Learning progressions are theoretical models that describe learning of scientific ideas and practices over time. These hypothetical progressions need to be tested and refined in order to productively inform instruction and assessment. In this paper, we report our attempts to revise a learning progression in genetics. In particular, we focused on two constructs that embody core ideas in classical genetics and one molecular construct. The revisions are based on analysis of pre- and postinterview data obtained from sixty 11th grade students before and after they engaged in a 10-week unit that addressed these concepts. We found that while many of the students held ideas that aligned with the progression, there were several distinct dimensions of student reasoning that were not captured and led to substantial revisions of the constructs including: (a) the splitting of the construct dealing with meiosis (E) into two subconstructs (E1-physical passage of genetic information and E2 -- the role of sex cells), (b) the addition of new levels to constructs dealing with the universal nature and organization of the genetic code (A) and construct (F). For Construct A, the lower levels were expanded to include ideas about the localization of DNA in cells and to include ideas about the composition of DNA that were not captured in the progression. Revisions to Construct F included the expansion of existing levels and the addition of modes of inheritance such as codominance and incomplete dominance. The research we present offers insights about a methodological approach that can be used to test and refine progressions, as well as insights about student learning in genetics as we further describe and expand the stepping-stone ideas in the progression and discuss further the multidimensional nature of learning progressions.