Both in New Zealand and abroad, the rate of workplace accidents is extremely high and leads to significant human and financial loss. While measurement alone will not solve the issue, organisations measuring health and safety-related behaviours and abilities can provide an important step in reducing potential adverse events. Unfortunately, the existing tools to measure factors are susceptible to issues around self-report or do not measure appropriate behaviours predictive of accident outcomes. As hazard recognition is a leading cause in workplace accidents, the Hazard Awareness Test (HAT) was developed as a gamified, objective tool to measure job candidates hazard recognition abilities. Past validation of the HAT showed promising findings, but the validation methods were concurrent rather than predictive. The present study had two main aims; to further validate the HAT as a criterion valid tool using a predictive validation method and to validate a shorter, more commercially viable version of the tool, the 4pHAT. Seventy participants completed the HAT/4pHAT and a lab-based hazard recognition task one week later. In addressing the criterion-related validity of the measures, the results indicated that the HAT and the 4pHAT are effective measures of hazard recognition ability. The implications of these results in an applied context are discussed.