Route choice in individuals - semantic network navigation

In a novel experimental task, individuals are asked to navigate from a start word to a goal word through a semantic network. In this forced-choice task, individuals perform with a high success rate (73%) and frequently navigate to the target in the minimal number of required steps (22%). We utilize these experimental results to explore different search and decision strategies. Our descriptive modeling results suggest individuals are not guessing at random (or utilizing only local information) and that knowledge of the global structure is necessary for individuals to succeed. We further show that a latent semantic space model, such as word association space, can capture much of the global semantic knowledge necessary to explain participant decisions. We suggest that performance in this task might capture some of the underlying structure of semantic memory and, importantly, search within memory.