The temporal dynamics of speeded decision making

This dissertation sheds light on the temporal dynamics of behavior in speeded decision making. Participants on reaction time (RT) tasks learn, get distracted, speed up, slow down, get confused, get bored, and eventually may start guessing. One can safely say that participants' behavior is dynamic. It seems therefore obvious that mean RT or accuracy are at best limited summaries of performance. This thesis shows that a study of the dynamics of speeded decision making requires an assessment of all the data, most notably the changes in entire RT distributions. More generally, this thesis shows that the analysis of psychological data should start from well-formulated ideas about the processes that might have generated the data. When the model that formalizes these ideas provides a reasonable account of the observed data, this constitutes a first step towards a deeper and more comprehensive understanding of behavior than is possible from the assessment of summary measures alone.