One-way analysis of variance (ANOVA) is performed either on a given set of data or on the set's corresponding sufficient statistics (nj , , sj 2).Statistical programs implement one-way ANOVA commands directly on a complete set of data.An uncomplicated technique is presented that allows the standard one-way ANOVA procedures to be invoked directly on the sufficient statistics.Sufficiency is utilized to derive two equations from which an equivalent surrogate set of data is constructed.The constructed set of data replaces the actual (unavailable) set of data.The technique can be generalized to the n-way ANOVA case.