Mouse-tracking is an increasingly popular process-tracing method. It builds on the assumption that the continuity of cognitive processing leaks into the continuity of mouse movements. Because this assumption is the prerequisite for meaningful reverse inference, it is an important question whether the assumed interaction between continuous processing and movement might be influenced by the methodological setup of the measurement. Here we studied the impacts of three commonly occurring methodological variations on the quality of mouse-tracking measures, and hence, on the reported cognitive effects. We used a mouse-tracking version of a classical intertemporal choice task that had previously been used to examine the dynamics of temporal discounting and the date-delay effect (Dshemuchadse, Scherbaum, & Goschke, 2013). The data from this previous study also served as a benchmark condition in our experimental design. Between studies, we varied the starting procedure. Within the new study, we varied the response procedure and the stimulus position. The starting procedure had the strongest influence on common mouse-tracking measures, and therefore on the cognitive effects. The effects of the response procedure and the stimulus position were weaker and less pronounced. The results suggest that the methodological setup crucially influences the interaction between continuous processing and mouse movement. We conclude that the methodological setup is of high importance for the validity of mouse-tracking as a process-tracing method. Finally, we discuss the need for standardized mouse-tracking setups, for which we provide recommendations, and present two promising lines of research toward obtaining an evidence-based gold standard of mouse-tracking. Keywords Mouse-tracking. Action dynamics. Process-tracing. Boundary conditions. Intertemporal choice Decision science has experienced a paradigmatic shift evolving its focus, methods, and approaches from an outcomebased perspective toward a more process-oriented paradigm (Oppenheimer & Kelso, 2015). This process paradigm acknowledges the temporal nature of basic mental processes and, hence, builds theories of choice incorporating perceptual, attentional, memory, and decisional processes. To test these process explanations, process-tracing methods are required. In the last 60 years, decision scientists introduced a variety of process-tracing methods to the field-for example, verbal protocols (e.g., Ericson & Simon, 1984), eye tracking (e.g., Russo & Rosen, 1975), and most recently, mouse-tracking (e.g., Dale, Kehoe, & Spivey, 2007; Spivey, Grosjean, & Knoblich, 2005) (for an overview, please see Schulte-Mecklenbeck et al., 2017). Whenever scientists apply such process-tracing methods, they rely on specific prerequisites and core concepts in order to conduct the reverse inference (Poldrack, 2006): Reverse inference describes the reasoning by which the presence of a particular cognitive process is inferred from a pattern of neuroimaging or behavioral data (cf. Heit, 2015). One prerequi...

No abstract

Mouse-tracking is an increasingly popular process-tracing method. It builds on the assumption that the continuity of cognitive processing leaks into the continuity of mouse movements. As this assumption is the prerequisite for a meaningful reverse inference, it is an important question whether the assumed interaction between continuous processing and movement might be influenced by the methodological setup of the measurement. Here, we study the impact of three commonly occurring methodological variations on the quality of mouse-tracking measures, and hence, the reported cognitive effects. We used a mouse-tracking version of a classical intertemporal choice task that had previously been used to examine the dynamics of temporal discounting and the date-delay effect (Dshemuchadse, Scherbaum, & Goschke, 2013). The data from this previous study also served as a benchmark condition in our experimental design. Between studies, we varied the starting procedure. Within the new study, we varied the response procedure and the stimuli position. The starting procedure had the strongest influence on common mouse-tracking measures and therefore the cognitive effects. The effect of the response procedure and the stimuli position was weaker and less pronounced. The results suggest that the methodological setup crucially influences the interaction between continuous processing and mouse movement. We conclude that the methodological setup is of high importance for the validity of mouse-tracking as a process-tracing method. Finally, we discuss the need for standardized mouse-tracking setups for which we provide recommendations, and present two promising lines of research to obtain an evidence-based gold standard of mouse-tracking.

No abstract

Eye-tracking allows researchers to infer cognitive processes from eye movements that are classified into distinct events. Parsing the events is typically done by algorithms. Here we aim at developing an unsupervised, generative model that can be fitted to eye-movement data using maximum likelihood estimation. This approach allows hypothesis testing about fitted models, next to being a method for classification. We developed gazeHMM, an algorithm that uses a hidden Markov model as a generative model, has few critical parameters to be set by users, and does not require human coded data as input. The algorithm classifies gaze data into fixations, saccades, and optionally postsaccadic oscillations and smooth pursuits. We evaluated gazeHMM’s performance in a simulation study, showing that it successfully recovered hidden Markov model parameters and hidden states. Parameters were less well recovered when we included a smooth pursuit state and/or added even small noise to simulated data. We applied generative models with different numbers of events to benchmark data. Comparing them indicated that hidden Markov models with more events than expected had most likely generated the data. We also applied the full algorithm to benchmark data and assessed its similarity to human coding and other algorithms. For static stimuli, gazeHMM showed high similarity and outperformed other algorithms in this regard. For dynamic stimuli, gazeHMM tended to rapidly switch between fixations and smooth pursuits but still displayed higher similarity than most other algorithms. Concluding that gazeHMM can be used in practice, we recommend parsing smooth pursuits only for exploratory purposes. Future hidden Markov model algorithms could use covariates to better capture eye movement processes and explicitly model event durations to classify smooth pursuits more accurately.