Effects of passing rates on driving behaviour in variable speed limit-controlled highways: Evidence of social pressure from a driving simulator study

Variable message signs on motorways can show dynamically changing speed limits for traffic safety or efficiency reasons. While the effects of variable speed limits have been studied in the literature, the effect of (drivers in) vehicles operating under different speed limit information ― caused by transitioning from one speed limit to another ― has not received any attention. During such a transition, drivers can either be overtaken by vehicles operating under a higher speed limit, or conversely, drivers overtake slower vehicles operating under a lower speed limit. In both scenarios, opposite forms of social pressure are expected to be exerted by the surrounding traffic. To investigate the effects of this social pressure on driving behaviour, this study analyses outcomes of driving simulator experiments where such passing rates were systematically varied. Forty-five participants performed three randomised drives, each reflecting different passing rates, i.e., the number of vehicles overtaking the driver, or vice versa, caused by a speed limit change. Passing rates varied from as low as 90 veh/h to as high as 360 veh/h. Increasing passing rates act as a proxy for increasing levels of social pressure applied to drivers. Statistical analyses are conducted using a Linear Mixed Model (LMM) and a Generalised Estimating Equations (GEE) approach, accounting for correlation caused by the panel nature of the data. The LMM indicates that the driving behaviour indicators are indeed affected under different passing rates and that these differences are statistically significant. Results indicate that drivers in higher passing rate scenario(s) tend to accelerate and drive faster compared to the low(er) passing rate scenario(s). Further, the GEE model for speed selection indicates that drivers from different age groups and gender select different speeds in response to the impact of social pressure caused by surrounding traffic. Similarly, the GEE model for speed variation within a driver suggests differential speed variations of age groups and gender under different passing rates. Overall, this study finds pronounced effects directly related to the imposed social pressure via the surrounding traffic, where high passing rates lead to significant speed variations that increase the chances of drivers engaging in safety-critical events