A sampling bias has occurred if the sample selected under or over represent certain groups that compose the target population. In this case, the sample is unrepresentative of the target population. This limits the extent to which generalisations can be made. To avoid this, the sample should be as large as possible