Abstract

This entry introduces event history and survival analysis, a set of methods adapted for data relating to the timing and occurrence of events. It starts with a review of the particularities of this type of data and the main concepts used to describe and model it. It then introduces two mathematical concepts that are the building blocks of survival analysis: the hazard rate and survivor functions. Next, it shows how these two concepts can be used to model time-to-event data. It reviews both discrete and continuous time models. Within continuous time models, three subclasses are discussed: proportional hazard models, accelerated failure time models, and the semiparametric Cox model. Finally, more complex topics such as unobserved heterogeneity, competing risks, and repeated events are briefly introduced.