In statistics, we are often interested in the population which is some entire set of entities that share a common trait. We often draw a sample which is a subset of the population to draw inferences about the population. The set of measurements taken from the sample form the sample data.
For example, we might be interested about drawing inferences about the number of hepatitis outbreaks in Springfield. The population would be all the inhabitants of Springfield within city limits. We can make a statistical inference about the population by drawing a sample which consists of choosing thirty people at random from the telephone book though this is not the perfect way to choose a sample. Why not ?
Often the quality of a sample is limited by both internal and external factors. For example, the quality of the measurements may impair the precision and accuracy of the inferences drawn from the sample data. In our hepatitis study, a positive case of hepatitis may be falsely recorded as negative. An external limiting factor is the cost to gather sample data.