In this article an educational action research study, based on a phenomenographic approach, is reported in which unexpected results have been possible to gather thanks to the inductive design of the study. The aim is to describe the ways in which contrasts of critical aspects of a learning object affect the students' generative learning found by analysing three learning studies based on the theory of variation. Variation is in this article defined as the varied ways a phenomenon can be discerned. By contrasting critical aspects (i.e. main features needed to understand a phenomenon) in a dimension of variation, the learning object (i.e. the targeted ability or knowledge taught) can change form and be experienced in different ways that influence the students' ability to learn. To investigate in which ways the contrasts affect the students' learning outcome was the primary focus of the study, but the results also show an interesting pattern of how students' learning outcomes in the short-term and long-term perspective are affected. In this study we have worked with learning study as a method, and the results are based on analyses of three learning studies made up of three lessons each. The results show how one pattern of contrasts allows the students to look critically upon their previous knowledge and make them find new ways of seeing the object of learning. This pattern has also been found to be more powerful in preparing students for future learning, since it seems to generate new learning (generative learning) after the learning situation itself.