Orthogonal

A model term in ANOVA is non-orthogonal if it has an efficiency factor equal to one. Essentially this means that it can be estimated independently of the other terms in the model.