Question

You have the position of political analyst at a TV station. In view of upcoming elections, you are asked to comment on the results from the 2016 elections. You have data on the percentage of votes received by Democratic candidates among all votes cast for each one of the 52 states. In addition, for each one of the states you have data on the unemployment rate, and you know whether President Obama appeared there to campaign for the party's candidates. You believe that the percentage of votes received by Democratic candidates in each state may be explained by the state's unemployment rate, by whether Obama campaigned in that state, and by the state's location in the country in five different regions: North, Central, South, West, East.

1. Write down a regression model for the determination of the percentage of votes for the Democratic candidate per state that expresses these beliefs. Interpret the coeffcient estimates of the model which you would obtain if you would actually use your data to estimate the regression model.

2. Describe how you would test the following hypotheses:

(a) Obama's campaigning didn't matter.

(b) The entire country voted uniformly: there are no regional differences.

(c) In addition you may think that the effect of Obama's campaigning in a state may be different depending on whether the state is in the North, Central, South, West or East. Write down a regression model for the determination of the percentage of votes for the Democratic candidate per state that also captures this effect. Interpret the coefficient estimates of the model which you would obtain if you would actually use your data to estimate the regression model. Describe how you would test the hypothesis that Obama's appearance had the same effect in all regions.

(No need for data since it is based on theory and how we solve it)

Answer 1