We examine, in the context of the Einstein-Podolsky-Rosen-Bohm gedankenexperiment, problems associated with state reduction and with nonlocal influences according to different interpretations of quantum mechanics, when attempts are made to apply these interpretations in the relativistic domain. We begin by considering the significance of retrodiction within four different interpretations of quantum mechanics, and show that three of these interpretations, if applied in a relativistic context, can lead to ambiguities in their description of a process. We consider ways of dealing with (...) these ambiguities, in particular focussing on the “preferred frame” hypothesis. We then re-examine an argument involving nonlocal measurements which claimed that the preferred frame hypothesis is not tenable, and show that this argument does not in fact necessitate a rejection of the preferred frame. We then suggest that, to avoid confusion, the preferred frame could be extended to cover unitary interactions as well as state reductions. We conclude with a brief examination of a proposal that state reduction should take effect across the backward light cone of a measurement event. (shrink)