Replication of Wayfinding Studies in Different Geographic Areas. A Simulation Study

Abstract

Replication of real-world wayfinding studies is not a trivial task. Even less if it is to be replicated in a different geographic environment. The selection of one or several routes is one of many decisions to be made. Only recently (2021), a reproducible, systematic and score-based approach for route selection for wayfinding experiments was published. Besides allowing for selecting a route within a selected experimental area, it claims to be able to find similar routes in different geographic areas. However, it remains unclear if similar, according to this route selection framework, routes lead to similar study results. In order to answer this question, an agent-based simulation comparing Turn-by-Turn and Free Choice Navigation approaches (between-subject design) is run in one European (Vienna) and one African (Djibouti City) city. First, a route in Vienna is selected and, second, the 5 most and the 5 least similar routes in Djibouti City are found. These routes are used in the simulation in order to scrutinize if more similar routes lead to more similar results regarding the arrival rate as a metric. The results suggest that the route selection framework is suitable for replication studies for the Turn-By-Turn navigation approach but needs further improvement for the Free Choice Navigation approach by adding features describing the neighborhood of the route.