On the interpretation of stellar disc observations in terms of diameters