Learning the complex skill of bronchoscopy involves the integration of cognitive domains and motor skills. The development of simulators has opened up new possibilities in bronchoscopy training. This study aimed at evaluating how effective the modelling example methodology is in training this skill and assessed its effect on cognitive load in learning. Methods: Forty-seven medical students participating in a simulator-based bronchoscopy training program were randomly allocated to a control group, receiving a video lesson, and the modelling example group. They were evaluated by the simulator’s metrics at different time points: pre-test, post-test and 15 days and 12 months after training. Cognitive load was assessed with the modified Paas scale. Results: Simulation-based training was effective for both groups, based on simulator metrics (p<0.05). The modelling example group outperformed the control group in all measures at post-test and after 15 days (p<0.001). After 12 months, there was a decline in skill in both groups, but the modelling example group performed better (p<0.001). Simulation-based training reduced cognitive load, more strongly so in the modelling example group (p<0.001). Conclusion: The modelling example group showed substantial benefits over the control group, both in reducing the cognitive load in learning and in retaining knowledge and skill after 15 days and 12 months.