Dirk U. Wulff, Jonas M. B. Haslbeck, Pascal J. Kieslich, Felix Henninger, Michael Schulte-Mecklenbeck
Mouse-tracking: Detecting types in movement trajectories

This chapter presents a new approach to analyzing mouse trajectories based on trajectory clustering that overcomes the limitations of aggregation-based analyses of movement trajectories. It focuses on movement trajectories recorded using the computer mouse. The chapter argues that the aggregate-level analysis underlying the approach risks obscuring important trial-level variability in movement trajectories that can paint a different picture of the underlying cognitive process than the aggregate-level results do. It also presents a novel procedure for the analysis of mouse and hand tracking data based on cluster analysis. The chapter discusses each of the steps in the context of identifying types in movement trajectories and demonstrate the usefulness of the approach using the data of M. J. Spivey et al. A fundamental challenge in using cluster analysis for mouse and hand trajectories is that trajectories are recorded with different numbers of points. Cluster algorithms practically never produce the exact same partitioning of trajectories for different data sets.