Module section description: Multivariate statistical methods simultaneously analyse data from several variables. In broad sense multivariate methods include all methods which require measurements from several variables, including multiple regression, ANOVA or t-test. Most methods for describing and testing associations could be called multivariate statistical methods in broad sense.

In statistical literature an method is called univariate if it just models the distribution of one variable (how it's distribution will change if one changes some other variable). But if one models several variables jointly, the corresponding method is called multivariate. For example, multivariate regression model is a model describing how the joint distribution of weigth and height depend on age. But multiple regression model describing how the weight depends on age and height would not be classified in statistical literature as multivariate model. Examples of multivariate methods in stricter sense include MANOVA, principal component analysis (describes how the set of observed variables depend on unobserved principal components), cluster analysis (several observed variables depend on unobserved variable containing cluster identificators) etc.