Discrete choice studies are increasingly used in urban planning to understand preferences and to make informed decisions based on its outcomes. Traditional discrete choice modelling approaches have evolved in a setting in which some very specific behavioural assumptions are made in specifying decision-making. These assumptions have given rise to the study of alternative process strategies in decision-making, such as majority of confirming dimensions (MCD), attribute non-attendance (ANA), or value learning (VL). In this paper, a stated choice experiment was designed to understand business location decisions, where a location specialist had to compare their current location with two alternative locations. After each choice task, respondents were asked whether they used ANA in processing the choice tasks, and at the end of the experiment a number of questions were asked to identify whether specific process heuristics were used such as MCD and VL. Choice models were estimated to compare the influence of including different stated heuristics responses as conditioning effects. The results show that the model which included the stated heuristics responses is superior in terms of the goodness of fit and in the estimates’ significance levels than the model that assumes everyone might be using the heuristics. The willingness to pay estimates derived from a traditional model were statistically equivalent to the ones derived from the stated multiple heuristics model. However, the median WTP derived from the stated multiple heuristics model was slightly lower and the confidence intervals higher than in the traditional model.

离散选择研究越来越多地用于城市规划中，以了解偏好并根据其结果做出明智的决定。传统的离散选择建模方法是在一种环境中发展的，在这种环境中，在指定决策时会做出一些非常具体的行为假设。这些假设引起了对决策中替代过程策略的研究，例如大多数确认维度（MCD），属性非出席（ANA）或价值学习（VL）。在本文中，设计了一个明确的选择实验来理解业务位置决策，其中位置专家必须将其当前位置与两个替代位置进行比较。每次选择任务完成后，都会询问受访者是否在处理选择任务时使用了ANA，在实验结束时，提出了许多问题，以确定是否使用了特定的过程启发式方法，例如MCD和VL。估计选择模型以比较包括不同陈述启发式响应为条件效应。结果表明，包含拟定启发式响应的模型在拟合优度和估计的显着性水平方面均优于假设每个人都可能使用启发式的模型。从统计上讲，从传统模型得出的支付意愿与从所述多重启发式模型得出的支付意愿在统计上是相等的。但是，与传统模型相比，从所述多重启发式模型得出的中位数WTP稍低，而置信区间则更高。