A number of the new generation of dark matter direct detection experim
ents, now underway, employ gamma-background rejection techniques to im
prove their performance. Event discrimination permits the setting of l
imits on dark matter event rates orders of magnitude below the raw bac
kground event rates. However, proper attention must be paid to systema
tic, as well as statistical errors, during data analysis. We illustrat
e the importance of using detectors with good intrinsic discrimination
if the impact of systematic errors is to be minimised.