S3E17: Logistic Regression: 2 Logit 2 Quit
Quantitude
Beware the Limits of Linear Models
Using a linear model for binary outcomes presents significant issues. First, linear models predict probabilities that can fall outside the valid range of 0 to 1, leading to nonsensical outcomes, such as probabilities exceeding one. Second, the residuals in such models are heteroscedastic and non-normally distributed, violating core assumptions of ordinary least squares regression. These inconsistencies highlight the inappropriate application of linear regression to binary data, where predictions can become unreasonable.
00:00
Transcript
Play full episode
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
