In most situations where data are collected, the underlying population (or process) is of more interest than the specific sample data. The characteristics of this population are usually unknown, so the sample data must be used to infer information about the population.

Using sample data to answer questions about an unknown population is called inference. Estimation is one branch of inference that tries to determine the value of a population parameter.

Since sample data are random, parameter estimates are also random quantities and an estimate is unlikely to be exactly equal to the unknown population parameter. A confidence interval is the most common way to describe the likely errors.