Students of all levels of education tend to have ‘justification schemes’ (Harel & Sowder, 1998) that are inconsistent with conventional validation methods. Yet, there is limited research knowledge about how mathematics instruction can support progressions in students’ justification schemes so that they better approximate conventional validation methods. In this talk, I will draw on findings from a four-year design experiment in an undergraduate mathematics course to present and exemplify an instructional intervention that has been successful in supporting progressions in students’ justification schemes. The notion of ‘cognitive conflict’ featured prominently in the theoretical framework that underpinned the design of the intervention.