(GLM) A common set of statistical assumptions upon which are based a full range of methods used to study one or more continuous dependent variables and one or more independent variables, whether they are continuous or categorical. The basic concept of GLM is that the relationship between the dependent variable and the independent variables is expressed as an equation that contains a term for the weighted sum of the values of the independent variables, plus a term for everything that is unknown (an error term). The least-squares criterion determines the weight for each independent variable.