Reference no: EM133080099

1. In a regression, explain why we include a constant term.

2. Briefly explain the major assumptions of classical linear regression model. Also discuss biasedness, efficiency and consistence of estimates.

3. Concisely discuss the classical linear regression model diagnostics focusing on causes, consequences, and solutions for Heteroscedasticity, autocorrelation, multicollinearity, non-normality as well as model specification error.

4. What is the difference between fixed effect and random effect estimators? Which test will assist you to choose between the two estimators?

5. What is a spurious regression? When such a regression does possibly occurs?

6. Distinguish logistic/probit estimator from tobit estimator. How do would you decide whether logistic/probit or tobit is an appropriate estimator to apply.

Part II: Application Section

1. Write down a stata command for generating a new variable

2. Write down a stata command useful for calculating summary statistics for a continuous variable

3. Write down a stata command useful for drawing a scatter plot between two variables say Y and X

4. Indicate a stata command for determining functional form of a continuous variable say Y

5. Write down a stata command to run a regression based on the following information - Dependent variable : y - Independent variables : r1, r2, r3, f1, f2 - Suppose that r's enter the model in logarithmic form while the f's enter the model in their level form.

6. Write down procedures for Breusch-Pagan / Cook-Weisberg and White tests for heteroskedasticity

7. Show procedures for Durbin-Watson test for autocorrelation in stata. Clearly indicate how to interpret the results.

8. Show the steps you need to follow to test for multicollinearity using stata command

9. How do you test for omitted variable bias using a stata command?

10. Outline Hausaman's test for choosing between fixed and random effect estimators.