Immersive virtual environments (VEs) are increasingly used to evaluate human responses to design variables. VEs provide a tremendous capacity to isolate and readily adjust specific features of an architectural or product design. They also allow researchers to safely and effectively measure performance factors and physiological responses. However, the success of this form of design-testing depends on the generalizability of response measurements between VEs and real-world contexts. At the current time, there is very limited research evaluating the consistency of human response data across identical real and virtual environments. Rendering tools were used to precisely replicate a real-world classroom in virtual space. Participants were recruited and asked to complete a series of cognitive tests in the real classroom and in the virtual classroom. Physiological data were collected during these tests, including electroencephalography (EEG), electrocardiography (ECG), electrooculography (EOG), galvanic skin response (GSR), and head acceleration. Participants' accuracy on the cognitive tests did not significantly differ between the real classroom and the identical VE. However, the participants answered the tests more rapidly in the VE. No significant differences were found in eye blink rate and heart rate between the real and VR settings. Head acceleration and GSR variance were lower in the VE setting. Overall, EEG frequency band-power was not significantly altered between the real-world classroom and the VE. Analysis of EEG event-related potentials likewise indicated strong similarity between the real-world classroom and the VE, with a single exception related to executive functioning in a color-mismatch task.