Emmanuel Ogundimu - Recent advances in sample selection models
|Starts:||14:00 11 May 2022|
|Ends:||15:00 11 May 2022|
|What is it:||Seminar|
|Organiser:||Department of Mathematics|
|Who is it for:||University staff, External researchers, Current University students|
Emmanuel Ogundimu, Assistant Professor of Statistics in the Department of Mathematical Sciences, Durham University is our speaker for the Statistics seminar series.
Title: Recent advances in sample selection models
Abstract: Sample selection arises when the outcome of interest is partially observed in a study. In the presence of sample selection, the observed data is not representative of the original population, even after controlling for explanatory variables. That is, data are missing not at random and the usual assumption that missing data process is independent of the outcome given the observed variables is not tenable. Although sophisticated statistical methods in the parametric, semiparametric, and non-parametric framework have been proposed to solve this problem, classical estimators introduced by James Heckman remain popular. Apart from the fact that these estimators are very sensitive to small deviations from the distributional assumptions, which are often not satisfied in practice, there is a need for the so-called exclusion restriction. That is, some of the variables affecting the missing data process do not affect the outcome. The drive to establish this requirement often leads to the inclusion of irrelevant variables in the model.
In this talk, I will motivate the problem of sample selection from non-random treatment allocation perspective and discuss the general modelling framework. In particular, the link between sample selection models and skew-distributions will be established. A new result on regularized sample selection model and some open problems in this framework will conclude the talk.
Organisation: Durham University
Travel and Contact Information
Zoom link: https://zoom.us/j/92947173491