tmoose's point is that if you vector sum the two stresses (as opposed to simple addition) then you'll get the same result as if you apply the resultant moment.
this is correct, though i think apply the resultant moment is more intuitive. this is a special case, 'cause the round section has no "special" direction for the major/minor principal axes.
While taking the resultant moment would work well enough for circular members, it is not sufficient for members which are non-circular since the determination of the section modulus normal to the resultant moment would be cumbersome. Chapter H of AISC 360 (available in the Steel Construction Manual or as a free download from AISC) is a useful for non-circular sections.