Eco 231 - Undergraduate Econometrics Homework 2
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
Homework 2
Eco 231 - Undergraduate Econometrics
Instructions: At the moment of submitting your homework, please use the following structure to write your do ile:
|
cls clear all
// Part a use HW2a.dta
/* Answer for part a */
// Part b use HW2b.dta
/* Answer for part b */ |
1. In this question you will work with the dataset HW2a.dta, which contains information about arrest records and characteristics of a group of individuals in 1986. All the variables in this dataset have been labeled. We are interested in understanding the determinants of crime and howefective incarceration is.
Using this information answer the following questions:
(a) Create the variable race equals to 1 if the individual is white, and 0 if not (black or hispanic). Show a table providing summary statistics (mean, standard deviation) of narr86, qemp86,inc86, tottime by race.
(b) Run the following regression:
narr86i = β0 + β1pcnvi + β2 avgseni + β3tottime i + β4ptime86i + β5 qemp86i + Ei
Interpret the result of each coefficient.
(c) Calculate the standard error for the estimate of avgsen. To do this, you will need to apply the formula seen in class and calculate all the relevant elements separately. (Notice that this is not equivalent to just reporting the value from the regression output).
(d) Using your results from part (c) compute a 95% conidence interval for the estimate of avgsen.
(e) Plot the residuals of the regression. Do the errors of the regression appear to be homoskedastic or heteroskedastic? [Hint: you can use the rvfplot command].
(f) Create a new variable avgsen2 equals to the average sentence lenght squared. Run the same regression of part (b) including avgsen2. How does the estimate of avgsen change?
(g) Run the following regression, employing heteroskedasticity-robust standard errors:
narr86i =α0 + α1pcnvi + α2 avgseni + α3 avgseni(2) + α4ptime86i + α5 qemp86i+
α6inc86i + α7race i + α8tottime i + νi
Is there any change in the estimated coe伍cient of tottime?
(h) Name one potential confounder in the speciication of part (g). Explain your reasoning.
2. The dataset HW2b.dta contains information about a job training experiment for a group of men conducted during the years 1976-1977. The objective in this question is to test whether participation in this training program had an efecton the unemployment probability and earnings in 1978.
(a) Create a table showing descriptive statistics (mean and standard deviation) of age, educ, black, married, lre74, lre75, lre78 according to job training status. Do you observe relevant diferences between both groups?
(b) Run a simple regression of lre78 on train. What is the estimated efect of participating in job training on real earnings in 1978?
(c) Calculate the standard error of the estimate of train. (As in Question 1 part (c), this is not the same as reporting the value from the regression output).
(d) Now add as controls to the regression in part (b) the variables re74, re75, educ, age, black, and hisp. Compare your results of the efect of job training with respect to part (b).
(e) Create a variable minor equal to 1 if the individual is black or hispanic. Employing this variable incorporate an interaction of minor and train in the regression you run on part (d).
(f) Repeat the regression of part (d) including controls for unemployment in 1974 and 1975. How does the estimate of train change?
(g) Repeat the speciication used in part (f), including now unem78 as the dependent variable. Interpret the value of the estimate for train in this case.
2024-03-24