Multivariate regression is a powerful statistical technique that allows researchers to explore multiple predictors simultaneously, to adjust for confounding, to test for interactions, and to improve predictions. Commonly used regression models include linear regression, logistic regression, and Cox regression. Linear regression is appropriate when the outcome variable of interest is continuous and normally distributed (although the latter assumption is critical only for small samples). For example, in a cross-sectional study of 3369 middle-aged and older men, researchers explored the relationship between 2 continuous variables, serum vitamin D levels and cognitive function, by using linear regression [1]. This article attempts to demystify linear regression by using mock and real data based on this example study.