Measure in statistics that refers to the variance of errors over a sample. Heteroskedasticity is present in samples where random variables display differing variabilities than other subsets of the variables. Such results can cause errors in regression analysis and other statistical measures in which statistical measures can be incorrectly justified.