Marketing decisions are often based upon measurements of the market place, which are usually constructed from sampled data. These data are obtained from mail surveys, face-to-face interviews, telephone surveys, interviews at shopping centers, etc. The recommended procedure is to capture data from a probability sampling scheme based on the assumption that the observations will be well balanced against all variables, independent of experimenter bias and necessary for valid inferences of unknown population parameters. The discussion in this paper compares the relevance of inferences from probabilistic samples to inferences from non-random samples, above all quota samples. It will be shown that inferences from quota samples could be more usefull than those from random samples and that quota samples are more informative in the presense of nonresponces.