This session will introduce the use of natural experiments in the social sciences to provide potentially stronger evidence for causal claims than in observational quantitative studies. The training will describe why study design matters for causal claims and will present a selection of case-studies that employ natural experiments. Students will be introduced to the quantitative techniques used to analyse natural experiments and will provided with a hands on opportunity to apply a differences in differences analysis. Students will use Stata and must be familiar with this software (including the use of syntax) and have experience and understanding of generalised linear models. The training will be delivered through pre-recorded videos although the instructor will run a live online introduction at the start and a Q&A at the end. Some pre-course reading will be provided.
Delivered by Dr Alan Marshall, Senior Lecturer in Quantitative Methods, School of Social and Political Science, University of Edinburgh