Econ 122B Problem Set 4 Name (print)_______________________ UCI ID_______________________________ Due on June 6 1. It has been conjectured that...

1. It has been conjectured that workplace bans induce smokers to quit by reducing their opportunities to smoke. Using data on a sample of 10,000 US indoor workers, you estimate the effect of workplace smoking bans on smoking. Suppose you estimate the following regression 2 1 2 3 45 smo smkban female education age age u ke =+ + + + + + αβ β β β β , where smoke is a dummy variable indicating whether a worker is a smoker or not; smkban is a dummy indicating whether there is a smoke ban at workplace, female is a dummy for female; education and age are in years. Here is what you estimate: 1) How would you interpret the estimated coefficient for smkban? 2) How would you interpret the estimated coefficient for education? 3) For a female aged 30 with 16 years of education who works at a place where smoke is allowed, the predicated value =0.157. How would you interpret it? 10. Consider the following estimated regression using a sample of . 18 Drinking =− + 0.50 0.05 0.30 , female D where Drinking is a dummy that indicates whether one drinks alcohol or not; female is a dummy that is equal to 1 if a person is female, and 0 otherwise; D18 is a dummy indicating if one exceeds the minimum legal drinking age 18. How would you interpret the estimated coefficient for D18? 2 1) How would you interpret the estimated coefficient for female? 2. The primary question of interest is this: Do athletes perform more poorly in school during the quarter their sport is in season? Suppose you have data for 200 football players from a large university for fall and spring quarters. Football players play their sport only in the fall. Suppose the academic ability levels of football players differ from each other. You try to estimate the following regression: 1 23 4 5 qrtGPA hsperc sat female fall load u =+ + + + + + αβ β β β β Where qrtGPAis quarter GPA; hsperc is high school percentile, sat is SAT score, female is a dummy indicating a female player, and fall is a dummy for the fall quarter, and load is the total number of classes taken in a quarter. 1) Compare OLS applied directly to the above regression with "difference" estimation using data differenced across the two quarters, which one do you prefer? Why? 2) If you do "difference" estimation, what variables will drop out? 3) Suppose the following is what you got. =-.065 - .12 , (.043) (.007) ∆ ∆ qrtGPA load where ∆ represents difference between fall and spring quarters. What is the estimated in-season effect? Is there a significant in-season effect? 4) How would you interpret the estimated coefficient -.12? perform more poorly in school during the quarter their sport is in season? Suppose you have data for 200 football players from a large university for fall and spring quarters. Football players play their sport only in the fall. Suppose the academic ability levels of football players differ from each other. You try to estimate the following regression: 1 23 4 5 qrtGPA hsperc sat female fall load u =+ + + + + + αβ β β β β Where qrtGPAis quarter GPA; hsperc is high school percentile, sat is SAT score, female is a dummy indicating a female player, and fall is a dummy for the fall quarter, and load is the total number of classes taken in a quarter. 1) Compare OLS applied directly to the above regression with "difference" estimation using data differenced across the two quarters, which one do you prefer? Why? 2) If you do "difference" estimation, what variables will drop out? 3) Suppose the following is what you got. =-.065 - .12 , (.043) (.007) ∆ ∆ qrtGPA load where ∆ represents difference between fall and spring quarters. What is the estimated in-season effect? Is there a significant in-season effect? 4) How would you interpret the estimated coefficient -.12?