我们总结了Brown University EDUC2360的经典案例，如果你有任何STATA代写的需求，请随时联系我们。

EDUC 2360

Problem Set 1

OVERVIEW

In this problem set, you will explore data from a well-known randomized controlled trial (RCT) in the mid-1970s, the National Supported Work Demonstration. This was a job training program designed to provide employment and counseling to low-income workers who lacked the basic skills to enter the labor market. The program was evaluated through an RCT, in which individuals in the treatment group were offered the full program and those in the control group received no additional services.

The NSW Demonstration has been evaluated in several influential papers. Most importantly, it has formed the basis for two analyses that compared the results from the RCT to those that would have been produced had the authors used a standard observational design. LaLonde (1986) explored whether non-experimental evaluations produced similar estimates of the treatment effect.

You will re-analyze these data. The problem set will lead you through the analysis. You should turn in a set of answers to the questions below, organizing your answers explicitly in response to the questions. In other words, I would include the question and type your response immediately below. Most of your responses can be a sentence or a brief paragraph, although you should frame your responses in text (e.g., don’t just put the number) and explain your reasoning (e.g., “Yes.” is not a sufficient response). While I do not require specific formats for tables, etc., in general they should look professional, formatted like something you would be proud to send to your supervisor.

Be sure to include answers to ALL of the sub-questions in each question.

You should also create a STATA log file that contains the commands and results of your final analysis. Append the log to your responses.

STATA should not be a barrier to your successful completion of this assignment. If you are struggling with STATA, let me know. In particular, if you tell me exactly what you want to get STATA to do, I can help you with the code to accomplish that. In other words, I care much more that you understand the concept.

Please note: You are welcome to complete the problem set with one partner. You team will turn in one memo.

How and what do I turn in?

  1. If you work in a pair please only turn in ONE assignment – it does not matter who turns it in.
  2. You should turn in a Word document (not a PDF).
    1. Provide numbered answers in a separate blank word document (not this doc).
    1. Append your log file to the end of this same document.
  3. Upload your answers/log file in a single document to the assignments page in Canvas. Please do NOT email me your answers, even if they are late.
  4. Please put names at the very END of the document
  5. Please

THE DATA

You will be using the data from the NSW Demonstration RCT, available in nsw.dta. This assignment will focus on African-American males, the group most affected by participation. Individuals were randomly assigned to the treatment and control groups in 1977. The earn_78 variable, which represents the individual’s earnings in 1978, is your outcome. The nsw variable indicates whether the individual received the NSW program or not (nsw=1 for individuals in the treatment group, 0 for individuals in the control group). It also includes an indicator for whether the individual dropped out of the program (noncompleter=1 for individuals who dropped out, 0 for individuals who completed the program, and is missing for individuals in the control group). All other predictors are pre-treatment covariates.  The attached codebook explains each variable.

THE TASK

In this assignment, you will estimate the effect of participating in the NSW program on individuals’ earnings in 1978.

  1. {10 points} Open the nsw.dta dataset in STATA. Create a table that compares the means of all pre-treatment covariates (age, educ, married, earn_75) for individuals in the treatment group to those in the control group. Do the treatment and control groups appear to be similar on these four characteristics? [NOTE: You do not need to create your table in STATA, although you are welcome to. You can do it in Word or Excel if you prefer. It should be a nice-looking table, like you might present to your supervisor.]
  • {10 points} Conceptually, what is the counterfactual in this study? How will we get an estimate of the average value of the counterfactual outcomes for the treatment group?
  • {5 points} Write out a regression model that would allow you to compare the average level of earnings in 1978 (the outcome) in the treatment and control groups. What parameter represents the average treatment effect?
  • {5 points} Fit your model in STATA. What is the effect of the program on 1978 earnings? Based on your results, would you reject the null hypothesis that the program has no effect?
  • {10 points} Add the students’ years of education and their 1975 earnings to your regression model and fit it in STATA [NOTE: you can assume that the functional form of these predictors is correct, and you do not need to transform them].
    • Why might you want to include pre-treatment covariates in this analysis?
    • How are your results different than in (3)?
    • Is this an unbiased estimate of the treatment effect (you can assume that the randomization was done correctly)? Why or why not?
  • {10 points} One hypothesis holds that the NSW program might have a bigger effect on the earnings of married individuals (who might be more motivated to obtain and hold a job) than non-married folks. Modify your regression model in (5) to test this hypothesis, that the effect of the program is different for married and non-married individuals. Write out your model. Fit the model in STATA, and present and interpret your results. What is the estimated treatment effect for married participants? For non-married participants?
VariableCodingDescription
nsw1=participated in NSW 0=did not participateTreatment indicator
ageYearsYears old
educYearsYears of schooling
married1=married 0=not marriedMarriage indicator
earn_75DollarsEarnings in 1975
earn_78DollarsEarnings in 1978
noncompleter1=started program but did not complete it0=completed program.=did not participate in NSWProgram completion indicator

CoursePear™是一家服务全球留学生的专业代写。
—-我们专注提供高质靠谱的美国、加拿大、英国、澳洲、新西兰代写服务。
—-我们专注提供Essay、统计、金融、CS、经济、数学等覆盖100+专业的作业代写服务。

STATA代写
STATA代写

CoursePear™提供各类学术服务，Essay代写，Assignment代写，Exam / Quiz助攻，Dissertation / Thesis代写，Problem Set代做等。