Having the same variance. A set, or a vector, of observations is homoscedastic if the variance of the random error is the same for all observations. This is one of the classical linear regression assumptions underlying the Gauss–Markov theorem. Violation of this assumption is called heteroscedasticity.