The distribution of an estimator is biased if its expected value does not equal the population mean. Individual values returned by estimators will usually not be exactly equal to the true value, and the determination of bias is a property of the sampling distribution.