Step 3: Interaction Model (model3)

1. Model Equation & Group Lines

\[\widehat{\text{Price}} = 54.2275 - 0.6203(\text{Mileage}) + 16.8630(\text{PorscheIND}) + 0.0309(\text{PorscheINT})\]

  • Jaguar (\(\text{PorscheIND} = 0\)):
    • Intercept: \(54.2275\)
    • Slope: \(-0.6203\)
    • Equation: \(\widehat{\text{Price}} = 54.2275 - 0.6203(\text{Mileage})\)
  • Porsche (\(\text{PorscheIND} = 1\)):
    • Intercept: \(54.2275 + 16.8630 = 71.0905\)
    • Slope: \(-0.6203 + 0.0309 = -0.5894\)
    • Equation: \(\widehat{\text{Price}} = 71.0905 - 0.5894(\text{Mileage})\)

2. Predictions at 30,000 Miles (\(\text{Mileage} = 30\))

  • Porsche: \(71.0905 - 0.5894(30) = 53.4084\) thousand dollars
  • Jaguar: \(54.2275 - 0.6203(30) = 35.6185\) thousand dollars
  • Predicted Difference (\(\text{Porsche} - \text{Jaguar}\)): \(53.4084 - 35.6185 = 17.7899\) thousand dollars

3. Model Selection (model2 vs. model3)

  • Visually, the two regression lines stay nearly parallel across the range of mileage (\(-0.6203\) vs. \(-0.5894\)).
  • In summary(model3), the p-value for the interaction term (PorscheINT) is \(p = 0.7803\) (\(> 0.05\)).
  • Conclusion: Because \(p = 0.7803\) is large, there is no convincing evidence that the slopes differ between Porsche and Jaguar. We should drop the interaction term and keep model2 (the parallel-lines model).