The course will be based on the book by Per Kragh Andersen and Lene Theil Skovgaard published by Springer-Verlag (2010). The course will deal with both linear regression for quantitative responses, logistic regression for binary responses and Cox regression for survival time responses. It will discuss how the effects of covariates are combined into a linear predictor when describing characteristics of the response variable, e.g. the mean, the failure probability or the hazard rate. It discusses two- and several-sample problems for all three types of outcome (the t-test and one-way ANOVA, chi-square tests and logrank tests), models with a linear effect or a non-linear effect (splines, polynomials) of a single quantitative covariate , multiple regression models, models with interaction as well as modeling strategies.