Analysis of variance assumptions
Analysis of variance and regression. Design of experiments

Factors whose influence (or lack of it) on measurement results is being investigated are called controlled factors. The study of their action is based on an experiment consisting of a set of measurements; the measurements are characterized by defined values (levels) of these factors.

The technique of analysis of variance assumes that the results of measurements are distributed according to the same Gaussian distribution, whatever the value of the factor under study that is likely to have an influence on the measurement results. This mathematical condition often discourages potential users of this type of analysis. That's why it seems necessary to consider the physical significance of this mathematical condition.