The second phase of the method, concerning statistical analysis of the data, begins with
the estimate of what we term the “consistency model”. It consists in defining a classical linear
model in which the parameters are estimated with the ordinary least squares method. In this first
model the dependent variable consists of scores attributed to global satisfaction of the box
scheme, while explicative variables consist of the average of scores attributed to each
macrocategory with due weighting assigned by the consumers. If the responses are consistent, the
parameters of all macrocategories must be statistically significant.