Breakout1

CS112_11_1_1.R

# If you are copy/pasting this to an R Notebook, select "Raw" from the GitHub page before you copy.

# Replication file of Section 5 in
# Iacus, King, Porro (2011), Multivariate Matching Methods 
# That Are Monotonic Imbalance Bounding, JASA, V 106, N. 493,
# p. 345-361

foo <- read.csv(url("https://course-resources.minerva.kgi.edu/uploaded_files/mke/00089202-1711/daughters.csv"))

# Table 1 in Ebonya
# Indep. vars
# anygirls: Any female children
# ngirls: Number of female children
# totchi: Total number of children

# white: White=1, 0 otherwise
# female: Female=1, 0 otherwise
# age: age
# srvlng: Service length (years)

# reg1 - reg9 are regional binary variables
# binary religious variables: none, Protestant, Catholic, OtherC, OtherR
# binary party variables: Dems, Repubs, OthParty
# DEPENDENT VARIABLE: nowtot

#### BREAKOUT 1 ####
## Perform genetic matching with the following variables:
## Dems, Repubs, Christian, age, srvlng, demvote
## Tr  = hasgirls
## Set pop.size = 100, set nboots = 250... complete the instructions below...

set.seed(2324); genout <- GenMatch(Tr = , X = cbind(), pop.size = , nboots = )

mout <- Match(Tr = , X = cbind(), Weight.matrix = )

MatchBalance( ~ , match.out = )

## If you obtain really high balance, consider rerunning with M = 2 or 3...
## If/when you are satisfied by balance achieved, then rerun Match() with Y included 
## and obtain the treatment effect estimate, the standard error, and the confidence interval.

mout <- Match(Y = , Tr = , 
              X = cbind(), Weight.matrix = )

summary(mout)

## If you get to the end and have extra time, you can try changing X (remember to change
## it in GenMatch, in Match, and also in the formula for MatchBalance().
## For example, you could try to control more finely for religion. You could try to control
## for race (e.g., binary "white" variable), or for region (reg1, reg2, etc.) 

## IMPORTANT: SAVE YOUR mout object!
## Follow this protocol: save(object_name, file = "object_filename") to save to working directory
## e.g., if you have an object named mmm, save(mmm, file = "mmm")
## To retrieve your object, you would type: load("object_filename")

`
https://gist.github.com/diamonaj/555eb605698b055fa76c1404c6c44bb7

If you are copy/pasting this to an R Notebook, select "Raw" from the GitHub page before you copy.

Replication file of Section 5 in

Iacus, King, Porro (2011), Multivariate Matching Methods

That Are Monotonic Imbalance Bounding, JASA, V 106, N. 493,

p. 345-361

foo <- read.csv(url("https://course-resources.minerva.kgi.edu/uploaded_files/mke/00089202-1711/daughters.csv"))

Table 1 in Ebonya

Indep. vars

anygirls: Any female children

ngirls: Number of female children

totchi: Total number of children

white: White=1, 0 otherwise

female: Female=1, 0 otherwise

age: age

srvlng: Service length (years)

reg1 - reg9 are regional binary variables

binary religious variables: none, Protestant, Catholic, OtherC, OtherR

binary party variables: Dems, Repubs, OthParty

DEPENDENT VARIABLE: nowtot

BREAKOUT 1

Perform genetic matching with the following variables:

Dems, Repubs, Christian, age, srvlng, demvote

Tr = hasgirls

Set pop.size = 20, set nboots = 200... complete the instructions below...

set.seed(2324);
genout <- GenMatch(Tr = foo$hasgirls, X = cbind(foo$Dems, foo$Repubs, foo$Christian, foo$age, foo$srvlng, foo$demvote), pop.size = 20, nboots = 200)

mout <- Match(Tr = foo$hasgirls, X = cbind(foo$Dems, foo$Repubs, foo$Christian, foo$age, foo$srvlng, foo$demvote), Weight.matrix = genout)

MatchBalance( foo$hasgirls ~ foo$Dems + foo$Repubs + foo$Christian + foo$age + foo$srvlng + foo$demvote + match.out = mout)

If you obtain really high balance, consider rerunning with M = 2 or 3...

If/when you are satisfied by balance achieved, then rerun Match() with Y included

and obtain the treatment effect estimate, the standard error, and the confidence interval.

mout <- Match(Y = foo$nowtot, Tr = foo$hasgirls, X = cbind(foo$Dems, foo$Repubs, foo$Christian, foo$age, foo$srvlng, foo$demvote), Weight.matrix = mout)

summary(mout)

If you get to the end and have extra time, you can try changing X (remember to change

it in GenMatch, in Match, and also in the formula for MatchBalance().

For example, you could try to control more finely for religion. You could try to control

for race (e.g., binary "white" variable), or for region (reg1, reg2, etc.)

IMPORTANT: SAVE YOUR mout object!

Follow this protocol: save(object_name, file = "object_filename") to save to working directory

e.g., if you have an object named mmm, save(mmm, file = "mmm")

To retrieve your object, you would type: load("object_filename")

`