[Methodological premises of statistical analysis of the data].

Most psychic disorders, such as anxiety, agitation and dementia are difficult to measure objectively since the scale of measure--psychometric tests--are often liable to subjective judgment--especially of clinicians or patients during self-rating scales. It become therefore necessary to use a complex way of measuring--a battery of tests--from which the results should be analyzed through multivariate statistic models. During this polycentric it was believed suitable to use the repeated multivariate analysis of variance, an instrument largely used in the data analysis of rating scale thanks to the flexibility of the generalized east squares methods. In this article the characteristics of this method and how to read its results will be discussed.