The variance sum law and its implications for modelling

Abstract

Basic understanding of linear model is primordial in quantitative psychology and modelling. As an advanced topic for students in psychology, modelling data according to given sets of parameters can be challenging. In this paper, we address the variance sum law which states how to combine variance of variables into a new one. The paper uses a progressive pedagogical approach, presenting step by step the statistical prerequisites. An example with code in R is offered to support the presentation. Some matrix algebra is gently introduced to stimulate readers to more advanced topics.