Why do drivers divert? Impact of graphical route information panels

This study applied a stated preference survey method to investigate the impact of graphical route information panels GRIPs on taxi drivers' diversion decisions. A cross-sectional ordered logit model and a panel data ordered logit model were developed. Main findings are: taxi drivers and mid-age drivers with high urban freeway use frequency are more likely to divert; drivers are more likely to divert if the original ramp is jammed and an accident occurs; employer-provided car drivers are less sensitive to accident and less likely to divert. On the modelling aspect, the panel data model does not provide substantially different model coefficients but more robust statistical inferences; the cross-sectional model tends to seriously overestimate t-test values for explanatory variables changing across drivers and tends to slightly underestimate t-test values for explanatory variables changing across trip scenarios and interaction terms. The findings have implications for GRIP assessment and for future response behaviour modelling effort.