Econ 302 Exam
Study Guide for Exam IV
Econ 302/BUS 302
The exam is similar in length and form to the previous 3 exams. Bring a tall, green skinny scantron to the exam, a calculator, and a pencil. Scratch paper will be provided, if needed. You will be provided with a copy of the t-distribution on page 920 in the text, and the F-distribution on page 925. The formulas for the full F test statistic and the partial F test statistic will be given on the exam. All other formulas you will be expected to know (excluding the adjusted R2).
Exam will cover chapter 16 and anything on multiple regression. It is assumed that past material, such as population model, regression equation, estimated regression equation, error term, population parameters, etc. are already known. …show more content…
Previous Information – Know these concepts from the last exam
• Population model, regression equation, estimated regression equation, error term, population parameters • When regressors are added to the model, what happens to the SSE, SSR, or SST?
What happens to the R2? What is the adjusted R2 and why is it necessary? The formula for the R2adj will be on the exam. How do you interpret the adjusted R2?
• What is a quadratic term? When should you add a quadratic term to the model? Know how to interpret the estimated coefficients to a quadratic term.
• What is an interaction term? When should you add an interactive term to the model? Know how to interpret the estimated coefficients to an interactive term.
• What are dummy variables? When should you add them to the model? Make sure you know how to add qualitative data to a regression, such as we did with Gender or Hospital Location.
How do interpret the estimated slope coefficients to Dummy Variables? Be able to find the different population models for different values of the qualitative data.
• You should know how to do individual significance test and a Full F-test.
• What is multicollinearity? What does it cause?
• Be able to take results from a regression in Excel and interpret the results. This includes
writing the estimated regression equation, predicting a value of y, testing each regressor for individual
• Know the 4 assumptions about the error term, and what violates each of them.
• Know how to detect if the error terms have been violated.
• Know how to correct for a violation of an error term.
• Key terms such as heteroskedasticity, homoskadisticity, serial correlation, autocorrelation, time series data, and cross-sectional data will be on the exam.
Chapter 16
• You will have to do a partial F test between a reduced model and a full model. You will need to make the null and alternative hypothesis, calculate the partial F test statistic, find the critical value from the F-distribution, and interpret the results. Be able to read excel outputs from 2 separate models and pull out the information needed to calculate the partial F test statistic
(formula will be on the exam). I posted an example using tax example.xls (located in the course documents section) below.
• Know how Backwards Elimination, Forward Selection, Stepwise Regression, and Best Subsets
Regression work. Why are they important? (16.4)
•
Know the Durbin Watson test. What is the null and alternative hypothesis? I will give a dL and dU, and the Durbin Watson test statistic. You will have to interpret what the results mean. Be able to reproduce the table with the results from the test.
Partial F-test Practice Problem using tax example.xls included in blackboard:
A researcher would like to determine the effect of changes on the marginal tax rate on the highest income bracket on real GDP growth rates. The first regression uses the marginal tax rate as the only regressor. The researcher then adds regressors for Tax Receipts as a percentage of GDP, Gov. Spending as a percentage of GDP, and the Federal Funds Rate. Test the joint significance of the 3 regressors added in the second model using alpha=0.05. The reduced model and full model are located in the tax example.xls file.
Answer: H0: β2=β3=β4=0. Ha: at least one β≠0. Test sta@s@c = 1.3785. Cri@cal Value (using 48 denominator degrees of freedom) = 2.798. Conclusion: The test statistic lies inside the critical value, Do not reject H0 with α=0.05. The three regressors added to the model were jointly insignificant.
Econ 302/BUS 302
The exam is similar in length and form to the previous 3 exams. Bring a tall, green skinny scantron to the exam, a calculator, and a pencil. Scratch paper will be provided, if needed. You will be provided with a copy of the t-distribution on page 920 in the text, and the F-distribution on page 925. The formulas for the full F test statistic and the partial F test statistic will be given on the exam. All other formulas you will be expected to know (excluding the adjusted R2).
Exam will cover chapter 16 and anything on multiple regression. It is assumed that past material, such as population model, regression equation, estimated regression equation, error term, population parameters, etc. are already known. …show more content…
Previous Information – Know these concepts from the last exam
• Population model, regression equation, estimated regression equation, error term, population parameters • When regressors are added to the model, what happens to the SSE, SSR, or SST?
What happens to the R2? What is the adjusted R2 and why is it necessary? The formula for the R2adj will be on the exam. How do you interpret the adjusted R2?
• What is a quadratic term? When should you add a quadratic term to the model? Know how to interpret the estimated coefficients to a quadratic term.
• What is an interaction term? When should you add an interactive term to the model? Know how to interpret the estimated coefficients to an interactive term.
• What are dummy variables? When should you add them to the model? Make sure you know how to add qualitative data to a regression, such as we did with Gender or Hospital Location.
How do interpret the estimated slope coefficients to Dummy Variables? Be able to find the different population models for different values of the qualitative data.
• You should know how to do individual significance test and a Full F-test.
• What is multicollinearity? What does it cause?
• Be able to take results from a regression in Excel and interpret the results. This includes
writing the estimated regression equation, predicting a value of y, testing each regressor for individual
• Know the 4 assumptions about the error term, and what violates each of them.
• Know how to detect if the error terms have been violated.
• Know how to correct for a violation of an error term.
• Key terms such as heteroskedasticity, homoskadisticity, serial correlation, autocorrelation, time series data, and cross-sectional data will be on the exam.
Chapter 16
• You will have to do a partial F test between a reduced model and a full model. You will need to make the null and alternative hypothesis, calculate the partial F test statistic, find the critical value from the F-distribution, and interpret the results. Be able to read excel outputs from 2 separate models and pull out the information needed to calculate the partial F test statistic
(formula will be on the exam). I posted an example using tax example.xls (located in the course documents section) below.
• Know how Backwards Elimination, Forward Selection, Stepwise Regression, and Best Subsets
Regression work. Why are they important? (16.4)
•
Know the Durbin Watson test. What is the null and alternative hypothesis? I will give a dL and dU, and the Durbin Watson test statistic. You will have to interpret what the results mean. Be able to reproduce the table with the results from the test.
Partial F-test Practice Problem using tax example.xls included in blackboard:
A researcher would like to determine the effect of changes on the marginal tax rate on the highest income bracket on real GDP growth rates. The first regression uses the marginal tax rate as the only regressor. The researcher then adds regressors for Tax Receipts as a percentage of GDP, Gov. Spending as a percentage of GDP, and the Federal Funds Rate. Test the joint significance of the 3 regressors added in the second model using alpha=0.05. The reduced model and full model are located in the tax example.xls file.
Answer: H0: β2=β3=β4=0. Ha: at least one β≠0. Test sta@s@c = 1.3785. Cri@cal Value (using 48 denominator degrees of freedom) = 2.798. Conclusion: The test statistic lies inside the critical value, Do not reject H0 with α=0.05. The three regressors added to the model were jointly insignificant.