ECON 523

Logo
photo: World Bank/Peter Kapuscinski (2015)

Instructor:
Pamela Jakiela

home
syllabus
schedule
stata


Empirical Exercise 1

This exercise makes use of the data set E1-CohenEtAl-data.dta, a subset of the data used in the paper Price Subsidies, Diagnostic Tests, and Targeting of Malaria Treatment: Evidence from a Randomized Controlled Trial by Jessica Cohen, Pascaline Dupas, and Simone Schaner, published in the American Economic Review in 2015. The authors examine behavioral responses to various discounts (“subsidies”) for malaria treatment, called “artemisinin combination therapy” or “ACT.” An overview of the randomized evalaution is available here.

The goal of this exercise is to review the different approaches to testing for differences in means across groups defined by a dummy variable, for example a randomly-assigned treatment. We will review the Stata command ttest, regress, and ci.

You can access the in-class activity (below) as a do file or pdf.

You can also access the main empirical exercise (also below) as a do file or pdf.


Getting Started

Create a do file that contains the following preliminaries:

// ECON 523:  PROGRAM EVALUATION FOR INTERNATIONAL DEVELOPMENT
// PROFESSOR PAMELA JAKIELA

/* preliminary stuff*/
clear all 
set more off
set seed 12345

You’ll also want to include a command to change the working directory so that any outputs are saved where you can find them later.

** change working directory as appropriate to where you want to save
cd "C:\Users\CookieMonster\Dropbox\ECON-523\exercises\E1-selection"

Because the data set is available on github, you can simply download it every time you want to use it. The following code in the do file will do this:

** load the data from the course website
webuse set https://pjakiela.github.io/ECON523/exercises
webuse E1-CohenEtAl-data.dta

If you prefer, you can instead download the data set using the link above and save it directly to your computer. In that case, you would use the use command to open the data set in Stata.


In-Class Activity

Extend your do file as you answer the following questions, so that you can run the code from start to finish and re-generate all your answers.

Question 1

How many variables are in the data set? (hint: use describe, or desc for short)

Question 2

How many observations are in the data set? (hint: use count)

Question 3

What does the variable act_any measure? (hint: use describe or codebook)

Question 4

What is the mean of act_any to three decimal places? (hint: use summarize, or sum for short)

Question 5

How many people received subsidized malaria treatment? (hint: use tabulate, or tab for short)

Question 6

What does the variable c_act measure?

Question 7

What is the mean of the variable c_act?

Question 8

What is the standard deviation of the variable c_act?

Question 9

What is the standard error of the mean of c_act? (hint: use the ci means command)

Question 10

What is the mean level of ACT use among those assigned to the treatment group? (hint: use an if statement)

Question 11

Variables starting with b_ are baseline characteristics (measured before the RCT). Use the summarize command to familiarize yourself with these variables. How many baseline variables are included in the data set? Which ones are missing data for some households in the sample? (hint: use sum b_*)

Question 12

We’re going to look at selection bias by comparing the level of educational attainment among households that choose to use ACT treatment when they have malaria. Use the ci means command to obtain the mean and standard error of b_h_edu when c_act==1 and when c_act==0. Using these quantities, calculate the estimated difference in means and its standard error.

Question 13

Now compare your results to what you obtain using the the ttest command.

Question 14

Why does the standard error you calculated using the output from ci means not match the results of the ttest command? How can you modify the ttest command so that your results line up with your answer to Question 12? (hint: look at the help file for ttest if needed)

Question 15

Confirm that you can also replicate your results from Q12 using the regress command withe the , vce(hc2) option.


Empirical Exercise

Create a new do file (with the same preliminaries at the top) to answer the following questions, so that you can run the code from start to finish and re-generate all your answers. Some questions only ask you to provide the correct Stata code; when a question asks for a numeric or verbal response, please provide your answer in a comment in your do file. Upload your finished do file to gradescope once you have finished.

Question 1: One Treatment Dummy

Part (a)

Summarize the mean level of ART use (the variable c_act) in the randomly assigned treatment group (act_any==1) and the randomly assigned comaprison group (act_any==0).

Part (b)

Conduct a t-test of the hypothesis that treatment (act_any) does not impact the likelihood of using ARTs (using the ttest command).

Part (c)

Test the hypothesis that treatment (act_any) does not impact the likelihood of using ARTs using the regress command.

Question 2: Multiple Treatments

Part (a)

The variable coartemprice indicates the randomly-assigned ACT price (and, implicitly, the associated level of price subsidy). What price/subsidy levels are included in the experiment?

Part (b)

If you place the code

bysort coartemprice:

before the summarize command, Stata will summarize your value of interest separately for each observed value of the variable coartemprice. What is the mean level of ART use at each subsidy level, and how do these levels compare to the level observed in the control group?

Part (c)

Now regress c_act on the dummies act40, act60, and act100, which indicate the three different randomly-assigned subsidy levels in the RCT. What do you expect the regression coefficients to be (based on your answer to Question 2b). Do they observed coefficients match your expectations?

Question 3: Pooling Treatment Arms

Part (a)

Can you figure out a way to get Stata to tell you how many treated observations are in each (of the three) treatment arms using a single line of Stata code?

Part (b)

Stata’s display command is useful for doing arithmetic. Calculate a weighted average of the regression coefficients from Question 2c, where the weights are the proportion of treated observations in each of the three arms.

Part (c)

Where have you seen this coefficient before?


This exercise is part of the module Selection Bias and the Experiental Ideal.