Assess fit by analyzing residuals

What you'll learn

1. Define a residual as observed minus predicted and compute residuals for data points
2. Construct a residual plot (residual versus x) with a reference line at zero
3. Interpret a residual plot - random scatter signals a good fit, a pattern signals the wrong model
4. Decide, from the residual plot, whether to keep the fitted model or refit with a different function family

Prerequisites

• Represent data on two quantitative variables with scatter plots
• Fit a function to data