King Soda, a nation-wide bottling company, notices that their bestselling product does not conform to the mean weight of 30 ounces in each container. Management suspects that the temperature in their three plants in Kansas City, Chicago, and Denver may explain the malfunction. A study of 40 bottles, some taken from each location, produces the data contained in the data set KING SODA. Fills are measured in ounces, temperature in Fahrenheit, and the cities in which the bottles were filled is specified.

Access the data file. Using your preferred software regress Fills on the other variables and determine the residuals. Notice that data for all three cities (dummy variables) is given. You need only two cities in the model our examination of dummy variables. Eliminate Denver from the model in your estimation process.

a. Plot the residuals and comment on the relevant features.

b. Does it appear the residuals are random?

c. What observations might you make about the potential for heteroscedasticity?

d. What other comments and remarks can you offer for the violations of all the basic assumptions of the OLS model?

e. Which variable, temperature or Chicago, has the greater impact on Fills?