Optimizing immersive learning in crisis management: cognitive and psychological mechanisms of VR training

Background Virtual Reality (VR) training holds significant promise for crisis education, yet the cognitive and psychological mechanisms through which it enhances higher-order decision-making skills remain inadequately explained. Grounded in Presence Theory and Cognitive Load Theory, this study investigates how VR training improves decision-making competence, examining the mediating role of immersion and the moderating effect of training duration. Methods A cross-sectional survey was conducted with 352 emergency response professionals from China, selected via stratified random sampling. Data were analyzed using a combination of Structural Equation Modeling (SEM) and Hayes’ PROCESS macro (Model 59) to test a moderated mediation model. Results The model demonstrated substantial predictive accuracy, explaining 35.1% of the variance in immersion and 23.8% in decision-making competence. VR training positively predicted decision-making ( β = 0.29, p < 0.001), with immersion serving as a significant partial mediator [indirect effect β = 0.094, 95% CI (0.036, 0.152)]. Training duration marginally moderated the immersion–decision-making link ( β = −0.14, p = 0.054), with simple slopes and Johnson-Neyman analyses confirming that the benefits of immersion are strongest at shorter durations and diminish as training extends. Conclusion This study empirically validates immersion as a key psychological mechanism in VR-based learning and identifies training duration as a temporal boundary condition shaping its effectiveness. Theoretically, it extends Presence Theory by linking immersive experience to higher-order cognitive performance; practically, it informs the design of VR-based training policies by emphasizing the coordination of cognitive load, temporal factors, and human–VR interaction design to optimize learning outcomes.