A sample which is carried out on the basis of a ‘stratification‘ of the population; unlike a stratified sample, however, the items within each stratum are not then selected by a simple random sample procedure, but by the interviewers themselves as they conduct the survey. Each interviewer would have a specified number of items of each kind, e.g. households and individuals by income level, shops by type of trade or firms by size, which he would have to interview, but the actual choice of the items is left to him. In the stratified sample, on the other hand, an objective procedure is used to select the items which the interviewer must then visit. The advantage of a quota sample is that it is generally cheaper: it avoids the costly process of drawing a simple random sample within each stratum of items. On the other, hand it may be biased and unreliable, since the interviewers will still have considerable discretion in choosing items even within the limits set by the quotas for the strata, so that conscious or unconscious biases may exist. In actual practice, because of their lower costs and greater speed, quota samples are by far the most prevalent form of sample.
Reference: The Penguin Dictionary of Economics, 3rd edt.