# The dataset SMOKE contains information on factors whichdetermine an individual’s smoking habits. The

The dataset SMOKE contains information on factors whichdetermine an individual’s smoking habits. The information includeson an individual’s race (a binary variable
white equalsone if individual is white and zero otherwise), age (
age),income (
income), years of education (
educ),statewide smoking restrictions in restaurants (binary variable
restaurn equal one if restaurants restrict smoking andzero otherwise), price of cigarettes (
cigpric), and thenumber of cigarettes an individual smokes (
cigs). Use this dataset from
http:/fmwww.bc.edu/ec-p/data/Wooldridge/smoke to estimatea linear probability (LPM) and probit model. smoker= β0 + β1 lcigpric +β2 educ + β3 age + β4 agesq + β5income
+δ
0
white
+ δ
1
restaurn
+ u where
smoker is a binary dependent variable equal toone if an individual smokes (cigs >=1) and zero otherwise,
lcigpric is the log of cigarette prices,
agesq isage2 and
u is the error term. You need to createthe binary dependent variable
smoker. Explain the difference in the estimated partial effects of
lcigpric, educ, restaurn, and
age for a 25 yearold on the probability of smoking between the linear probabilitymodel and probit models. Calculate and interpret the averagepartial effect (APE), and partial effect at the average (PEA) for
lcigpric, educ, restaurn, and
age for a 25 yearold in the probit model?
Hint: First test forheteroscedasticity for educ, income age agesq variables in the LPMmodel using
hettest
afterregress.
.
Then estimate both models withheteroscedasticity-robust standard errors using the option
vce(robust)
for both models.
Calculate the average partial effects for the probitmodel using the
margins, dydx (*)
commandafter probit.
Im having an unsually difficult time with this question,any help would be greatly appreciated Attached