A lump - sample is a form of random selection. In a lump - sample the population is divided into many small (often geographically delimited ) subpopulations, called chunks (eg schools for a survey of students ). This process starts as opposed to stratified sample before so that only a portion of the lump is randomly selected and enters the sample. The elements of the lumps are then, however, fully ( all students in the selected schools are asked ).
Compared to the simple random sample is often achieved with a cluster sample a significant cost reduction per unit collected, for which, however, an increased sampling error, which is also referred to as a cluster effect, must be taken into account.