To help you complete this asesment, you may use resources (e.g., books, lecture notes,
online guides, etc.) and/or work with other students in the clas, but each person must hand in
their own work (i.e., you are not alowed to copy text from someone else—that is
plagiarism!).
R Markdown
You should upload your R Markdown file directly to MOLE, which wil provide me with
your R script, results, and any notes that you include to answer each question.
To create an R Markdown file, you should instal the rmarkdown package via ‘pacman’ from
the command line:
> pacman::p_load(rmarkdown)
And then render your R script. from the command line (NOT from the R script. you wish to
run!):
> rmarkdown::render("your R script. file location and name here.R")
Make sure you comment out (place a # in front of) any lines that wil cause errors for the
markdown proces (e.g., ‘setwd()’, ‘file.choose()’, ‘rmarkdown::render()’, comments, etc.).
You should start your R script. with the following header:
#' ---
#' title: "SMI 606 Assessment 2"
#' author: "Anonymous"
#' date: "`r format(Sys.Date())`"
#' ---
Remember, that at its most basic form, R Markdown is just an R script. in which comments #
are replaced by #' as in the following:
#' This is an example of a heading comment using the R Markdown formatting.
Mexican Elections and the Effect of the Conditional Cash Transfer Program
In this asesment, you’ll be analyzing data from a study about the electoral impact of
Progresa, Mexico’s conditional cash transfer (CCT) programme. CCT programmes aim to
reduce poverty by making welfare conditional upon the receivers’ actions (i.e., the
government only transfers money to individuals who met specific criteria). For example, the
CCT provided cash payments to families in exchange for regular school atendance, health
clinic visits, and nutrition support. Cash payments were made from the government directly
to families to decrease overhead and corruption. Eligible vilages were randomly asigned to
receive the program either 21 or 6 months before the 2000 Mexican presidential election.
You can think of the treatment as vilages which were asigned to the CCT 21 months before
the election because they would have sufficient time to implement the programe.
The data for this asesment are available as the CSV file progresa.csv (as part of the
QS datasets on MOLE). Each observation in the data represents a precinct, and for each
precinct the file contains information about its treatment status, the outcomes of interest,
socioeconomic indicators, and other precinct characteristics (se below for more details about
the variables included in the dataset). The sample contains precincts that had at most one
participating vilage in the evaluation.
Names and Descriptions of Variables in the CCT Program (Progresa) Data
Name Description
treatment
1 = treatment (precinct contains a vilage that received CCT at 21 months);
0 = control (vilage received CT at 6 months)
pri2000s
PRI votes in the 200 election as a share of precinct population above 18
pri2000v
Official PRI vote share in the 200 election
t2000
Turnout in the 200 election as a share of precinct population above 18
t2000r
Official turnout in the 200 election
pri1994
Total PRI votes in the 194 presidential election
pan1994
Total PAN votes in the 194 presidential election
prd1994
Total PRD votes in the 194 presidential election
pri1994s
Total PRI votes in the 194 election as a share of precinct population above 18
pan1994s
Total PAN votes in the 194 election as a share of precinct population above 18
prd1994s
Total PRD votes in the 1994 election as a share of precinct population above 18
pri1994v
Official PRI vote share in the 1994 election
pan1994v
Official PAN vote share in the 1994 election
prd1994v
Official PRD vote share in the 1994 election
t1994
Turnout in the 194 election as a share of precinct population above 18
t1994r
Official turnout in the 194 election
votos1994
Total votes cast in the 194 presidential election
avgpoverty
Precinct Avg of Vilage Poverty Index
pobtot1994
Total Population in the precinct
villages
Number of vilages in the precinct
Imagine that you’ve been hired as a consultant by the PRI-controlled Mexican government to
determine the electoral impact of the CCT programe. In particular, party leaders want to
know whether exposure to the CCT increased turnout and vote share for the incumbent party.
They provided you the progresa.csv dataset and asked you to answer the following
questions.
1. To begin, which measures of 1) voter turnout and 2) support for the PRI (in 2000)
would you use for your analyses (list the specific variable names)? Explain your
reasoning and discuss any statistical analyses that you conducted to inform. your
decision. Do the diferent versions of turnout and party support measure the same
thing? Discuss.
2. The government believes that poor voters should have turned out in higher numbers in
the 2000 election and cast their balot for the PRI. Estimate the sample average
treatment efects (SATEs) of exposure to the CCT using a diference-in-means
estimator. Did the CCT programe increase voter turnout and PRI vote share in the
2000 election? Discuss your results.
3. Next, using a linear regresion approach, estimate the same sample average treatment
efects by regresing each outcome variable—turnout and vote share—on the
treatment dummy variable (i.e., you wil have two separate regresion models).
Interpret your results and compare them to your findings in Question 2.
4. The dataset also includes results from the prior election. Is there any evidence that
poorer voters would support the PRI (and hence one of the reasons for introducing
Progresa)? Examine whether precinct poverty is asociated with support for any of
the three major political parties in the 1994 election (check the correlations). Now
regres each party’s vote share on average poverty levels (rescaled from 0 to 1).
Interpret the coeficient for average poverty for each of the three regresion models.
Do precincts with higher levels of poverty tend to vote for any particular party?
Discuss.
5. Do you se any isues with the 1994 voter turnout measures? Discuss how you
investigated this question and how it might impact your earlier analyses.