In this paper we analyze statistical properties of stated preference designs for the scheduling model. The scheduling model is frequently used for estimating the willingness to pay (WTP) for a reduction in schedule delay early or schedule delay late. These WTP values can be used to calculate the costs of travel time variability. Creating a design is not an easy task, since the attributes shown to the respondents are different from the model attributes that are estimated. We derive a general method to account for this. We apply the theoretical results to the scheduling model and create an algorithm to design the choice experiment using measures for efficiency (S-efficiency and WTP-efficiency) for a design dependent on the reference travel time of the traveler. In a simulation we show that the designs based on these criteria perform on average better than designs used in the literature in terms of efficiency of the WTP. Surprisingly the orthogonal design performs really well.