Homogeneity of Variance

The term homogeneity of variance, which is also often referred to as homoskedasticity, is defined as the assumption that any distribution or comparison of distributions shares the same level of variance within the particular group of data points. Most statistical tests assume that a comparison of the level of variability of groups or to a hypothetical distribution will exhibit similarity, or at a minimum demonstrate a lack of significant deviation from the expected distribution. The importance of the assumption lies in the nature of the tests that compare between-group variability (usually the mean of each group) to the level of within-group variability (often expressed as the error term, the average mean square error/within, or weighted average standard deviation). This entry provides discussions of two types ...