The aim of this paper is to better understand the effect of task complexity in experimental design on individual choice behavior. Although the scale factor in discrete choice models is generally normalized, studies have consistently pointed out that the imposed randomness on individual choice behavior varies as a function of task complexity. Therefore, several researchers incorporated the effect of task complexity by parameterizing the scale factor as an exponential function of task complexity. Consequently, choice behavior becomes more random with increased task complexity. Moreover, sensitivity to task complexity increases at an exponential rate with decreasing task complexity. In this paper, to relax these rigorous assumptions, the authors propose an alternative function that can be viewed as a generalization of the commonly used exponential function. The newly suggested approach is examined for both conventional utility-maximization and regret-minimization models. To assess the empirical performance of the proposed methodology, they designed a stated preference survey with seven task scenarios, differing in their degree of complexity. Estimation results evidence that the generalized parameterization of the scale factor as a function of task complexity better represents individuals’ stated choices in both utility and regret models.
|Title of host publication||Transportation Research Board 98th Annual Meeting|
|Number of pages||22|
|Publication status||Published - Jan 2019|
|Event||98th Annual Meeting of the Transportation Research Board - Walter E. Washington Convention Center, Washington, D.C., United States|
Duration: 13 Jan 2019 → 17 Jan 2019
|Conference||98th Annual Meeting of the Transportation Research Board|
|Period||13/01/19 → 17/01/19|