Flaw population, importance

All this implies a selection procedure from a larger pool of individuals. In all studies of normal populations it is important, if the results are to be valid and generally applicable, that the sample investigated is representative of the whole identified population both in terms of the lead distribution, and in social and demographic indices. Flaws in the research design, sampling bias, and a high or biased drop-out rate can all result in a non-representative sample. There are three main points in the selection of a sample for study, where biases can be important. 

---