Logistic Regression with R

Course Overview

Who this course is for

If you have completed the Linear Regression Short Course, you have the foundation to build a wide range of statistical models for data analysis. However, linear regression is not suitable for all types of data, and it is common to encounter data in which the outcome variable is binary (e.g., success/failure, yes/no). Logistic regression is a powerful tool for modeling binary outcomes, and this course will guide you through its concepts and applications.

What you will learn

Understand the difference between linear and logistic regression.
Recognize when logistic regression is appropriate for binary outcomes.
Translate between probability, odds, and log-odds.
Fit logistic models in R with glm() and read the output.
Interpret coefficients as odds ratios and communicate effects clearly.
Assess model performance with likelihood measures and classification metrics.
Make predictions and evaluate classification accuracy.

Prerequisites

You have completed the Linear Regression short course (or equivalent).
You can run basic R scripts and read simple model summaries.

Course map

When to Use Logistic Regression when to use logistic regression (and when linear regression fails).
Introduction to Logistic Regression (GLMs, odds/logit, logistic function).
Fitting Logistic Regression Models logistic regression models (estimation, interpretation, inference).
?sec-assessing model fit (deviance, model comparison, pseudo R-squared).
Predictions from Logistic Regression Models from logistic regression models (probabilities, thresholds, ROC/AUC).
?sec-extensions and mini case study (optional: multivariable models, interactions, reporting).

How to use this book

The course is designed to be completed sequentially: successive chapters build on prior concepts. However, it is possible to jump to specific sections as needed.
There is a glossary at the end for quick reference of key terms.
Each chapter includes excersises which should be attempted before continuing.