We present a comparison of SCUBA-2 850-$��$m and Herschel 70--500-$��$m observations of the L1495 filament in the Taurus Molecular Cloud with the goal of characterising the SCUBA-2 Gould Belt Survey (GBS) data set. We identify and characterise starless cores in three data sets: SCUBA-2 850-$��$m, Herschel 250-$��$m, and Herschel 250-$��$m spatially filtered to mimic the SCUBA-2 data. SCUBA-2 detects only the highest-surface-brightness sources, principally detecting protostellar sources and starless cores embedded in filaments, while Herschel is sensitive to most of the cloud structure, including extended low-surface-brightness emission. Herschel detects considerably more sources than SCUBA-2 even after spatial filtering. We investigate which properties of a starless core detected by Herschel determine its detectability by SCUBA-2, and find that they are the core's temperature and column density (for given dust properties). For similar-temperature cores, such as those seen in L1495, the surface brightnesses of the cores are determined by their column densities, with the highest-column-density cores being detected by SCUBA-2. For roughly spherical geometries, column density corresponds to volume density, and so SCUBA-2 selects the densest cores from a population at a given temperature. This selection effect, which we quantify as a function of distance, makes SCUBA-2 ideal for identifying those cores in Herschel catalogues that are closest to forming stars. Our results can now be used by anyone wishing to use the SCUBA-2 GBS data set.

MNRAS accepted, 19 pages, 14 figures