Dog “aggression” in the veterinary practice is commonplace. Therefore, student knowledge and education about dog behaviour and the ability to interpret “aggressive” behaviour is important from a human injury prevention and dog welfare perspective. The study aimed to compare first-year veterinary students’ perceived safest proximity to both an “aggressive” and non-reactive simulated dog, both before and after a teaching intervention about canine behaviour and a handling practical. It also examined student confidence and their ability to identify “aggressive” behaviours. Forty first year veterinary students took part in two surveys. Each survey included two videos: one of a simulated dog displaying “aggressive” behaviour, based on the ‘Canine Ladder of Aggression’; and another displaying non-reactive (passive behaviours without reaction to the participants) behaviours. Each video depicted the slow and consistent approach towards the virtual dog within a virtual indoor environment, and participants were asked to press stop if or when they would stop approaching the dog. In the “aggressive” scenario, there was a reduction in the approach-stop time from survey 1 (median = 17.8 s) to survey 2 (median = 15.2 s) in the intervention group (p = 0.018) but not in the control group (p = 0.147). Regarding confidence, there was a significant increase in the self-reported confidence rating relating to a participant’s ability to interpret canine behaviour in both the control (p = 0.011) and intervention (p = 0.003). In conclusion, these results indicate that students using approach-stop videos stayed further away from an “aggressive” virtual dog model if they had undertaken a canine behaviour educational intervention. This novel approach has the potential for further use in teaching and assessment of student knowledge and behaviour which may otherwise be difficult to demonstrate.