The explanatory variables in a general linear model partially explain why the response values vary in a data set. An analysis of variance table gives numerical values for:

The analysis of variance (anova) table provides a different approach to testing whether explanatory variables are needed in a general linear model. The ordinary t-tests for the significance of model parameters are equivalent to anova tests, but the anova approach also provides tests of hypotheses that cannot be assessed with t-tests.

Analysis of variance also explains problems that arise when the explanatory variables are highly correlated — multicollinearity.