The potential lack of realism in stated-preference surveys is particularly acute in contexts where disaggregate real-world data is challenging to obtain. Mobility-on-Demand (MOD) services present one such context. The MOD context is unique due to factors such as service reliability (difference in stated vs. actual travel and waiting time) and current mode inertia which affect the choice of MOD services and are difficult to infer from revealed preference data. Further, travel mode choices are repetitive and constitute a relatively easy choice situation. Consequently, individuals may utilize simple non-compensatory strategies. In this study, we design a survey to mimic real-world choice sets using a joint revealed and stated- (RP-SP) preference survey approach. We construct the complete journey of individuals taking into account departure time, access and egress mode, current primary mode and origin–destination pair. A Choquet Integral (CI)-based choice model with endogeneity correction is employed, thereby allowing to approximate non-compensatory behaviour. Results confirm the presence of non-compensatory behaviour across all mode users (car, public transport and bike). Reliability and inertia effects are most pronounced for car users including the potential for a combined departure time-mode shift towards MOD. Owing to non-compensatory behaviour and inertia, higher travel costs cannot be fully compensated by shorter waiting and travel times and a differential pricing strategy may be required to increase MOD market share. Failure to account for common unobserved factors between the RP and SP choices results in inflated attribute importance.

中文翻译：

---

通过情境感知调查和非补偿性策略了解对按需出行服务的偏好

在很难获得分类的现实世界数据的情况下，陈述偏好调查潜在缺乏现实性的情况尤其严重。按需移动（MOD）服务就是这样一种背景。由于服务可靠性（规定与实际旅行和等待时间的差异）和当前模式惯性等因素影响 MOD 服务的选择，并且很难从显示的偏好数据中推断出，MOD 上下文是独特的。此外，出行模式选择是重复的并且构成相对容易的选择情况。因此，个人可以利用简单的非补偿性策略。在本研究中，我们设计了一项调查，使用联合揭示和陈述（RP-SP）偏好调查方法来模拟现实世界的选择集。我们考虑出发时间、进出模式、当前主要模式和出发地-目的地对来构建个人的完整旅程。采用具有内生性校正的基于 Choquet Integral (CI) 的选择模型，从而可以近似非补偿行为。结果证实所有模式用户（汽车、公共交通和自行车）都存在非补偿性行为。可靠性和惯性效应对于汽车用户来说最为明显，包括组合出发时间模式向 MOD 转变的可能性。由于非补偿性行为和惯性，较短的等待和出行时间无法完全补偿较高的出行成本，因此可能需要采取差异化定价策略来增加 MOD 市场份额。未能考虑 RP 和 SP 选择之间常见的未观察到的因素会导致属性重要性夸大。