Abstract
This research study is focused on the investigation of the covariation of magnitudes in modelling physical phenomena. Starting from the analysis of the data provided by two teaching experiments, we extend the theory of the six levels of covariational reasoning (Thompson & Carlson, 2017) in order to describe those situations in which not only variables are involved, but also parameters. We aim at introducing and defining the construct of second order covariation (Arzarello, 2019). In particular, the here proposed analysis is focused on detecting clear signs of transition toward a higher order of covariation and on the importance of technological supports for a full achievement of it.