Consider a simple logistic regression model that relates 30 day mortality after acute myocardial infarction to two predictors: sex (coded 0 for men, 1 for women) and age (coded in years). This model assumes that the distribution of 30 day mortality follows a binomial distribution, which is quite natural for a binary outcome. Other assumptions are linearity in age and additivity of the effects of sex and age.