Virtual Reality applied to teaching the assembly of prefabricated housing in graduate studies: a mixed analysis of usability and performance

Abstract

Teaching the assembly of prefabricated housing requires mastering both procedural sequencing and spatial understanding to position and connect structural components accurately. This paper evaluates a guided immersive Virtual Reality (VR) workflow designed to simulate the structural assembly of a prefabricated house as a safe, repeatable learning activity. In a single-session pilot with n=12 construction graduate students, participants completed the VR assembly training task and responded to a post-experience questionnaire (Likert 1-7) along with open-ended questions. At the same time, the VR software automatically logged objective performance indicators (time, attempts, successes/errors, accuracy, and efficiency). Survey results showed consistently high ratings across usability/conditions, perceived learning, and transfer intention (an overall rating of 6.25 out of 7). Software logs indicated an average completion time of 10.69 ± 2.65 minutes and a mean accuracy of 0.61 ± 0.18, with significant variability in error-driven rework (ranging from 3 to 40), suggesting diverse execution strategies. Correlation analyses confirmed strong internal links among objective metrics (e.g., accuracy vs. errors/min, ρ=-0.984), while perception-performance associations were generally weak; notably, perceived learning increased with time/attempts and decreased with accuracy/efficiency, suggesting a “productive struggle” pattern during early exposure. Qualitative feedback emphasized improvements in room-scale safety, practice recurrence, and task complexity. Overall, the study offers a replicable triangulation framework that advances VR construction-process education from satisfaction-based assessment to measurable procedural performance and practical design improvements.