The impact of virtual experiments on students’ scientific inquiry skills: evidence from a three-level meta-analysis

Virtual experiments (VE) represent a crucial practical medium for science experiment instruction, yet empirical inconsistencies persist regarding their effectiveness in fostering students’ scientific inquiry skills relative to physical experiments (PE). Accordingly, this study followed the PRISMA 2020 statement and systematically retrieved empirical literature published from 2000 to 2025. A total of 21 experimental and quasi-experimental studies (41 effect sizes, N = 3,931) were ultimately included. A three-level meta-analysis model was employed to correct for dependency bias from nested effect sizes, compare the intervention effects of non-immersive VE versus PE on students’ scientific inquiry skills, and examine the moderating roles of educational level, discipline type, intervention duration, and inquiry skill dimensions. Results revealed that VE exerted a significant small-to-moderate positive effect on students’ scientific inquiry skills relative to PE (Hedges’ g = 0.454, p < 0.05). Moderating effect analyses indicated no significant overall moderating effects for the four variables. Exploratory subgroup analyses demonstrated that the positive effects of VE within subgroups were significant at the senior high school level (Hedges’ g = 0.690, p < 0.05), in physics disciplines (Hedges’ g = 0.603, p < 0.05), and for 4‒8-week interventions (Hedges’ g = 0.590, p < 0.05), whereas no significant advantage was observed for the dimension of designing controlled experiments (Hedges’ g = 0.056, p > 0.05). Building on these findings, this study explains the differentiated functions of VE in optimizing cognitive load and PE in enhancing embodied experience from the perspective of cognitive-embodied dynamic balance. It elucidates the inherent complementary mechanism and potentially applicable teaching scenarios of the two approaches, providing quantitative empirical evidence for resolving controversies over the effectiveness of virtual and physical experiments and refining the analytical framework of cognitive-embodied complementarity in instruction.