Public Finance and Cost Benefit
1. What happens to the MDE (the Minimum detectable effect) if we increase the variance of the error term? What does that say about the value of including control variables in our regression model?
2. Suppose that we want to be more careful and decide that you will only reject the null hypothesis at the 1% level instead of 5% level in a two taiked test. How does this change MDE?
3. For a given MDE, how does deciding to reject the null only at the 1% instead of at the 5% level affect your chances of accepting the alternative hypothesis given that it's true?