Session 2: Linear and logistic regression as Generalized Linear Models

Lecture

Learning objectives

1. define generalized linear models (GLM)
2. define linear and logistic regression as special cases of GLMs
3. distinguish between additive and multiplicative models
4. define Pearson and deviance residuals
5. describe application of the Wald test

Outline

1. Brief overview of multiple regression (Vittinghoff 4.1-4.3)
2. Linear Regression as a GLM (Vittinghoff 4.1-4.3)
3. Logistic Regression as a GLM (Vittinghoff 5.1-5.3)
4. Statistical inference for logistic regression (Vittinghoff 5.1-5.3)

Lab

Learning Objectives

1. Define "tidy" data
2. Load a dataset in R and perform basic exploratory data analysis
3. Create descriptive "Table 1" of a study sample using the tableone package
4. Create a customized bar plot using the ggplot2 package

Exercises

1. Load the contraceptive use dataset into R
2. Create an "Epi Table 1" of sample characteristics
3. Create a barplot stratified by age and showing the relative proportions of participants using contraceptives among those who do and do not want more children.
4. Repeat the barplot showing percentages instead of counts