Although redshift-space distortions only affect inferred distances and not angles, they still distort the projected angular clustering of galaxy samples selected using redshift dependent quantities. From an Eulerian view-point, this effect is caused by the apparent movement of galaxies into or out of the sample. From a Lagrangian view-point, we find that projecting the redshift-space overdensity field over a finite radial distance does not remove all the anisotropic distortions. We investigate this effect, showing that it strongly boosts the amplitude of clustering for narrow samples and can also reduce the significance of baryonic features in the correlation function. We argue that the effect can be mitigated by binning in apparent galaxy pair-centre rather than galaxy position, and applying an upper limit to the radial galaxy separation. We demonstrate this approach, contrasting against standard top-hat binning in galaxy distance, using sub-samples taken from the Hubble Volume simulations. Using a simple model for the radial distribution expected for galaxies from a survey such as the Dark Energy Survey (DES), we show that this binning scheme will simplify analyses that will measure baryon acoustic oscillations within such galaxy samples. Comparing results from different binning schemes has the potential to provide measurements of the amplitude of the redshift-space distortions. Our analysis is relevant for other photometric redshift surveys, including those made by the Panoramic Survey Telescope & Rapid Response System (Pan-Starrs) and the Large Synoptic Survey Telescope (LSST).