(ANCOVA) an extension of the analysis of variance procedure that statistically eliminates the effect of variables irrelevent to the question of interest (i.e., the test of the null hypothesis); specifically, this procedure applies to the case where the dependent variable (i.e., the variable being analyzed) is a function of one or more independent variables (covariates) that have not been controlled in the experimental or sampling unit —note the analysis seeks to discover whether the variation of the dependent variable between classes is due to class effects or to its dependence on the other variates that themselves vary between classes; class effects are determined by adjusting values of the dependent variable to a common level of the covariates This definition last updated 10/23/2008