A population is a set of all members who share a certain, common characteristic; a set of all people or things that are of interest in a particular study.

The size of a population is determined by the number of its members.

In practice, it is rare for all members of a population to be available or such a study would be expensive, which is why research is most often done on a sample.

A sample is a smaller set of people or things that are selected from a larger set, i.e. the population. A sample is, therefore, a subset of the population.

The main purpose of a sample is to draw conclusions about the population based on the results of measurements or tests performed on that sample.

A statistic refers to a characteristic of the sample, and a parameter refers to a characteristic of the population.

In order to draw the best possible conclusions about the population based on the sample, the sample must represent the population well. The ideal sample in a research study is a random sample.

A simple random sample has the following characteristics:

  • each member of the population has the same probability of being selected for the sample;
  • the selection of each member of the sample is independent, i.e. it does not affect the selection of any other member; and
  • each sample of a given size has the same probability of being selected as any other sample of the same size.

Example: Population and Sample

If someone from human resources at "Company X" is interested in examining the motivation for work of its employees, then everyone who works in that organization would constitute the population.

If the research concerned the motivation for work of all employees in the industry to which the organization belongs, then the employees of that one organization would only be a biased subset of that industry, not a legitimate sample of it – a true sample would need to include employees selected from across the entire industry.

It is difficult to imagine that the employees of just one organization are representative of the entire industry, which is why conclusions drawn from them cannot safely be generalized to the industry as a whole.