Mechanics and Techniques Problems
9.1. A regional express delivery company has asked you to estimate the price of shipping a package
based on the durability of the package. You randomly sample the packages, making sure
that you get packages that are all about the same size and are being shipped about
the same distance. The company rates the durability of a package as either ”durable”,
”semifragile” or ”fragile”. Your data on fifteen packages is in the file C09 Shipping.xls
[.rda].
- Formulate a multiple regression model to predict the cost of shipping a package as a
function of its durability.
- Interpret the regression coefficients and the quality of your model.
- According to your model, what type of package is the most expensive to ship? Which
is the least expensive to ship?
- Use your model to predict the cost of shipping a semifragile package.
- Why is it important that the packages sampled in the data are all ”about the same
size” and ”shipped about the same distance”?
9.2. Consider the housing data in C09 Homes.xls [.rda]. We are going to build a model using the
location and style of the home, along with some of the numerical variables, to see how these affect
the price, and whether they are significant. You may want to use a table like the one below to
record your work.
- First create dummy variables for the location and the style variables.
- Formulate a multiple regression model using the location data and the numerical
variables Age, Size, Taxes, and Baths. Comment on the interpretation of this model
and its quality. Finally, comment on whether this model proves the old adage ”The
three most important things in real estate are location, location, location.”
- Formulate a multiple regression model using the style data and the numerical variables
Age, Size, Taxes, and Baths. Comment on the interpretation of this model and its
quality. Compare it to the location model you created in part b.
- Formulate a multiple regression model using the same numerical variables as before,
and using both the style and location data. How does this model compare with the
previous two models?
- Which of the models (just the numerical, numerical plus location, numerical plus style,
or numerical plus style and location) would you recommend that the realtor use for
making pricing decisions? Why?