They are coded according to the scheme described in Section 17.4. For a technology company, technology = 1 and the other dummy variables will be 0. For a financial services company, financial = 1 and the others will be 0. If a company is “other,” all six values will be 0. Thus, we need only six variables to describe the seven possible categories.
1. Generate a correlations matrix that includes variables 2–9 and 11–16. (Variable 10 would be meaningless here because its values are 1–7 for the 7 company categories.) Do the correlations “make sense” in terms of which correlations are positive and which are negative?
2. Using y = stock price (variable 8) as the dependent variable, carry out a conventional multiple regression analysis using variables 2–7, 9, and 11–16. Excel users, see note 1, which follows. Examine the partial regression coefficients to see whether their signs match the signs of the corresponding correlation coefficients associated with stock price. What percentage of the variation in stock prices is explained by the set of predictor variables?
3. Repeat step 2, but this time perform a stepwise regression analysis. Minitab and Excel users should refer to the following notes. Identify the variables that were introduced and interpret the printout. What percentage of the variation in stock prices is explained by the reduced set of predictor variables?

  • CreatedSeptember 08, 2015
  • Files Included
Post your question