If n=1, i.e. at one supervision in a cell, not all zero hypotheses can be checked as the average square as in this case there cannot be a speech about interaction of factors drops out the Q3 component of the total amount of squares of deviations, and with it also.
Division of dispersion into parts or components is the cornerstone of the dispersive analysis. The variation caused by influence of the factor which is been the basis for group is characterized by intergroup dispersion σ It is a measure of a variation of private averages for groups round general average and is determined by a formula:
They say that the equipment of the dispersive analysis is "robastny". This term used by statisticians means that these assumptions can be somewhat broken, but despite it, the equipment can be used.