anirudhjayaraman/HW02.R

## HW02.R
# Set working directory to local directory where the data is kept
setwd("~/IGIDR/Development Economics - MIT/Homework Assignment 02")
# read data
migueldata = read.csv("ted_miguel_worms.csv", header = TRUE)
attach(migueldata)
# Question 6
# How many observations are there per pupil? (Enter a whole number of 0 or higher)?
length(migueldata$pupid)
length(unique(migueldata$pupid))
# Question 7
# What percentage of the pupils are boys? (Answers within 0.50 percentage points of the correct answer will be accepted. For instance, 67 would be accepted if the correct answer is 67.45%)
mean(sex, na.rm = TRUE)
# Question 8
# What percentage of pupils took the deworming pill in 1998? (Answers within 0.50 percentage points of the correct answer will be accepted. For instance, 67 would be accepted if the correct answer is 67.45%)
mean(pill98, na.rm = TRUE)
# Question 9
# Was the percentage of schools assigned to treatment in 1998 greater than or less than the percentage of pupils that actually took the deworming pill in 1998?
mean(treat_sch98, na.rm = TRUE)
mean(treat_sch98, na.rm = TRUE) > mean(pill98, na.rm = TRUE) # Ans = Greater Than
# Question 10
# Which of the following variables from the dataset are dummy variables? (Check all that apply.)
summary(migueldata)
# Question 11
# Using the data, find and enter the difference in outcomes (Y: school participation) between students who took the pill and students who did not in 1998. (Enter your answer as a difference in proportions. For instance, if the proportion in one group is 0.61 and the proportion in the other group is 0.54, enter 0.07. Answers within 0.05 of the correct answer will be accepted. For instance, 0.28 would be accepted if the correct answer is 0.33.)
took_pill_98 = mean(migueldata[migueldata$pill98 == 1,]$totpar98, na.rm = TRUE)
no_pill_98 = mean(migueldata[migueldata$pill98 == 0,]$totpar98, na.rm = TRUE)
diff = took_pill_98 - no_pill_98
diff
# Question 12
# Since schools were randomly assigned to the deworming treatment group, the estimate calculated in the previous answer is an unbiased estimate of taking the pill on school attendance.
# False
# Explanation
# The estimated impact of 13 percentage points calculated in the previous answer might not be a good estimate of the effect of taking the pill. Many students in the randomly assigned treatment schools did not actually take the pills, so those who took the pills would not have been randomly selected at all. For instance, kids who attend school more anyway might have been more likely to be there when the pills were handed out, meaning that omitted variables would be correlated with taking the pill and future school attendance. This would bias the estimate upward i.e. the 13 percentage point difference might overstate the impact of deworming on attendance.
# Question 13
# Using the data, find and enter the difference in outcomes (Y: school participation) between students in treatment schools and students not in treatment schools in 1998, regardless of whether or not they actually took the pill. (Enter your answer as a difference in proportions. For instance, if the proportion in one group is 0.61 and the proportion in the other group is 0.54, enter 0.07. Answers within 0.05 of the correct answer will be accepted. For instance, 0.28 would be accepted if the correct answer is 0.33.)
in_treatment_sch = mean(migueldata[migueldata$treat_sch98 == 1,]$totpar98, na.rm = TRUE)
non_treatment_sch = mean(migueldata[migueldata$treat_sch98 == 0,]$totpar98, na.rm = TRUE)
diff_treatment_sch = in_treatment_sch - non_treatment_sch
diff_treatment_sch
# Question 14
# Using the data, calculate the difference in the probability of taking the pill given that a student was in a treatment school and the probability of taking it if a student was not in a treatment school. (Enter your answer as a difference in proportions. For instance, if the proportion in one group is 0.61 and the proportion in the other group is 0.54, enter 0.07. Answers within 0.05 of the correct answer will be accepted. For instance, 0.28 would be accepted if the correct answer is 0.33.)
pr_pill_treatment_sch = mean(migueldata[migueldata$treat_sch98 == 1,]$pill98, na.rm = TRUE)
pr_pill_no_treatment_sch = mean(migueldata[migueldata$treat_sch98 == 0,]$pill98, na.rm = TRUE)
diff_pr_pill_treatment_sch = pr_pill_treatment_sch - pr_pill_no_treatment_sch

# Question 15
# Using the data, derive the Wald Estimator of taking the pill on school attendance. (Enter your answer as a difference in proportions. For instance, if the proportion in one group is 0.61 and the proportion in the other group is 0.54, enter 0.07. Answers within 0.05 of the correct answer will be accepted. For instance, 0.28 would be accepted if the correct answer is 0.33.)
waldRatio = diff_treatment_sch/diff_pr_pill_treatment_sch
waldRatio
	# Set working directory to local directory where the data is kept
	setwd("~/IGIDR/Development Economics - MIT/Homework Assignment 02")
	# read data
	migueldata = read.csv("ted_miguel_worms.csv", header = TRUE)
	attach(migueldata)
	# Question 6
	# How many observations are there per pupil? (Enter a whole number of 0 or higher)?
	length(migueldata$pupid)
	length(unique(migueldata$pupid))
	# Question 7
	# What percentage of the pupils are boys? (Answers within 0.50 percentage points of the correct answer will be accepted. For instance, 67 would be accepted if the correct answer is 67.45%)
	mean(sex, na.rm = TRUE)
	# Question 8
	# What percentage of pupils took the deworming pill in 1998? (Answers within 0.50 percentage points of the correct answer will be accepted. For instance, 67 would be accepted if the correct answer is 67.45%)
	mean(pill98, na.rm = TRUE)
	# Question 9
	# Was the percentage of schools assigned to treatment in 1998 greater than or less than the percentage of pupils that actually took the deworming pill in 1998?
	mean(treat_sch98, na.rm = TRUE)
	mean(treat_sch98, na.rm = TRUE) > mean(pill98, na.rm = TRUE) # Ans = Greater Than
	# Question 10
	# Which of the following variables from the dataset are dummy variables? (Check all that apply.)
	summary(migueldata)
	# Question 11
	# Using the data, find and enter the difference in outcomes (Y: school participation) between students who took the pill and students who did not in 1998. (Enter your answer as a difference in proportions. For instance, if the proportion in one group is 0.61 and the proportion in the other group is 0.54, enter 0.07. Answers within 0.05 of the correct answer will be accepted. For instance, 0.28 would be accepted if the correct answer is 0.33.)
	took_pill_98 = mean(migueldata[migueldata$pill98 == 1,]$totpar98, na.rm = TRUE)
	no_pill_98 = mean(migueldata[migueldata$pill98 == 0,]$totpar98, na.rm = TRUE)
	diff = took_pill_98 - no_pill_98
	diff
	# Question 12
	# Since schools were randomly assigned to the deworming treatment group, the estimate calculated in the previous answer is an unbiased estimate of taking the pill on school attendance.
	# False
	# Explanation
	# The estimated impact of 13 percentage points calculated in the previous answer might not be a good estimate of the effect of taking the pill. Many students in the randomly assigned treatment schools did not actually take the pills, so those who took the pills would not have been randomly selected at all. For instance, kids who attend school more anyway might have been more likely to be there when the pills were handed out, meaning that omitted variables would be correlated with taking the pill and future school attendance. This would bias the estimate upward i.e. the 13 percentage point difference might overstate the impact of deworming on attendance.
	# Question 13
	# Using the data, find and enter the difference in outcomes (Y: school participation) between students in treatment schools and students not in treatment schools in 1998, regardless of whether or not they actually took the pill. (Enter your answer as a difference in proportions. For instance, if the proportion in one group is 0.61 and the proportion in the other group is 0.54, enter 0.07. Answers within 0.05 of the correct answer will be accepted. For instance, 0.28 would be accepted if the correct answer is 0.33.)
	in_treatment_sch = mean(migueldata[migueldata$treat_sch98 == 1,]$totpar98, na.rm = TRUE)
	non_treatment_sch = mean(migueldata[migueldata$treat_sch98 == 0,]$totpar98, na.rm = TRUE)
	diff_treatment_sch = in_treatment_sch - non_treatment_sch
	diff_treatment_sch
	# Question 14
	# Using the data, calculate the difference in the probability of taking the pill given that a student was in a treatment school and the probability of taking it if a student was not in a treatment school. (Enter your answer as a difference in proportions. For instance, if the proportion in one group is 0.61 and the proportion in the other group is 0.54, enter 0.07. Answers within 0.05 of the correct answer will be accepted. For instance, 0.28 would be accepted if the correct answer is 0.33.)
	pr_pill_treatment_sch = mean(migueldata[migueldata$treat_sch98 == 1,]$pill98, na.rm = TRUE)
	pr_pill_no_treatment_sch = mean(migueldata[migueldata$treat_sch98 == 0,]$pill98, na.rm = TRUE)
	diff_pr_pill_treatment_sch = pr_pill_treatment_sch - pr_pill_no_treatment_sch

	# Question 15
	# Using the data, derive the Wald Estimator of taking the pill on school attendance. (Enter your answer as a difference in proportions. For instance, if the proportion in one group is 0.61 and the proportion in the other group is 0.54, enter 0.07. Answers within 0.05 of the correct answer will be accepted. For instance, 0.28 would be accepted if the correct answer is 0.33.)
	waldRatio = diff_treatment_sch/diff_pr_pill_treatment_sch
	waldRatio