A confidence interval is a statistical interval estimate that is expected, under repeated sampling, to contain an unknown population parameter at a specified confidence level. It expresses uncertainty around an estimate rather than a single point value.