ABSTRACT
This chapter generalizes the univariate regression model outlined previously to multiple or multivariate regression, a statistical tool that allows us to examine how a number of explanatory variables can be used to explain causal relationships. It introduces additional important concepts associated with estimating a multiple regression model. The chapter utilizes model selection criteria to help select significant variables to be included in the final model. A multiple regression model containing both variables (stock and private real estate market returns) would allow us to quantify the impact of the private market on real estate investment trust returns whilst controlling for general stock market influences. In a multiple regression model, the chapter examines and quantifies the impact of age on house prices controlling for size. Maximising the value of is considered a soft rule since the authors may still end up with a large model and coefficients which make almost meaningless (marginal) contributions.
