Event History 2: Understanding the Cox Model and Semi-parametric Analysis
London School of Economics
Dr Ben Wilson
Entry (no or almost no prior knowledge)
View in Google Maps (WC2A 2AE)
69 Aldwych, Columbia House
The lecture will begin with a brief overview of event history models (note that this overview will be much shorter than in ‘Event History 1’). This will include an outline of the differences between parametric, semi-parametric and non-parametric event history analysis.
After this, the rest of the lecture will focus on Cox models. These models measure time in ‘continuous units’, and make the assumption of proportional hazards. The importance of this assumption will be explained in the lecture, alongside a review of the advantages and disadvantages of the Cox model, which has become the most popular event history model in many areas of research. As with discrete-time models, Cox models can also be extended to allow a variety of advanced applications, including models that allow for unobserved heterogeneity. These models are semi-parametric because the functional form of the baseline hazard remains unspecified. All of these terms will be explained in the lecture, which will aim to provide the audience with a solid introduction to the most salient issues for applied research.
Computer seminar: In the afternoon, a computer-based seminar will allow participants to apply the techniques covered in the lecture. In addition, the aim of the seminar is to reinforce participants’ understanding of the lecture content and to introduce important aspects of applied research, including the preparation of data for analysis. Although time will be limited, there will also be an opportunity to discuss any questions relating to event history models, including questions that relate to participants’ own research.
This is the second of two MY560 courses on event history modelling (also referred to as survival analysis, duration analysis, or hazard modelling). It is not necessary to take both courses, and you are welcome to sign up for either or both.
This event will be delivered in person at the LSE campus, Columbia House room 8.13 from 10 am to 1 pm.
Website and registration:
Event History Analysis
Related publications and presentations: