From Simple to Multiple Linear Regression
What you'll learn: How to expand from one predictor variable to many, allowing you to model more complex real-world relationships.
The Core Idea
Until now, you've worked with simple linear regression: predicting an outcome using a single predictor variable (like predicting salary from years of experience). But real-world phenomena rarely depend on just one factor. Multiple linear regression extends the same least squares framework to include several predictor variables simultaneously.
Think of it like upgrading from a one-dimensional thermometer to a multi-sensor weather station. Instead of just temperature, you now measure humidity, wind speed, and pressure—all contributing to your weather forecast.
The Mathematical Extension
Simple regression uses:
y = β₀ + β₁x + ε
Multiple regression extends this to:
y = β₀ + β₁x₁ + β₂x₂ + ... + βₚxₚ + ε
Each predictor x₁, x₂, ..., xₚ gets its own coefficient β₁, β₂, ..., βₚ, representing its unique contribution while holding all other predictors constant. The intercept β₀ still represents the expected value when all predictors equal zero.
Why This Matters
Multiple regression allows you to:
- Control for confounding variables you learned about in partial correlation
- Model realistic scenarios (house prices depend on square footage, bedrooms, location, age...)
- Isolate the effect of each predictor independently
- Improve prediction accuracy by capturing more variation in the outcome
The same assumptions, diagnostics, and least squares principles still apply—they just operate in higher-dimensional space. You'll still minimize residuals, check for linearity and homoscedasticity, and interpret R-squared, but now with multiple predictors working together.
Key Takeaway: Multiple linear regression extends simple regression from one predictor to many, using the same least squares framework to estimate how each variable independently contributes to the outcome.