In statistics, a sample randomly taken from an investigated population, usually known as a random sample. To avoid having bais from our response and for it to have the best chance of it being indicative of the entire population, our sample must be random. This random sample chosen must contain subjects related to the data in the population we what to obtain a result from.
