In classical statistics, linear regression is regarded as the ‘gateway to predictive modelling’. For decades students have been taught about regression from theory to practice simply because it still one of the most versatile and simple ways to understand and predict the effect of key factors on critical outcomes.

In this video Jarlath Quinn shows how you can move beyond simple linear regression to model curvilinear relationships using techniques such as variable transformations and quadratic regression before finally exploring how log-log regression can be used to model price elasticity of demand.

The video explores moderation analysis, which enables analysts to identify interaction effects that alter the relationship between a dependent and independent variable.

This video explores mediation analysis – an analytical approach used to test if a third factor could represent the underlying cause of a relationship between an independent and dependent variable.