One of the cornerstones in resuscitation training is defibrillation safety, inadvertent “shocking” of the patient when another person has contact with the patient may have a range of safety consequences. The objective of the study was to assess visual and verbal safety checks by paramedic students prior to defibrillation. This was a prospective observational mannequin study of defibrillation safety during a simulated cardiac arrest by paramedic students. The study was conducted in the lounge room of the Department of Community Emergency Health & Paramedic Practice simulation flat, a replica of a complete flat where prehospital simulations are conducted. Each student completed two 10-min cardiac arrest simulations with multiple defibrillation attempts. Each student and an independent Faculty member rated the simulation safety performance using a defibrillation safety self-assessment (DSSA) form. Twenty-four (20 %) students participated in the study with 14 (58 %) being female. For scenario one agreement between student and assessor proved significant for “scanning the incident scene” for all three defibrillation attempts, with agreement ranging from 29 % (p = 0.044) to 47 % (p = 0.007), and stating “stand clear” for defibrillation attempt one and three with the agreement ranging from 47 % (p = 0.007) to 100 % (p < 0.001). For scenario two agreement between student and assessor proved significant for “charging eye contact” for all three defibrillation attempts, with agreement ranging from 40 % (p = 0.043) to 53 % (p = 0.003), and “scanning the scene to ensure all persons are clear of the patient” before defibrillation attempt one and two with agreement ranging from 29 % (p = 0.044) to 46 % (p < 0.007). The results of this study suggest student perception of their performance and what they actually do is vastly different. Further studies using video recording glasses are required so students can gain an accurate and realistic sense of their defibrillation safety performance.