What Are Fitted Values?
You'll learn: What fitted values are and how they relate to your regression line's predictions.
The Core Idea
When you run a simple linear regression, you estimate a line of best fit through your data points. Fitted values (also called predicted values) are the Y values that your regression line predicts for each X value in your dataset.
Think of it this way: you have actual observed data points scattered around, and you've drawn the best-fitting straight line through them. For every X value in your original data, the fitted value is what your model thinks Y should be—it's the height of the regression line at that specific X.
Why "Fitted" Values?
They're called "fitted" because they come from the line you've fit to your data. Unlike your actual observed Y values (which vary above and below the line), fitted values always sit exactly on the regression line. Every single one.
The Relationship
If you recall from computing the slope (β₁) and intercept (β₀), your regression equation is:
Ŷ = β₀ + β₁X
For each observation i in your dataset:
- Fitted value (Ŷᵢ) = β₀ + β₁Xᵢ
- Actual value (Yᵢ) = what you measured
- Residual (eᵢ) = Yᵢ - Ŷᵢ (the difference you learned about earlier)
Fitted values are essential because they represent your model's "best guess" at each data point, and comparing them to actual values tells you how well your model performs.
Key Takeaway: Fitted values are the predicted Y values generated by your regression line; they lie exactly on the line and represent what your model expects Y to be for each X in your dataset.