An Introduction to Good Practices in Cognitive Modeling Andrew Heathcote, Scott D. Brown and Eric-Jan Wagenmakers Abstract Cognitive modeling can provide important insights into the underlying causes of behavior, but the validity of those insights rests on careful model devel- opment and checking. We provide guidelines on five important aspects of the prac- tice of cognitive modeling: parameter recovery, testing selective influence of exper- imental manipulations on model parameters, quantifying uncertainty in parameter estimates, testing and displaying model fit, and selecting among different model parameterizations and types of models. Each aspect is illustrated with examples. 1 Introduction One of the central challenges for the study of the human mind is that cognitive processes cannot be directly observed. For example, most cognitive scientists feel confident that people can shift their attention, retrieve episodes from memory, and accumulate sensory information over time; unfortunately, these processes are latent and can only be measured indirectly, through their impact on overt behavior, such as task performance. Another challenge, one that exacerbates the first, is that task performance is often the end result of an unknown combination of several different cognitive processes. Consider the task of deciding quickly whether an almost vertical line tilts slightly Andrew Heathcote University of Newcastle, School of Psychology, University Avenue, Callaghan, 2308, NSW, Aus- tralia e-mail: Andrew.Heathcote@newcastle.edu.au Scott D. Brown University of Newcastle, School of Psychology, University Avenue, Callaghan, 2308, NSW, Aus- tralia e-mail: Scott.Brown@newcastle.edu.au Eric-Jan Wagenmakers University of Amsterdam, Department of Psychological Methods, Weesperplein 4, 1018 XA Am- sterdam, The Netherlands e-mail: EJ.Wagenmakers@gmail.com 1