Speaking up is an important yet challenging aspect of health professional communication. To overcome social-cognitive influences and improve speaking up, an intervention based on Kolb's experiential learning cycle was developed, which integrated Virtual Simulation, curriculum, and practice speaking up. The present study investigated if integrating Virtual Simulation influenced Respiratory Therapy students' ability to challenge a physician compared to a control condition at multiple time points during training. A multi-institutional longitudinal randomized control trial was conducted. Students from two schools completed a Virtual Simulation or No Virtual Simulation before classroom instruction on speaking up and an in-person simulation requiring speaking up. After three-to-six months and post-clinical placement, students completed a second simulation requiring speaking up. The student's ability to speak up and use CUS (Concerned, Uncomfortable, Safety Issue) was measured. No significant effects for the intervention were observed across time points, p>.05, with a small effect for using CUS, ϕ=.28. During the study, two unexpected findings emerged with theoretical and practical implications. The multi-institutional design created a natural experiment that allowed for the identification of instructor effects on speaking up and Bloom's Two-Sigma problem. Observations were also made related to perceptual limitations that diminish the ability to speak up. Single speaking-up interventions continue to appear to be ineffective. To substantially influence behaviour, consistent mentorship through a "champion" is likely necessary to train for and create a culture of speaking up. Training in situational awareness is also likely needed to counter human perceptual limitations in complex situations.