Clustering eye movement transitions reveals latent cognitive strategies

  • Šimon Kucharský University of Amsterdam
  • Ingmar Visser University of Amsterdam
  • Gabriela-Olivia Truțescu Leiden University
  • Paulo Guirro Laurence Mackenzie Presbyterian University
  • Martina Zaharieva University of Amsterdam
  • Maartje E. J. Raijmakers University of Amsterdam
Keywords: Cognitive strategies, eye-tracking, unsupervised clustering, latent groups, scanpaths

Abstract

In cognitive tasks, solvers can adopt different strategies to process information which may lead to different response behavior. These strategies might elicit different eye movement patterns which can thus provide substantial information about the strategy a person uses. However, these strategies are usually hidden and need to be inferred from the data. After an overview of existing techniques which use eye movement data for the identification of latent cognitive strategies, we present a relatively easy to apply unsupervised method to cluster eye movement recordings to detect groups of different solution processes that are applied in solving the task. We test the method's performance using simulations and demonstrate its use on two examples of empirical data. Our analyses are in line with presence of different solving strategies in a Mastermind game, and suggest new insights to strategic patterns in solving Progressive matrices tasks.

Published
26-02-2020
How to Cite
Kucharský, Šimon, Visser, I., Truțescu, G.-O., Laurence, P., Zaharieva, M., & Raijmakers, M. (2020). Clustering eye movement transitions reveals latent cognitive strategies. Journal of Eye Movement Research, 13(1). https://doi.org/10.16910/jemr.13.1.1
Section
Articles