Reaction time (RT) is one of the dependent variables usually available to cognitive scientists that is probably used most often in psychological experiments. Making inferences based on RT data usually resolves in calculating algebraic RT means across the conditions generated through the adoption of a particular experimental design, and applying the most suited statistical test to detect significant discrepancies between these means. In addition to this traditional procedure, RTs can be submitted to more sophisticated analyses based on properties of the functions describing the RT distribution, which are recently receiving much attention in several psychological fields. These specific analyses, labelled distributional analyses, are the focus of the present review. The mathematical functions shown to provide the best fits to observed RT families are first considered, and a justification for the use of the ex-Gaussian distribution is provided. A second section is dedicated to an overview of the basic findings consequent to the adoption of distributional analyses in the context of cognitive studies. The theoretical relevance of the application of these type of analyses is discussed in the final section of the present review.