Nowadays, in the educational context, emphasis has been placed on the collection and
analysis of data by scientists from several areas of knowledge, such as: psychology and
economics. These professionals analyze data and their results can be used, for example, to
aid in the decision making of a public policy. However, educational measures have become
very popular and can encompass all the multidimensionality contained in the educational
process, from teaching and learning to social interaction in the classroom. Understanding
of data analysis in a classroom needs to be done by teachers and pedagogues who know
exactly the empirical signicance of the variability of a particular measured variable. In
this sense, this thesis conceptualizes, discusses, denes and applies Educometrics, which
is recognized as an area of knowledge that makes use of multivariate statistical models to
analyze data related to educational contexts. After establishing the concept of Educometrics,
we apply some mathematical models in the learning context of teaching in distance
learning. A sample of 791 students answered the updated QEOn questionnaire for three
courses and the factorial structure of the questionnaire was valid from the application
of factorial analysis. The principal component analysis and the Parafac2, bilinear and
multilinear models, respectively, were applied and able to identify intrinsic behaviors in
relation to the 34 statements contained in the QEOn questionnaire. As a conclusion, the
application of models that allow pedagogical intervention in the classroom as it is the key
of the support provided by the educometrics developed throughout this thesis.
Keywords: Educometrics. Multivariate Statistics. Multilinear