The effect of virtual reality on behavioral, affective, and cognitive skills: A meta-analysis study

The aim of this study is to examine the effect of virtual reality (VR) technology on skill development through cognitive, affective, and behavioral dimensions using the meta-analysis method. Based on the inclusion criteria, 53 primary studies from the international literature were selected, and the data were analyzed using the meta-analysis approach. The analysis, conducted with a random-effects model, evaluated the overall effect size, heterogeneity level, publication bias, and moderator variables. Findings indicate that VR applications have a generally moderate effect on skill development. The effect was found to differ significantly depending on skill type (cognitive, affective, behavioral), research method, application duration, immersion level, educational level, VR technology type, publication year, and data analysis methods. When the results were examined in terms of moderator analyses, a more detailed picture emerged. According to skill type, VR applications had a high-level effect on affective skills (g = 1.101) and moderate-level effects on cognitive (g = 0.328) and behavioral skills (g = 0.478). In terms of research methods, both true experimental (g = 0.473) and quasi-experimental studies (g = 0.523) showed similar, moderate, and significant effects, indicating that the method did not create a determining difference. Regarding application duration, short-term studies showed low effects (g = 0.168), while medium- and long-term applications increased the effect (g = 0.475; g = 0.596), with the highest impact observed in very long-term applications (g = 0.840). In terms of educational level, the strongest effects were found at the primary school (g = 1.660) and middle school (g = 1.051) levels, while the effects were lower at high school and graduate levels. Regarding VR technology type, fully immersive systems produced higher effects, whereas semi-immersive and non-immersive applications yielded relatively limited impacts. Analysis by publication year showed that recent studies (2021–2023) demonstrated higher effect sizes (g = 0.740). Finally, when examined by data analysis methods, no significant difference was found between quantitative and mixed-method (quantitative + qualitative) studies.