Building a less intimidating cycling environment for women: A structural equation modeling analysis based on a VR-based laboratory experiment

<b>Background:</b> Security perception significantly affects women's intention to use bikes. This area is under-researched in a context where female cyclists face harassment in a setting of low- or middle-Income countries.<b>Objectives:</b> This study aims to contribute to understanding the determinants of security perception as a latent construct by establishing the relationship between perceived security, individual attributes, and built environment features.<b>Methods:</b> (i) Participants: a total of 208 women in Tehran participated in a laboratory-like experiment (age ranging from 18 to 42 years [M = 28.8 ± 14.15], 63% with a household having at least one bike, 75% never use a bike for daily trips, and 67% are without any experience of harassment and insecurity). (ii) Apparatus: a two-degree-of-freedom VR-based bicycle simulator comprising sixteen scenarios that combine various states of built environment factors is employed. (iii) Procedure: the test consists of several stages. The experiment for each participant took an average of 26 minutes. The experiments were conducted over 50 days. (iv) Modeling: a structural equation model (SEM) was developed and estimated using the partial least squares technique.<b>Findings:</b> Informal surveillance (i.e., people's presence) is the most important determinant of perceived security. The second most important factor is <i>evidence of incivility</i> (trash in the streets, abandoned cars, damaged urban furniture, graffiti, etc.). The third noteworthy factor is the type of built environment, such as obstacles, vegetation, and underpasses. Contrary to expectations and paradoxically, official surveillance by police forces has a minor impact on women's perceived security. It is a phenomenon in theocratic societies, where the government enforces severe legal constraints against women. The least influencing but still statistically significant factor is lighting conditions.<b>Practical implications:</b> Several solutions for increasing the perception of security for female cyclists, especially for new projects, are proposed. We also offered and partially presented the development of a spatiotemporal performance dashboard based on the time-varying perceived security index.