Question:
Off-road motorcycles (often called dirt bikes) are a segment (about 18%) of the growing motorcycle market. Because dirt bikes offer great variation in features, they are a good market segment to study to learn about which features account for the cost (manufacturers suggested retail price, MSRP) of a bike. Researchers collected data on 2005-model dirt bikes (lib .stat.cmu.edu/datasets/dirtbike_aug.csv). Their original goal was to study market differentiation among brands. (The Dirt on Bikes: An Illustration of CART Models for Brand Differentiation, Jiang Lu, Joseph B. Kadane, and Peter Boatwright). In Chapter 17, Exercises 41, 42, and 43 dealt with these data, but several bikes were removed from those data to simplify the analysis. Now well take on the full set. Heres a regression model and some associated graphs.
Well, in honesty, weve removed one luxury handmade bike whose
MSRP was $19,500 as a clearly identified outlier.
a) List aspects of this regression model that lead to the conclusion that it is likely to be a useful model.
b) What aspects of the displays indicate that the model is a good one?
Transcribed Image Text:
Dependent variable is: MSRP R-squared-91.0% R-squared (adjusted)-90.5% s606.4 with 100 6 94 degrees of freedom Source Regression 349911096 Residual Sum of Squares df Mean Square F-ratio 5 69982219 94 367733 190 34566886 Coeff -5514.66 826.2 Variable Intercept Bore Clearance Engine Strokes-315.812 89.83 -3.52 Total Weight Wheelbase SE(Coeff) t-ratio P-value -6.67 <0,0001 83.7950 145 13.6 <0.0001 2.93 0.0042 0.0007 -13.8502 3.0174.59 0.0001 0.0008 152.617 52.02 119.138 34.26 3.48 2 750 2 750 -750 t 2000 4000 6000 Predicted 750 -750 -1.25 0.00 125 Nscores 10 0.00 0.08 Leverages 0.16