We investigated how student understanding of complex science topics was impacted when 27 teachers switched from typical to inquiry instruction in a delayed cohort comparison design study. For the same set of science topics, the teachers used typical methods of instruction in the first year and online, visualization rich inquiry units in the second year. Both cohorts of students were tested on knowledge integration at the end of both school years. We obtained students' knowledge integration estimates by applying an Item Response Theory analysis based on a Rasch Partial Credit Model. We used a mixed effects analysis of variance to investigate effects related to inquiry instruction, teaching context, and science course. We found significant main effects of inquiry instruction and teaching context as well as significant interaction effects between inquiry instruction and science course and between inquiry instruction and teaching context on student knowledge integration. We triangulate these findings with teacher surveys, interview transcripts and project records to explore potential factors associated with successful implementation of inquiry instruction. (Contains 3 figures and 3 tables.)