This study focuses on the issue of theoretically incongruent learning patterns found among subgroups of students. The longitudinal study was done among first year university students in a student-oriented learning environment. Students completed the Inventory of Learning Styles (ILS) twice, once in the first, and once in the third trimester. This inventory assesses learning strategies, learning orientations and learning conceptions. Students' learning patterns in the third trimester were far more dissonant than in the first trimester. Data for two subgroups of students were analysed separately. Both groups showed learning patterns dissonant from theory, especially in the relations between orientations and conceptions on the one hand, and learning strategies on the other hand. This raises the question that, when dissonant patterns are observed among groups of students that do not conform to what is expected theoretically, when is it necessary to revise theory and when can real dissonance be confirmed? After revising theory, dissonant patterns may disappear, because they no longer deviate from what can be expected theoretically.