On page 23 of the “Forecasting the 1992 Election” article are the results of four regression models. The first three have the same dependent variable, but have differing model specifications.
1. Explain the differences between the first three regressions. What is the dependent variable? What are the explanatory variables?
2. Interpret the regression coefficients of each model.
3. Can you conduct appropriate t-tests on the coefficients of each model (exclude the constant)? Why or why not?
4. Evaluate the three regressions model’s goodness of fit.
5. Using the first three regressions, conduct an appropriate F-test to test whether:
a.. the party seeking a third term variable is significant.
b. either the seat change in midterm election or the incumbent received at least 60% of the primary vote variables are significant.
Note: List both the restricted and unrestricted models and the respective hypotheses that you’re using for a and b.
6. Obama won the 2012 election w/ 51.1% of the popular vote. Would the LBR1, Abramowitz and LBR2 models have predicted an Obama win? Note: Obama’s July 2012 approval level was 46%, GDP growth in the first 6 months of 2012 averaged 1.7% and the number of congressional seats gained by Republicans in the 2010 midterm election was 69.