The aim of this chapter is to show how to obtain information about a phenomenon concerning a collective (population) observing just a part of it (the sample). The objective is to investigate real phenomena as they are, via a sample, and not build a sample with the aim of verifying our hypothesis at all costs. It is easier to sample badly than to sample well (Moore 2000). In fact the diffi culties about getting information via a sample survey or opinion polling, concern both “ethical” and “methodological” aspects that are typical of the statistics. This chapter focuses on the characteristics and properties that a subset of the population must have to be termed a ‘sample’ and therefore estimating the unknown parameters of the population.