Building on your Computer Exercise #1 – Remember to attach all your relevant computer outputs to your work.
Now we move to the multiple regression framework.
(a) Think about the concept of omitted variable bias – does the simple regression you run in Exercise #1 give an unbiased estimate of the relationship between X and Y? Think of one other variable in your dataset, if omitted, that can potentially cause your estimate in Exercise 1 to be biased.
(b) Think of three additional variables you might want to add as controls in your regression model. At most two of the variables have to be qualitative variables (dummy or categorical). You could use just one dummy variable if you so choose.
(c) Think of what other modifications you might want to add to your regression model and different types of specifications/functional forms (logs, maybe?) you might want to add to improve your model.
Using Eviews:
1. Add the new variable in (a) above to you simple regression in Exercise #1 and run a regression with the two independent variables. (Regression 1)
2. Add the three variables in (b) above to your new multiple regression model (You should have 5 independent variables in this regression). Run a regression with the 5 independent variables. (Regression 2)
3. Try 3 different functional/specification forms on your regression in (2) above, this is in addition to your original linear model) (You might try logs, non-linearity, quadratic etc.). (Regression 3,4,5)
Note that for this computer Exercise #2, you would have to run 5 regressions in all [For the purposes of your report/ instructions below let’s label them as Regression 1(the one with 2 independent variables) Regression 2 (the original linear regression with 5 independent variables at levels) and Regressions 3, 4 and 5 (the 3 regressions, each with a different functional form/specification of the model with the 5 independent variables
Report on Exercise #2 – Requiremnets/Questions for Assignment
• Interpret Regression 1 fully (3 lines) (2 point)
• Referring to your simple regression from Task 1 and Regression 1 from this Task, provide a discussion on whether and why the omission of the new variable from your simple regression may have biased the estimated effect of your original independent variable. (3 lines) (2 point)
• Report and interpret Regression 2 (4 lines) (2 point)
• Test the significance of your qualitative variable(s) at 5%. Be sure to explain your conclusions fully. (1 point)
• Test if the independent variables in Regression 1 are jointly significant in Regression 2. (2 point)
• Report and interpret Regressions 3, Regression 4 and Regression 5 (6 lines) (4 point)
• Between Regression 2, Regression 3, Regression 4 and Regression 5, pick the estimated regression equation that best fits your data. What criteria did you use to come to this conclusion? (3 lines) (2 point)
• Test the overall significance of the “best” model you have chosen as your ‘best’ model (1 point)
Last Completed Projects
| topic title | academic level | Writer | delivered |
|---|