Question 11 A new operating system is installed in every workstation at a large company. The claim of the operating system manufacturer is that the time to shut down and turn on the machine will be much faster. To test it an employee selects 32 machines and tests the combined shut down and restart time of each machine before and after the new operating system has been installed. The mean of the differences (before - after) is 13.5 seconds with a standard deviation of 18.8 seconds. The degrees of freedom for the t-test equal: 0 0 32 31 62 64 Question 12 A marketing manager has developed a regression model to predict quarterly sales of her company's mid-weight microfiber jackets based on price and amount spent on advertising. An intern suggests that she include indicator (dummy) variables for each quarter. How would you code the variables? (How many dummy variables do you need? What values would they have? Choose the correct alternative. There should be 3 variables. The variables could be spring (1 for spring, O for others), summer (1 for summer, O for others), and fall (1 for fall, 0 for others). There should be 1 indicator variable for one season. The season chosen does not matter, but it should have a value of 1 for that season and 0 for the others. There should be 1 variable. The values could be 0 for spring, 1 for summer, 2 for fall, and 3 for winter. There should be 4 variables. The variables could be spring (1 for spring, 0 for others), summer (1 for summer, 0 for others), fall (1 for fall, 0 for others), and winter (1 for winter, 0 for others). Question 13 A marketing manager has developed a regression model to predict quarterly sales of her company's mid-weight microfiber jackets based on price and amount spent on advertising. An intern suggests that she include indicator (dummy) variables for each quarter. Do you think a regression with the indicator variables would model jacket sales better than one without those predictors? No, because adding more than one indicator variable for different seasons will result in collinear independent variables, which can cause problems with the model. Yes, because adding any term to a regression model will always make a significant increase in the accuracy of the model. 0 No, because the original model has enough variables to accurately capture the change in sales with the seasons. Yes, because sales are likely seasonal, and indicator variables for the seasons can model this behavior well. Question 14 A marketing manager has developed a regression model to predict quarterly sales of her company's mid-weight microfiber jackets based on price and amount spent on advertising. An intern suggests that he include indicator (dummy) variables for each quarter. Does the intern's suggestion make sense? Choose the best alternative. C 0 Sales of jackets probably have a seasonal component. One indicator variable can model this well. The jacket prices probably change with the season. Three Indicator variables can model this well. The suggestion does not make sense. The variation with season would be captured by the linear regression model already. The amount spent on advertising probably changes with the season. A seasonal variable can model this well. Question 15 A scatterplot of US Gross ($Million) vs. Budget ($Million) for Comedy and Adventure movies showed (technology enabled us to show red colour points for Comedy movies, and blue colour points for Adventure movies), suggesting two different linear models - one for Comedy pictures, and another one for Adventure pictures. We want to code the movie genre by using an indicator variable ((Use Adventure as the base level). Choose the correct alternative. Comedy 1, Adventure = 0 Comedy C, Adventure = A Comedy 0, Adventure = 1 Comedy 1, Adventure = 2