populations usually too large to study all of it so we select a sample from it that is representative of everyone - sample drawn from population will be representative of target population so generalisation of findings becomes possible (giving high population validity)