The course covers some foundational statistics that are necessary for going forward. We then cover the classical linear regression model (CLRM). We discuss inference and estimation using the CLRM, as well as common violations of the CRLM assumptions, particularly endogeneity, heteroscedasticity, non-linear errors and omitted variable bias. We introduce participants to binary dependent variable models (Linear probability model, probit, logit) and discuss further issues one can run into with regression estimation. South African survey data will be used throughout to estimate the models covered. In all sections of the course, attention will be paid to becoming familiar with the use of Stata.
Prerequisites: Course participants should be familiar with Stata and with basic statistical tools. Don't know Stata? Have a look at our Introduction to Stata course.