Introduction: Bystander cardiopulmonary resuscitation (CPR) training is inconsistent among students and the public. Existing CPR teaching courses are costly, time-consuming, and inconsistent. This study aimed to determine the association between overall CPR competency and two teaching modules, a group-directed video instruction module versus an instructor-led traditional classroom instruction module. Methods: This randomized prospective interventional study involved first year medical students of Universiti Sains Malaysia Health Campus from November 2018 until January 2019. Pass-fail scores representing the overall CPR, individual skill performance, and willingness to perform CPR for strangers and family members were collected. Factors associated with reluctance to perform CPR were assessed in a questionnaire. Results: A total of 99 participants were included, 50 in the group-directed video instruction as the intervention module and 49 in the traditional classroom instruction as the control module. There was no statistical significance between the pass and fail outcomes for both video module (p=0.436). Participants in both modules performed similarly in 8 out of 12 individual CPR skills. There was a significant difference in the distribution of skill scores between the pass and fail outcomes (p=<0.001). The intervention module is non-inferior compared to the control module, in relation to CPR willingness rates for strangers (p=0.999) and family members (p=0.117) after the training. Conclusions: The group-directed video self-instruction method is as effective as the instructor-led traditional classroom method to help participants to be competent and willing to perform CPR. It can be used as an independent or supplementary teaching tool for first-time learners and refreshers, especially in a group setting when teaching materials are limited.