A single-photon incident on a beam splitter produces an entangled field state, and in principle could be used to violate a Bell inequality, but such an experiment (without postselection) is beyond the reach of current experiments. Here we consider the somewhat simpler task of demonstrating Einstein-Podolsky-Rosen (EPR) steering with a single photon (also without postselection). We demonstrate that Alice's choice of measurement on her portion of the entangled state can affect Bob's portion of the entangled state in his laboratory, in a sense rigorously defined by us and Doherty [ Phys. Rev. Lett. 98 140402 (2007)]. Previous work by Lvovsky and coworkers [ Phys. Rev. Lett. 92 047903 (2004)] has addressed this phenomenon (which they called remote preparation) experimentally using homodyne measurements on a single photon. Here we show that, unfortunately, their experimental parameters do not meet the bounds necessary for a rigorous demonstration of EPR steering with a single photon. However, we also show that modest improvements in the experimental parameters, and the addition of photon counting to the arsenal of Alice's measurements, would be sufficient to allow such a demonstration.