The complete set of cases that could be included in a valid analysis. For example, if studying US Citizens, then the population is all US Citizens at the time of the study.
In practice, the definition of the population is often surprisingly difficult. For example, in a study of voting in the 2052 presidential election, the population is everybody that will vote in the 2052 presidential election. As the example is intended to illustrate, it is often the case that the population of interest is not knowable. For example, we will not know who voted in the 2052 election until after it has been completed, and perhaps not even then.
Much of the art and science of experimental design and survey design relates to selecting cases such that they represent a meaningful population. For example:
- A study of the cat food buyer population may only include people that have purchased cat food in California during the past 3 months from a supermarket.
- A study looking at attitudes towards chewing gum may include people that do not chew chewing gum based on various criteria of whether or not they may potentially chew chewing gum.