Use Vote1.dta to answer the following questions. Suppose I want to study the effect of how candidate A’s expenditures affect the vote received by candidate A.
1. State the null and alternate hypothesis
2. Run a regression with voteA as the dependent variable, and ln(expendA), ln(expendB), and prtystrA as regressors.
3. How does expendA affect voteA? Do you reject the null hypothesis?
4. Now run the following regressions:
a. with voteA as the dependent variable, and ln(expendA), ln(expendB), democA as regressors.
b. with voteA as the dependent variable, and ln(expendA), ln(expendB), democA, and prtystrA as regressors.
5. What is the effect of expendA on voteA? Compare your answers from (4) with that from (3). Does your answer change?
Use the data CEOSal.dta to answer the following questions.
1. Suppose I want to test if CEO salary is affected by the ROE (returns to equity) of the company. State the null and alternate hypothesis.
2. Run a regression with salary as the dependent variable, and roe and sales as regressors. What is the effect of roe on salary? Do you reject the null?
3. Run a regression with ln(salary) as the dependent variable, and roe and sales as regressors. What is the effect of roe on salary? Do you reject the null?
4. Now use ln(salary) as the dependent variable, and roe, sales, indus, finance, and consprod as regressors. What is the effect of roe on salary? Do you reject the null?
5. Do a test of joint significance of whether finance, indus and consprod should be included in the model.