Bootstrap confidence intervals for effects of electronic peak load pricing on public transport demand

In the public transport system of the city of Münster / Germany electronic ticketing has been introduced. Subsequently flexible electronic fares in public transport have been established in order to conduct peak load pricing. The pricing scheme for the new annual season ticket named FlexAbo consists of a low basic fare per month, an incremental fare per day with trips during the defined peak time limited by a maximum monthly fare.Before launching the FlexAbo a large scale stated choice survey among annual season ticket holders was conducted. They were asked to assume the FlexAbo would replace their tickets. Within the experimental choice context the incremental fare and the maximum fare per month varied according to an orthogonal design. Holders of the seasonal ticket MSAbo and the job ticket were asked whether they would still ride during peak times, delay peak-trips to off-peak times, go by car instead, ride by bicycle or even walk. The shares of shifted peak-trips were forecasted applying a discrete choice model.The results of shifted peak-trips are presented as well as their associated bootstrap confidence intervals. Percentile and bias corrected confidence intervals are compared. Additionally, all results are provided in dependency of different sample sizes. It is shown that even for small shifted shares notable differences between the two techniques for constructing confidence intervals only occur in combination with small sample sizes. Further, the width of the confidence intervals substantially decreases with sample size.