Microscopic pedestrian simulation models are based on local interactions between agents. Many interactions occur between walkers, with many factors of influence. There is a need for observations of individuals facing interactions in crowds to better understand them and improve the level of realism of simulation algorithms. We explore Virtual Reality (VR) as an experimental tool to perform such observations with an accurate control of experimental conditions. However, the bias introduced in the collected data through the VR-system must be evaluated. We present the effort we have made in the last few years to evaluate such bias.
