<p>Mixed Reality (MR), including virtual and augmented reality, is increasingly implemented for Vocational Education and Trainings (VETs), prompting research to establish their effectiveness in professional settings. This paper systematizes this research for behavioral, cognitive, and affective training outcomes respectively, while investigating the role of the training technology, content and context. The systematic literature review covers <i>k</i> = 53 studies, followed by a synthesis of the results through three-level meta-analyses with exploratory moderator analyses. When compared to control groups, there are significant positive effects for the use of MR in VETs for behavioral (<i>d</i> = 0.40), cognitive (<i>d</i> = 0.84), and affective training outcomes (<i>d</i> = 0.65). These effects remained comparable when considering the subset of studies with equal training contents (<i>d</i>_{<i>behavioral</i>} = 0.40, <i>d</i>_{<i>cognitive</i>} = 0.68, <i>d</i>_{<i>affective</i>} = 0.71) but not when considering the subsets where two MR conditions were compared (<i>d</i>_{<i>behavioral</i>} = 0.04, <i>d</i>_{<i>cognitive</i>} = −0.31, <i>d</i>_{<i>affective</i>} =&#xa0;−0.51). Furthermore, results indicate that within MR VETs behavioral (<i>d</i> = 1.79), cognitive, (<i>d</i> = 1.29), and affective (<i>d</i> = 1.31) training outcomes improve from pre-to-post-intervention, but results at follow-up were more ambivalent. The subsequently conducted moderator analyses highlighted that the effectiveness of MR VETs is mostly independent of the investigated technological, content, and context features. Overall, these results underline the promising effects of MR training for different kinds of training outcomes but also pointed to a lack of comparative research which implements and compares instructional features to support active knowledge construction and investigates the interaction of immersive technologies with instructional methods.</p>