Interpreting Bivariate Regression Coefficients: Going Beyond The Average

Abstract

Statistics, econometrics, investment analysis, and data analysis classes often review the calculation of several types of averages, including the arithmetic mean, geometric mean, harmonic mean, and various weighted averages.  This note shows how each of these can be computed using a basic regression framework.  By recognizing when a regression model is computing one of these averages, students can properly interpret these types of regressions.  Similarly, by seeing how these averages fit into a common framework, students can have a better understanding of the different calculations.