Process industry is known for its complexity and sensitivity with critical procedures saturated with demanding human-machine interfaces that may induce human errors thus resulting in abnormal situations. Abnormal situations may lead to near misses and even to severe accidents, which can result in loss of production and even in casualties and fatalities. This paper aims at abridging the gap between the highly demanding human machine interfaces and the training methods employed in the process industry by experimentally analyzing the effectiveness of distinct training methods in a virtually simulated abnormal situation. The performance of operators is measured by means of suitable Key Performance Indicators (KPIs) applied to the specific case study. In particular, we analyze experimentally two distinct training methods based respectively on a Power Point presentation and a 3D virtual environment. The positive outcomes of this approach consist in increasing the reliability, cost effectiveness, environmental friendliness, and safety of the process. This work is the result of the interaction between chemical engineers and experimental psychologists, which may open new horizons to scientific research.